223031012
/
DAMASK_EICMD
mirror of https://github.com/achalhp/DAMASK_EICMD.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
DAMASK_EICMD/code/mesh.f90

5328 lines
227 KiB
Fortran
Raw Normal View History

introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
! Copyright 2011-13 Max-Planck-Institut für Eisenforschung GmbH
added copyright text to all f90 (free) format files
2011-04-04 19:39:54 +05:30
!
! This file is part of DAMASK,
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
! the Düsseldorf Advanced Material Simulation Kit.
added copyright text to all f90 (free) format files
2011-04-04 19:39:54 +05:30
!
! DAMASK is free software: you can redistribute it and/or modify
! it under the terms of the GNU General Public License as published by
! the Free Software Foundation, either version 3 of the License, or
! (at your option) any later version.
!
! DAMASK is distributed in the hope that it will be useful,
! but WITHOUT ANY WARRANTY; without even the implied warranty of
! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
! GNU General Public License for more details.
!
! You should have received a copy of the GNU General Public License
! along with DAMASK. If not, see <http://www.gnu.org/licenses/>.
!
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
added version information to all files do NOT edit text like this: $Id: constitutive_phenopowerlaw.f90 406 2009-08-31 14:13:10Z MPIE\f.roters $
2009-08-31 20:39:15 +05:30
!* $Id$
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!> @author Franz Roters, Max-Planck-Institut für Eisenforschung GmbH
!> @author Philip Eisenlohr, Max-Planck-Institut für Eisenforschung GmbH
!> @author Christoph Koords, Max-Planck-Institut für Eisenforschung GmbH
!> @author Martin Diehl, Max-Planck-Institut für Eisenforschung GmbH
!> @author Krishna Komerla, Max-Planck-Institut für Eisenforschung GmbH
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!> @brief Sets up the mesh for the solvers MSC.Marc, Abaqus and the spectral solver
!--------------------------------------------------------------------------------------------------
module mesh
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
use, intrinsic :: iso_c_binding
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use prec, only: pReal, pInt
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
added function mesh_FEtoCPelement(FEid)
2007-03-21 21:48:33 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
private
divergence_correction for basic solver variants has now 3 possibilities: 0: uncorrected, slope per sidelength (physical dimension) e = res/dim 1: corrected by sidelength, slope per unitlength e = res/1 2: corrected such that distance between FPs e = 1 alway regarding the medium length of x,y,z direction
2013-01-16 16:10:53 +05:30
integer(pInt), public, protected :: &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_NcpElems, & !< total number of CP elements in mesh
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
mesh_NelemSets, &
mesh_maxNelemInSet, &
mesh_Nmaterials, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_Nnodes, & !< total number of nodes in mesh
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_Ncellnodes, & !< total number of cell nodes in mesh (including duplicates)
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_Ncells, & !< total number of cells in mesh
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_maxNnodes, & !< max number of nodes in any CP element
mesh_maxNips, & !< max number of IPs in any CP element
mesh_maxNipNeighbors, & !< max number of IP neighbors in any CP element
mesh_maxNsharedElems, & !< max number of CP elements sharing a node
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_maxNcellnodes, & !< max number of cell nodes in any CP element
preparation for introduction of plain mode and restructuring of calc modes for CPFEM_general
2013-02-28 02:11:14 +05:30
mesh_Nelems !< total number of elements in mesh
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
divergence_correction for basic solver variants has now 3 possibilities: 0: uncorrected, slope per sidelength (physical dimension) e = res/dim 1: corrected by sidelength, slope per unitlength e = res/1 2: corrected such that distance between FPs e = 1 alway regarding the medium length of x,y,z direction
2013-01-16 16:10:53 +05:30
integer(pInt), dimension(:,:), allocatable, public, protected :: &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_element, & !< FEid, type(internal representation), material, texture, node indices as CP IDs
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_sharedElem, & !< entryCount and list of elements containing node
mesh_nodeTwins !< node twins are surface nodes that lie exactly on opposite sides of the mesh (surfaces nodes with equal coordinate values in two dimensions)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
divergence_correction for basic solver variants has now 3 possibilities: 0: uncorrected, slope per sidelength (physical dimension) e = res/dim 1: corrected by sidelength, slope per unitlength e = res/1 2: corrected such that distance between FPs e = 1 alway regarding the medium length of x,y,z direction
2013-01-16 16:10:53 +05:30
integer(pInt), dimension(:,:,:,:), allocatable, public, protected :: &
third entry in mesh_ipNeighborhood stores the neighbor_index that points from each neighbor back to the central ip; needed in nonlocal model
2012-10-24 19:33:02 +05:30
mesh_ipNeighborhood !< 6 or less neighboring IPs as [element_num, IP_index, neighbor_index that points to me]
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
real(pReal), public, protected :: &
mesh_unitlength !< physical length of one unit in mesh
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
real(pReal), dimension(:,:), allocatable, public :: &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_node, & !< node x,y,z coordinates (after deformation! ONLY FOR MARC!!!)
mesh_cellnode !< cell node x,y,z coordinates (after deformation! ONLY FOR MARC!!!)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
divergence_correction for basic solver variants has now 3 possibilities: 0: uncorrected, slope per sidelength (physical dimension) e = res/dim 1: corrected by sidelength, slope per unitlength e = res/1 2: corrected such that distance between FPs e = 1 alway regarding the medium length of x,y,z direction
2013-01-16 16:10:53 +05:30
real(pReal), dimension(:,:), allocatable, public, protected :: &
mesh_ipVolume, & !< volume associated with IP (initially!)
mesh_node0 !< node x,y,z coordinates (initially!)
real(pReal), dimension(:,:,:), allocatable, public, protected :: &
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
mesh_ipArea !< area of interface to neighboring IP (initially!)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
divergence_correction for basic solver variants has now 3 possibilities: 0: uncorrected, slope per sidelength (physical dimension) e = res/dim 1: corrected by sidelength, slope per unitlength e = res/1 2: corrected such that distance between FPs e = 1 alway regarding the medium length of x,y,z direction
2013-01-16 16:10:53 +05:30
real(pReal), dimension(:,:,:), allocatable, public :: &
mesh_ipCoordinates !< IP x,y,z coordinates (after deformation!)
real(pReal),dimension(:,:,:,:), allocatable, public, protected :: &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_ipAreaNormal !< area normal of interface to neighboring IP (initially!)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
divergence_correction for basic solver variants has now 3 possibilities: 0: uncorrected, slope per sidelength (physical dimension) e = res/dim 1: corrected by sidelength, slope per unitlength e = res/1 2: corrected such that distance between FPs e = 1 alway regarding the medium length of x,y,z direction
2013-01-16 16:10:53 +05:30
logical, dimension(3), public, protected :: mesh_periodicSurface !< flag indicating periodic outer surfaces (used for fluxes)
preparation for introduction of plain mode and restructuring of calc modes for CPFEM_general
2013-02-28 02:11:14 +05:30
removed #define Marc and substituted #ifdef Marc by #ifdef Marc4DAMASK which is defined by either the submit scripts or DAMASK_marcXXXX.f90
2013-04-30 15:19:30 +05:30
#ifdef Marc4DAMASK
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), private :: &
hypoelasticTableStyle, & !< Table style (Marc only)
initialcondTableStyle !< Table style (Marc only)
#endif
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), dimension(2), private :: &
mesh_maxValStateVar = 0_pInt
character(len=64), dimension(:), allocatable, private :: &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_nameElemSet, & !< names of elementSet
mesh_nameMaterial, & !< names of material in solid section
mesh_mapMaterial !< name of elementSet for material
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), dimension(:,:), allocatable, private :: &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_cellnodeParent, & !< cellnode's parent element ID, cellnode's intra-element ID
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_mapElemSet !< list of elements in elementSet
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), dimension(:,:), allocatable, target, private :: &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_mapFEtoCPelem, & !< [sorted FEid, corresponding CPid]
mesh_mapFEtoCPnode !< [sorted FEid, corresponding CPid]
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt),dimension(:,:,:), allocatable, private :: &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_cell !< cell connectivity for each element,ip/cell
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(:,:,:), allocatable, private :: &
FE_nodesAtIP, & !< map IP index to node indices in a specific type of element
FE_ipNeighbor, & !< +x,-x,+y,-y,+z,-z list of intra-element IPs and(negative) neighbor faces per own IP in a specific type of element
FE_cell, & !< list of intra-element cell node IDs that constitute the cells in a specific type of element geometry
FE_cellface !< list of intra-cell cell node IDs that constitute the cell faces of a specific type of cell
real(pReal), dimension(:,:,:), allocatable, private :: &
FE_cellnodeParentnodeWeights !< list of node weights for the generation of cell nodes
integer(pInt), dimension(:,:,:,:), allocatable, private :: &
FE_subNodeOnIPFace
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
logical, private :: noPart !< for cases where the ABAQUS input file does not use part/assembly information
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
#ifdef Spectral
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
include 'fftw3.f03'
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
#endif
added function mesh_FEtoCPelement(FEid)
2007-03-21 21:48:33 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! These definitions should actually reside in the FE-solver specific part (different for MARC/ABAQUS)
started envIP building
2007-03-22 20:18:58 +05:30
! Hence, I suggest to prefix with "FE_"
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), parameter, public :: &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_Nelemtypes = 13_pInt, &
FE_Ngeomtypes = 10_pInt, &
FE_Ncelltypes = 4_pInt, &
FE_maxNnodes = 20_pInt, &
FE_maxNips = 27_pInt, &
FE_maxNipNeighbors = 6_pInt, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
FE_maxmaxNnodesAtIP = 8_pInt, & !< max number of (equivalent) nodes attached to an IP
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_maxNmatchingNodesPerFace = 4_pInt, &
FE_maxNfaces = 6_pInt, &
FE_maxNcellnodes = 64_pInt, &
FE_maxNcellnodesPerCell = 8_pInt, &
FE_maxNcellfaces = 6_pInt, &
FE_maxNcellnodesPerCellface = 4_pInt
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
integer(pInt), dimension(FE_Nelemtypes), parameter, public :: FE_geomtype = & !< geometry type of particular element type
int([ &
1, & ! element 6 (2D 3node 1ip)
2, & ! element 125 (2D 6node 3ip)
3, & ! element 11 (2D 4node 4ip)
4, & ! element 27 (2D 8node 9ip)
- added element type 54 for macro (2d 8node 4ip reduced integration) - automatically raising error for unknown element type inside mapping function
2013-04-10 15:08:40 +05:30
3, & ! element 54 (2D 8node 4ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
5, & ! element 134 (3D 4node 1ip)
6, & ! element 157 (3D 5node 4ip)
6, & ! element 127 (3D 10node 4ip)
7, & ! element 136 (3D 6node 6ip)
8, & ! element 117 (3D 8node 1ip)
9, & ! element 7 (3D 8node 8ip)
9, & ! element 57 (3D 20node 8ip)
10 & ! element 21 (3D 20node 27ip)
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
],pInt)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(FE_Ngeomtypes), parameter, public :: FE_celltype = & !< cell type that is used by each geometry type
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
int([ &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1, & ! element 6 (2D 3node 1ip)
2, & ! element 125 (2D 6node 3ip)
2, & ! element 11 (2D 4node 4ip)
2, & ! element 27 (2D 8node 9ip)
3, & ! element 134 (3D 4node 1ip)
4, & ! element 127 (3D 10node 4ip)
4, & ! element 136 (3D 6node 6ip)
4, & ! element 117 (3D 8node 1ip)
4, & ! element 7 (3D 8node 8ip)
4 & ! element 21 (3D 20node 27ip)
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
],pInt)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(FE_Ngeomtypes), parameter, public :: FE_dimension = & !< dimension of geometry type
introduced dedicated calculation of interface normals in IP neighborhood. (2D elements did not work so far..!) fixed definitions of some (2D) interface planes to have all normals pointing outward. removed "int, private :: i" defined for spectral compilation.
2013-02-05 18:57:37 +05:30
int([ &
2, & ! element 6 (2D 3node 1ip)
2, & ! element 125 (2D 6node 3ip)
2, & ! element 11 (2D 4node 4ip)
2, & ! element 27 (2D 8node 9ip)
3, & ! element 134 (3D 4node 1ip)
3, & ! element 127 (3D 10node 4ip)
3, & ! element 136 (3D 6node 6ip)
3, & ! element 117 (3D 8node 1ip)
3, & ! element 7 (3D 8node 8ip)
3 & ! element 21 (3D 20node 27ip)
],pInt)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(FE_Nelemtypes), parameter, public :: FE_Nnodes = & !< number of nodes that constitute a specific type of element
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
int([ &
3, & ! element 6 (2D 3node 1ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
6, & ! element 125 (2D 6node 3ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
4, & ! element 11 (2D 4node 4ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
8, & ! element 27 (2D 8node 9ip)
8, & ! element 54 (2D 8node 4ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
4, & ! element 134 (3D 4node 1ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
5, & ! element 157 (3D 5node 4ip)
10, & ! element 127 (3D 10node 4ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
6, & ! element 136 (3D 6node 6ip)
8, & ! element 117 (3D 8node 1ip)
8, & ! element 7 (3D 8node 8ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
20, & ! element 57 (3D 20node 8ip)
20 & ! element 21 (3D 20node 27ip)
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
],pInt)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(FE_Ngeomtypes), parameter, public :: FE_Nfaces = & !< number of faces of a specific type of element geometry
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
int([ &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
3, & ! element 6 (2D 3node 1ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
3, & ! element 125 (2D 6node 3ip)
4, & ! element 11 (2D 4node 4ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
4, & ! element 27 (2D 8node 9ip)
4, & ! element 134 (3D 4node 1ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
4, & ! element 127 (3D 10node 4ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
5, & ! element 136 (3D 6node 6ip)
6, & ! element 117 (3D 8node 1ip)
6, & ! element 7 (3D 8node 8ip)
6 & ! element 21 (3D 20node 27ip)
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
],pInt)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(FE_Ngeomtypes), parameter, private :: FE_NmatchingNodes = & !< number of nodes that are needed for face matching in a specific type of element geometry
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
int([ &
3, & ! element 6 (2D 3node 1ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
3, & ! element 125 (2D 6node 3ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
4, & ! element 11 (2D 4node 4ip)
4, & ! element 27 (2D 8node 9ip)
4, & ! element 134 (3D 4node 1ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
4, & ! element 127 (3D 10node 4ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
6, & ! element 136 (3D 6node 6ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
8, & ! element 117 (3D 8node 1ip)
8, & ! element 7 (3D 8node 8ip)
8 & ! element 21 (3D 20node 27ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
],pInt)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(FE_maxNfaces,FE_Ngeomtypes), parameter, private :: &
FE_NmatchingNodesPerFace = & !< number of matching nodes per face in a specific type of element geometry
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([ &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
2,2,2,0,0,0, & ! element 6 (2D 3node 1ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
2,2,2,0,0,0, & ! element 125 (2D 6node 3ip)
2,2,2,2,0,0, & ! element 11 (2D 4node 4ip)
2,2,2,2,0,0, & ! element 27 (2D 8node 9ip)
3,3,3,3,0,0, & ! element 134 (3D 4node 1ip)
3,3,3,3,0,0, & ! element 127 (3D 10node 4ip)
3,4,4,4,3,0, & ! element 136 (3D 6node 6ip)
4,4,4,4,4,4, & ! element 117 (3D 8node 1ip)
4,4,4,4,4,4, & ! element 7 (3D 8node 8ip)
4,4,4,4,4,4 & ! element 21 (3D 20node 27ip)
],pInt),[FE_maxNipNeighbors,FE_Ngeomtypes])
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(FE_maxNmatchingNodesPerFace,FE_maxNfaces,FE_Ngeomtypes), &
parameter, private :: FE_face = & !< List of node indices on each face of a specific type of element geometry
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
reshape(int([&
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1,2,0,0 , & ! element 6 (2D 3node 1ip)
2,3,0,0 , &
3,1,0,0 , &
corrected mesh_build_elements (unit) so that state variables are read correctly from input file
2007-04-25 13:03:24 +05:30
0,0,0,0 , &
added element types 11 (4node 2D linear full) and 27 (8node 2D quadratic full)
2007-10-12 19:18:29 +05:30
0,0,0,0 , &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
0,0,0,0 , &
1,2,0,0 , & ! element 125 (2D 6node 3ip)
2,3,0,0 , &
3,1,0,0 , &
0,0,0,0 , &
0,0,0,0 , &
0,0,0,0 , &
1,2,0,0 , & ! element 11 (2D 4node 4ip)
added element types 11 (4node 2D linear full) and 27 (8node 2D quadratic full)
2007-10-12 19:18:29 +05:30
2,3,0,0 , &
3,4,0,0 , &
4,1,0,0 , &
0,0,0,0 , &
0,0,0,0 , &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1,2,0,0 , & ! element 27 (2D 8node 9ip)
Element type 21 (20-node hexagon with triquadric interpolation functions) is now included. 'mesh_get_nodeElemDimensions', mesh_build_elements', 'mesh_build_sharedElems' now allow for more than 14 nodes per elements which are then written in more than one line in the 'connectivity' field of the input file
2009-04-02 18:30:51 +05:30
2,3,0,0 , &
3,4,0,0 , &
4,1,0,0 , &
added element types 11 (4node 2D linear full) and 27 (8node 2D quadratic full)
2007-10-12 19:18:29 +05:30
0,0,0,0 , &
added two more element types (157 and 136)
2008-06-17 14:41:54 +05:30
0,0,0,0 , &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1,2,3,0 , & ! element 134 (3D 4node 1ip)
added two more element types (157 and 136)
2008-06-17 14:41:54 +05:30
1,4,2,0 , &
2,3,4,0 , &
1,3,4,0 , &
0,0,0,0 , &
0,0,0,0 , &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1,2,3,0 , & ! element 127 (3D 10node 4ip)
1,4,2,0 , &
2,4,3,0 , &
1,3,4,0 , &
0,0,0,0 , &
0,0,0,0 , &
1,2,3,0 , & ! element 136 (3D 6node 6ip)
added two more element types (157 and 136)
2008-06-17 14:41:54 +05:30
1,4,5,2 , &
2,5,6,3 , &
1,3,6,4 , &
4,6,5,0 , &
Element type 21 (20-node hexagon with triquadric interpolation functions) is now included. 'mesh_get_nodeElemDimensions', mesh_build_elements', 'mesh_build_sharedElems' now allow for more than 14 nodes per elements which are then written in more than one line in the 'connectivity' field of the input file
2009-04-02 18:30:51 +05:30
0,0,0,0 , &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1,2,3,4 , & ! element 117 (3D 8node 1ip)
Element type 21 (20-node hexagon with triquadric interpolation functions) is now included. 'mesh_get_nodeElemDimensions', mesh_build_elements', 'mesh_build_sharedElems' now allow for more than 14 nodes per elements which are then written in more than one line in the 'connectivity' field of the input file
2009-04-02 18:30:51 +05:30
2,1,5,6 , &
3,2,6,7 , &
4,3,7,8 , &
4,1,5,8 , &
reworked build_ipNeighborhood added C3D8R reduced integration hexahedral element
2010-05-10 20:24:59 +05:30
8,7,6,5 , &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1,2,3,4 , & ! element 7 (3D 8node 8ip)
reworked build_ipNeighborhood added C3D8R reduced integration hexahedral element
2010-05-10 20:24:59 +05:30
2,1,5,6 , &
3,2,6,7 , &
4,3,7,8 , &
4,1,5,8 , &
1) added element type 57 (C3D20R) reduced quadratic hexahedral 2) need discussion about the present omission of the FE_nodesAtIP for type 8 --> added comment into source code
2010-08-17 04:23:24 +05:30
8,7,6,5 , &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1,2,3,4 , & ! element 21 (3D 20node 27ip)
1) added element type 57 (C3D20R) reduced quadratic hexahedral 2) need discussion about the present omission of the FE_nodesAtIP for type 8 --> added comment into source code
2010-08-17 04:23:24 +05:30
2,1,5,6 , &
3,2,6,7 , &
4,3,7,8 , &
4,1,5,8 , &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
8,7,6,5 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_maxNmatchingNodesPerFace,FE_maxNfaces,FE_Ngeomtypes])
integer(pInt), dimension(FE_Ngeomtypes), parameter, private :: FE_Ncellnodes = & !< number of cell nodes in a specific geometry type
int([ &
3, & ! element 6 (2D 3node 1ip)
7, & ! element 125 (2D 6node 3ip)
9, & ! element 11 (2D 4node 4ip)
fixed bug in cell generation: "FE_Ncellnodes" was not correct for elem type 17 and 136 too much memory allocated for mesh_cell
2013-04-18 23:26:14 +05:30
16, & ! element 27 (2D 8node 9ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
4, & ! element 134 (3D 4node 1ip)
15, & ! element 127 (3D 10node 4ip)
fixed bug in cell generation: "FE_Ncellnodes" was not correct for elem type 17 and 136 too much memory allocated for mesh_cell
2013-04-18 23:26:14 +05:30
21, & ! element 136 (3D 6node 6ip)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
8, & ! element 117 (3D 8node 1ip)
27, & ! element 7 (3D 8node 8ip)
64 & ! element 21 (3D 20node 27ip)
],pInt)
integer(pInt), dimension(FE_Ncelltypes), parameter, private :: FE_NcellnodesPerCell = & !< number of cell nodes in a specific cell type
int([ &
3, & ! (2D 3node)
4, & ! (2D 4node)
4, & ! (3D 4node)
8 & ! (3D 8node)
],pInt)
integer(pInt), dimension(FE_Ncelltypes), parameter, private :: FE_NcellnodesPerCellface = & !< number of cell nodes per cell face in a specific cell type
int([&
2, & ! (2D 3node)
2, & ! (2D 4node)
3, & ! (3D 4node)
4 & ! (3D 8node)
],pInt)
integer(pInt), dimension(FE_Ngeomtypes), parameter, public :: FE_Nips = & !< number of IPs in a specific type of element
int([ &
1, & ! element 6 (2D 3node 1ip)
3, & ! element 125 (2D 6node 3ip)
4, & ! element 11 (2D 4node 4ip)
9, & ! element 27 (2D 8node 9ip)
1, & ! element 134 (3D 4node 1ip)
4, & ! element 127 (3D 10node 4ip)
6, & ! element 136 (3D 6node 6ip)
1, & ! element 117 (3D 8node 1ip)
8, & ! element 7 (3D 8node 8ip)
27 & ! element 21 (3D 20node 27ip)
],pInt)
integer(pInt), dimension(FE_Ncelltypes), parameter, public :: FE_NipNeighbors = & !< number of ip neighbors / cell faces in a specific cell type
int([&
3, & ! (2D 3node)
4, & ! (2D 4node)
4, & ! (3D 4node)
6 & ! (3D 8node)
],pInt)
integer(pInt), dimension(FE_Ngeomtypes), parameter, private :: FE_maxNnodesAtIP = & !< maximum number of parent nodes that belong to an IP for a specific type of element
int([ &
3, & ! element 6 (2D 3node 1ip)
1, & ! element 125 (2D 6node 3ip)
1, & ! element 11 (2D 4node 4ip)
2, & ! element 27 (2D 8node 9ip)
4, & ! element 134 (3D 4node 1ip)
1, & ! element 127 (3D 10node 4ip)
1, & ! element 136 (3D 6node 6ip)
8, & ! element 117 (3D 8node 1ip)
1, & ! element 7 (3D 8node 8ip)
4 & ! element 21 (3D 20node 27ip)
],pInt)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
integer(pInt), dimension(FE_Nelemtypes), parameter, private :: MESH_VTKELEMTYPE = &
int([ &
5, & ! element 6 (2D 3node 1ip)
22, & ! element 125 (2D 6node 3ip)
9, & ! element 11 (2D 4node 4ip)
23, & ! element 27 (2D 8node 9ip)
23, & ! element 54 (2D 8node 4ip)
10, & ! element 134 (3D 4node 1ip)
10, & ! element 157 (3D 5node 4ip)
24, & ! element 127 (3D 10node 4ip)
13, & ! element 136 (3D 6node 6ip)
12, & ! element 117 (3D 8node 1ip)
12, & ! element 7 (3D 8node 8ip)
25, & ! element 57 (3D 20node 8ip)
25 & ! element 21 (3D 20node 27ip)
],pInt)
integer(pInt), dimension(FE_Ncelltypes), parameter, private :: MESH_VTKCELLTYPE = &
int([ &
5, & ! (2D 3node)
9, & ! (2D 4node)
10, & ! (3D 4node)
12 & ! (3D 8node)
],pInt)
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
public :: &
mesh_init, &
mesh_FEasCP, &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_build_cellnodes, &
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_build_ipVolumes, &
mesh_build_ipCoordinates, &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_cellCenterCoordinates, &
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
mesh_init_postprocessing, &
added two new scripts: "marc_extractData" is a simple variant of the "postResult" script that extracts all data from a macro *.t16 file (quadratic elements not yet supported) and writes it into an asciitable "addDataToGeometry" searches for corresponding *.txt and *.vtk file in a given directory and adds the data from the *.txt file as SCALARS to the *.vtk file first script needed an additional function in mesh that returns the corresponding node of a given ip in a specific element
2013-04-29 02:26:17 +05:30
mesh_get_Ncellnodes, &
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
mesh_get_unitlength, &
added two new scripts: "marc_extractData" is a simple variant of the "postResult" script that extracts all data from a macro *.t16 file (quadratic elements not yet supported) and writes it into an asciitable "addDataToGeometry" searches for corresponding *.txt and *.vtk file in a given directory and adds the data from the *.txt file as SCALARS to the *.vtk file first script needed an additional function in mesh that returns the corresponding node of a given ip in a specific element
2013-04-29 02:26:17 +05:30
mesh_get_nodeAtIP
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#ifdef Spectral
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
public :: &
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mesh_spectral_getGrid, &
mesh_spectral_getSize, &
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_regrid, &
mesh_nodesAroundCentres, &
mesh_deformedCoordsFFT, &
mesh_deformedCoordsLinear, &
mesh_volumeMismatch, &
mesh_shapeMismatch
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#endif
private :: &
#ifdef Spectral
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_spectral_getHomogenization, &
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mesh_spectral_count, &
mesh_spectral_mapNodesAndElems, &
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_spectral_count_cpSizes, &
mesh_spectral_build_nodes, &
mesh_spectral_build_elements, &
spectral_build_neighborhood now "hidden" in spectral compile part. (should not interfere with Marc/Abaqus compile anymore)
2013-04-22 14:31:58 +05:30
mesh_spectral_build_ipNeighborhood, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#endif
removed #define Marc and substituted #ifdef Marc by #ifdef Marc4DAMASK which is defined by either the submit scripts or DAMASK_marcXXXX.f90
2013-04-30 15:19:30 +05:30
#ifdef Marc4DAMASK
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_marc_get_tableStyles, &
mesh_marc_count_nodesAndElements, &
mesh_marc_count_elementSets, &
mesh_marc_map_elementSets, &
mesh_marc_count_cpElements, &
mesh_marc_map_Elements, &
mesh_marc_map_nodes, &
mesh_marc_build_nodes, &
mesh_marc_count_cpSizes, &
mesh_marc_build_elements, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#endif
#ifdef Abaqus
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_abaqus_count_nodesAndElements, &
mesh_abaqus_count_elementSets, &
mesh_abaqus_count_materials, &
mesh_abaqus_map_elementSets, &
mesh_abaqus_map_materials, &
mesh_abaqus_count_cpElements, &
mesh_abaqus_map_elements, &
mesh_abaqus_map_nodes, &
mesh_abaqus_build_nodes, &
mesh_abaqus_count_cpSizes, &
mesh_abaqus_build_elements, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#endif
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_get_damaskOptions, &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_build_cellconnectivity, &
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_build_ipAreas, &
mesh_build_nodeTwins, &
mesh_build_sharedElems, &
mesh_build_ipNeighborhood, &
mesh_tell_statistics, &
FE_mapElemtype, &
mesh_faceMatch, &
mesh_build_FEdata, &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_write_cellGeom, &
mesh_write_elemGeom, &
mesh_write_meshfile, &
mesh_read_meshfile
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
contains
added function mesh_FEtoCPelement(FEid)
2007-03-21 21:48:33 +05:30
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief initializes the mesh by calling all necessary private routines the mesh module
!! Order and routines strongly depend on type of solver
!--------------------------------------------------------------------------------------------------
introduced dedicated calculation of interface normals in IP neighborhood. (2D elements did not work so far..!) fixed definitions of some (2D) interface planes to have all normals pointing outward. removed "int, private :: i" defined for spectral compilation.
2013-02-05 18:57:37 +05:30
subroutine mesh_init(ip,el)
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
renamed interface from "mpie_interface" to "DAMASK_interface"
2011-05-11 22:31:03 +05:30
use DAMASK_interface
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use, intrinsic :: iso_fortran_env ! to get compiler_version and compiler_options (at least for gfortran 4.6 at the moment)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
use IO, only: &
reporting current time for most module inits
2013-02-25 22:04:59 +05:30
IO_timeStamp, &
further improvements on implementing the plain mode
2013-04-16 22:37:27 +05:30
IO_error, &
last version still did not compile: IO_write_jobFile was not used by mesh_init
2013-04-19 18:32:37 +05:30
IO_write_jobFile, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#ifdef Abaqus
IO_abaqus_hasNoPart, &
#endif
#ifdef Spectral
IO_open_file
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
use numerics, only: &
divergence_correction
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#else
IO_open_InputFile
#endif
further improvements on implementing the plain mode
2013-04-16 22:37:27 +05:30
use numerics, only: &
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
usePingPong, &
numerics_unitlength
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
use FEsolving, only: &
FEsolving_execElem, &
FEsolving_execIP, &
calcMode, &
lastMode, &
modelName
overhaul to allow for multiphase/homogenization setup.
2009-03-04 17:18:54 +05:30
added function calls in init()
2007-04-10 16:52:53 +05:30
implicit none
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), parameter :: fileUnit = 222_pInt
introduced dedicated calculation of interface normals in IP neighborhood. (2D elements did not work so far..!) fixed definitions of some (2D) interface planes to have all normals pointing outward. removed "int, private :: i" defined for spectral compilation.
2013-02-05 18:57:37 +05:30
integer(pInt), intent(in) :: el, ip
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
integer(pInt) :: j
numerics.config is now optional; if this file does not exist in the working directory, standard values from numerics.f90 are used; either way the output file tells which values were taken. added some output to the init subroutines of debug.f90, FEsolving.f90, lattice.f90, material.f90 and mesh.f90.
2009-06-18 19:58:02 +05:30
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
write(6,'(/,a)') ' <<<+- mesh init -+>>>'
write(6,'(a)') ' $Id$'
write(6,'(a15,a)') ' Current time: ',IO_timeStamp()
made code standard conform to Fortran 2008 (ignoring warning concerning comments beyond character 132). Basically, changing "x" format specifier to "Nx" ("1x") plus removing $ format specifier added compiler switches for gfortran and ifort to check for standard conformity old gnu compilers <4.4 are not longer supported because they don't provide the c binding for fftw
2012-02-01 00:48:55 +05:30
#include "compilation_info.f90"
removed unnecsessary omp statements, removed one transpose in constitutitve
2013-01-08 16:39:20 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
if (allocated(mesh_mapFEtoCPelem)) deallocate(mesh_mapFEtoCPelem)
if (allocated(mesh_mapFEtoCPnode)) deallocate(mesh_mapFEtoCPnode)
if (allocated(mesh_node0)) deallocate(mesh_node0)
if (allocated(mesh_node)) deallocate(mesh_node)
if (allocated(mesh_element)) deallocate(mesh_element)
if (allocated(mesh_cell)) deallocate(mesh_cell)
if (allocated(mesh_cellnode)) deallocate(mesh_cellnode)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
if (allocated(mesh_cellnodeParent)) deallocate(mesh_cellnodeParent)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
if (allocated(mesh_ipCoordinates)) deallocate(mesh_ipCoordinates)
if (allocated(mesh_ipArea)) deallocate(mesh_ipArea)
if (allocated(mesh_ipAreaNormal)) deallocate(mesh_ipAreaNormal)
if (allocated(mesh_sharedElem)) deallocate(mesh_sharedElem)
if (allocated(mesh_ipNeighborhood)) deallocate(mesh_ipNeighborhood)
if (allocated(mesh_ipVolume)) deallocate(mesh_ipVolume)
if (allocated(mesh_nodeTwins)) deallocate(mesh_nodeTwins)
if (allocated(FE_nodesAtIP)) deallocate(FE_nodesAtIP)
if (allocated(FE_ipNeighbor)) deallocate(FE_ipNeighbor)
if (allocated(FE_cellnodeParentnodeWeights)) deallocate(FE_cellnodeParentnodeWeights)
if (allocated(FE_subNodeOnIPFace)) deallocate(FE_subNodeOnIPFace)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
call mesh_build_FEdata ! get properties of the different types of elements
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
mesh_unitlength = numerics_unitlength ! set physical extent of a length unit in mesh
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#ifdef Spectral
call IO_open_file(fileUnit,geometryFile) ! parse info from geometry file...
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
call mesh_spectral_count(fileUnit)
call mesh_spectral_mapNodesAndElems
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
call mesh_spectral_count_cpSizes
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
call mesh_spectral_build_nodes(fileUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
call mesh_spectral_build_elements(fileUnit)
added specialized ipNeighborhood calculation for spectral solver case
2013-04-19 21:34:20 +05:30
call mesh_get_damaskOptions(fileUnit)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
call mesh_build_cellconnectivity
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
mesh_cellnode = mesh_build_cellnodes(mesh_node,mesh_Ncellnodes)
added specialized ipNeighborhood calculation for spectral solver case
2013-04-19 21:34:20 +05:30
call mesh_build_ipCoordinates
call mesh_build_ipVolumes
call mesh_build_ipAreas
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
call mesh_spectral_build_ipNeighborhood(fileUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#endif
removed #define Marc and substituted #ifdef Marc by #ifdef Marc4DAMASK which is defined by either the submit scripts or DAMASK_marcXXXX.f90
2013-04-30 15:19:30 +05:30
#ifdef Marc4DAMASK
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
call IO_open_inputFile(fileUnit,modelName) ! parse info from input file...
call mesh_marc_get_tableStyles(fileUnit)
call mesh_marc_count_nodesAndElements(fileUnit)
call mesh_marc_count_elementSets(fileUnit)
call mesh_marc_map_elementSets(fileUnit)
call mesh_marc_count_cpElements(fileUnit)
call mesh_marc_map_elements(fileUnit)
call mesh_marc_map_nodes(fileUnit)
call mesh_marc_build_nodes(fileUnit)
call mesh_marc_count_cpSizes(fileunit)
call mesh_marc_build_elements(fileUnit)
added specialized ipNeighborhood calculation for spectral solver case
2013-04-19 21:34:20 +05:30
call mesh_get_damaskOptions(fileUnit)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
call mesh_build_cellconnectivity
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
mesh_cellnode = mesh_build_cellnodes(mesh_node,mesh_Ncellnodes)
added specialized ipNeighborhood calculation for spectral solver case
2013-04-19 21:34:20 +05:30
call mesh_build_ipCoordinates
call mesh_build_ipVolumes
call mesh_build_ipAreas
call mesh_build_nodeTwins
call mesh_build_sharedElems
call mesh_build_ipNeighborhood
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#endif
#ifdef Abaqus
call IO_open_inputFile(fileUnit,modelName) ! parse info from input file...
noPart = IO_abaqus_hasNoPart(fileUnit)
call mesh_abaqus_count_nodesAndElements(fileUnit)
call mesh_abaqus_count_elementSets(fileUnit)
call mesh_abaqus_count_materials(fileUnit)
call mesh_abaqus_map_elementSets(fileUnit)
call mesh_abaqus_map_materials(fileUnit)
call mesh_abaqus_count_cpElements(fileUnit)
call mesh_abaqus_map_elements(fileUnit)
call mesh_abaqus_map_nodes(fileUnit)
call mesh_abaqus_build_nodes(fileUnit)
call mesh_abaqus_count_cpSizes(fileunit)
call mesh_abaqus_build_elements(fileUnit)
reordered (and partly redistributed) error message identifiers, deleted those which are not in use anymore. all constitutive as well numerics now raises an error, if an unknown keyword is found in the respective config file
2012-02-13 23:11:27 +05:30
call mesh_get_damaskOptions(fileUnit)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
call mesh_build_cellconnectivity
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
mesh_cellnode = mesh_build_cellnodes(mesh_node,mesh_Ncellnodes)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
call mesh_build_ipCoordinates
call mesh_build_ipVolumes
call mesh_build_ipAreas
call mesh_build_nodeTwins
call mesh_build_sharedElems
call mesh_build_ipNeighborhood
added specialized ipNeighborhood calculation for spectral solver case
2013-04-19 21:34:20 +05:30
#endif
reordered (and partly redistributed) error message identifiers, deleted those which are not in use anymore. all constitutive as well numerics now raises an error, if an unknown keyword is found in the respective config file
2012-02-13 23:11:27 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
close (fileUnit)
added specialized ipNeighborhood calculation for spectral solver case
2013-04-19 21:34:20 +05:30
call mesh_tell_statistics
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
call mesh_write_meshfile
call mesh_write_cellGeom
call mesh_write_elemGeom
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
if (usePingPong .and. (mesh_Nelems /= mesh_NcpElems)) call IO_error(600_pInt) ! ping-pong must be disabled when havin non-DAMASK-elements
added function calls in init()
2007-04-10 16:52:53 +05:30
introduced dedicated calculation of interface normals in IP neighborhood. (2D elements did not work so far..!) fixed definitions of some (2D) interface planes to have all normals pointing outward. removed "int, private :: i" defined for spectral compilation.
2013-02-05 18:57:37 +05:30
FEsolving_execElem = [ 1_pInt,mesh_NcpElems ]
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
if (allocated(FEsolving_execIP)) deallocate(FEsolving_execIP)
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
allocate(FEsolving_execIP(2_pInt,mesh_NcpElems)); FEsolving_execIP = 1_pInt
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
forall (j = 1_pInt:mesh_NcpElems) FEsolving_execIP(2,j) = FE_Nips(FE_geomtype(mesh_element(2,j)))
added a new module called numerics.f90 which reads in all numerical "parameters" from the file numerics.config (also being added). From now on this file has to be located in the working directory of the FEM-model and has to contain all necessary parameters.
2009-06-15 18:41:21 +05:30
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
if (allocated(calcMode)) deallocate(calcMode)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
allocate(calcMode(mesh_maxNips,mesh_NcpElems))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
calcMode = .false. ! pretend to have collected what first call is asking (F = I)
introduced dedicated calculation of interface normals in IP neighborhood. (2D elements did not work so far..!) fixed definitions of some (2D) interface planes to have all normals pointing outward. removed "int, private :: i" defined for spectral compilation.
2013-02-05 18:57:37 +05:30
calcMode(ip,mesh_FEasCP('elem',el)) = .true. ! first ip,el needs to be already pingponged to "calc"
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
lastMode = .true. ! and its mode is already known...
added function calls in init()
2007-04-10 16:52:53 +05:30
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
end subroutine mesh_init
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Gives the FE to CP ID mapping by binary search through lookup array
!! valid questions (what) are 'elem', 'node'
!--------------------------------------------------------------------------------------------------
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt) function mesh_FEasCP(what,myID)
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
use IO, only: &
IO_lc
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
added function calls in init()
2007-04-10 16:52:53 +05:30
implicit none
character(len=*), intent(in) :: what
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myID
added function calls in init()
2007-04-10 16:52:53 +05:30
integer(pInt), dimension(:,:), pointer :: lookupMap
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt) :: lower,upper,center
added function calls in init()
2007-04-10 16:52:53 +05:30
mesh_FEasCP = 0_pInt
select case(IO_lc(what(1:4)))
case('elem')
lookupMap => mesh_mapFEtoCPelem
case('node')
lookupMap => mesh_mapFEtoCPnode
case default
return
added a new module called numerics.f90 which reads in all numerical "parameters" from the file numerics.config (also being added). From now on this file has to be located in the working directory of the FEM-model and has to contain all necessary parameters.
2009-06-15 18:41:21 +05:30
endselect
added function calls in init()
2007-04-10 16:52:53 +05:30
lower = 1_pInt
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
upper = int(size(lookupMap,2_pInt),pInt)
added function calls in init()
2007-04-10 16:52:53 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (lookupMap(1_pInt,lower) == myID) then ! check at bounds QUESTION is it valid to extend bounds by 1 and just do binary search w/o init check at bounds?
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
mesh_FEasCP = lookupMap(2_pInt,lower)
added function calls in init()
2007-04-10 16:52:53 +05:30
return
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
elseif (lookupMap(1_pInt,upper) == myID) then
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
mesh_FEasCP = lookupMap(2_pInt,upper)
added function calls in init()
2007-04-10 16:52:53 +05:30
return
endif
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do while (upper-lower > 1_pInt) ! binary search in between bounds
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
center = (lower+upper)/2_pInt
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
if (lookupMap(1_pInt,center) < myID) then
added function calls in init()
2007-04-10 16:52:53 +05:30
lower = center
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
elseif (lookupMap(1_pInt,center) > myID) then
added function calls in init()
2007-04-10 16:52:53 +05:30
upper = center
else
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
mesh_FEasCP = lookupMap(2_pInt,center)
added function calls in init()
2007-04-10 16:52:53 +05:30
exit
added a new module called numerics.f90 which reads in all numerical "parameters" from the file numerics.config (also being added). From now on this file has to be located in the working directory of the FEM-model and has to contain all necessary parameters.
2009-06-15 18:41:21 +05:30
endif
enddo
added function calls in init()
2007-04-10 16:52:53 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
end function mesh_FEasCP
added function calls in init()
2007-04-10 16:52:53 +05:30
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!> @brief Split CP elements into cells.
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
!> @details Build a mapping between cells and the corresponding cell nodes ('mesh_cell').
!> Cell nodes that are also matching nodes are unique in the list of cell nodes,
!> all others (currently) might be stored more than once.
!> Also allocates the 'mesh_node' array.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
subroutine mesh_build_cellconnectivity
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
implicit none
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(:), allocatable :: &
matchingNode2cellnode
integer(pInt), dimension(:,:), allocatable :: &
cellnodeParent
integer(pInt), dimension(mesh_maxNcellnodes) :: &
localCellnode2globalCellnode
integer(pInt) &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
e,t,g,c,n,i, &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
matchingNodeID, &
localCellnodeID
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
!*** Count cell nodes (including duplicates) and generate cell connectivity list
allocate(mesh_cell(FE_maxNcellnodesPerCell,mesh_maxNips,mesh_NcpElems)) ; mesh_cell = 0_pInt
allocate(matchingNode2cellnode(mesh_Nnodes)) ; matchingNode2cellnode = 0_pInt
allocate(cellnodeParent(2_pInt,mesh_maxNcellnodes*mesh_NcpElems)) ; cellnodeParent = 0_pInt
mesh_Ncellnodes = 0_pInt
mesh_Ncells = 0_pInt
do e = 1_pInt,mesh_NcpElems ! loop over cpElems
t = mesh_element(2_pInt,e) ! get element type
g = FE_geomtype(t) ! get geometry type
c = FE_celltype(g) ! get cell type
localCellnode2globalCellnode = 0_pInt
mesh_Ncells = mesh_Ncells + FE_Nips(g)
do i = 1_pInt,FE_Nips(g) ! loop over ips=cells in this element
do n = 1_pInt,FE_NcellnodesPerCell(c) ! loop over cell nodes in this cell
localCellnodeID = FE_cell(n,i,g)
if (localCellnodeID <= FE_NmatchingNodes(g)) then ! this cell node is a matching node
matchingNodeID = mesh_element(4_pInt+localCellnodeID,e)
if (matchingNode2cellnode(matchingNodeID) == 0_pInt) then ! if this matching node does not yet exist in the glbal cell node list ...
mesh_Ncellnodes = mesh_Ncellnodes + 1_pInt ! ... count it as cell node ...
matchingNode2cellnode(matchingNodeID) = mesh_Ncellnodes ! ... and remember its global ID
cellnodeParent(1_pInt,mesh_Ncellnodes) = e ! ... and where it belongs to
cellnodeParent(2_pInt,mesh_Ncellnodes) = localCellnodeID
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
endif
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_cell(n,i,e) = matchingNode2cellnode(matchingNodeID)
else ! this cell node is no matching node
if (localCellnode2globalCellnode(localCellnodeID) == 0_pInt) then ! if this local cell node does not yet exist in the global cell node list ...
mesh_Ncellnodes = mesh_Ncellnodes + 1_pInt ! ... count it as cell node ...
localCellnode2globalCellnode(localCellnodeID) = mesh_Ncellnodes ! ... and remember its global ID ...
cellnodeParent(1_pInt,mesh_Ncellnodes) = e ! ... and it belongs to
cellnodeParent(2_pInt,mesh_Ncellnodes) = localCellnodeID
endif
mesh_cell(n,i,e) = localCellnode2globalCellnode(localCellnodeID)
endif
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo
enddo
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
enddo
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
allocate(mesh_cellnodeParent(2_pInt,mesh_Ncellnodes))
allocate(mesh_cellnode(3_pInt,mesh_Ncellnodes))
forall(n = 1_pInt:mesh_Ncellnodes)
mesh_cellnodeParent(1,n) = cellnodeParent(1,n)
mesh_cellnodeParent(2,n) = cellnodeParent(2,n)
endforall
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
deallocate(matchingNode2cellnode)
deallocate(cellnodeParent)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
end subroutine mesh_build_cellconnectivity
!--------------------------------------------------------------------------------------------------
!> @brief Calculate position of cellnodes from the given position of nodes
!> Build list of cellnodes' coordinates.
!> Cellnode coordinates are calculated from a weighted sum of node coordinates.
!--------------------------------------------------------------------------------------------------
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
function mesh_build_cellnodes(nodes,Ncellnodes)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
implicit none
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
integer(pInt), intent(in) :: Ncellnodes !< requested number of cellnodes
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
real(pReal), dimension(3,mesh_Nnodes), intent(in) :: nodes
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
real(pReal), dimension(3,Ncellnodes) :: mesh_build_cellnodes
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
integer(pInt) &
e,t,n,m, &
localCellnodeID
real(pReal), dimension(3) :: &
myCoords
mesh_build_cellnodes = 0.0_pReal
fixed bug in cell generation: "FE_Ncellnodes" was not correct for elem type 17 and 136 too much memory allocated for mesh_cell
2013-04-18 23:26:14 +05:30
!$OMP PARALLEL DO PRIVATE(e,localCellnodeID,t,myCoords)
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
do n = 1_pInt,Ncellnodes ! loop over cell nodes
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
e = mesh_cellnodeParent(1,n)
localCellnodeID = mesh_cellnodeParent(2,n)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
t = mesh_element(2,e) ! get element type
myCoords = 0.0_pReal
do m = 1_pInt,FE_Nnodes(t)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
myCoords = myCoords + nodes(1:3,mesh_element(4_pInt+m,e)) &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
* FE_cellnodeParentnodeWeights(m,localCellnodeID,t)
enddo
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_build_cellnodes(1:3,n) = myCoords / sum(FE_cellnodeParentnodeWeights(:,localCellnodeID,t))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
enddo
fixed bug in cell generation: "FE_Ncellnodes" was not correct for elem type 17 and 136 too much memory allocated for mesh_cell
2013-04-18 23:26:14 +05:30
!$OMP END PARALLEL DO
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
end function mesh_build_cellnodes
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Calculates IP volume. Allocates global array 'mesh_ipVolume'
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!> @details The IP volume is calculated differently depending on the cell type.
!> 2D cells assume an element depth of one in order to calculate the volume.
!> For the hexahedral cell we subdivide the cell into subvolumes of pyramidal
!> shape with a cell face as basis and the central ip at the tip. This subvolume is
!> calculated as an average of four tetrahedals with three corners on the cell face
!> and one corner at the central ip.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
subroutine mesh_build_ipVolumes
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
use math, only: &
math_volTetrahedron, &
math_areaTriangle
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
implicit none
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt) :: e,t,g,c,i,m,f,n
real(pReal), dimension(FE_maxNcellnodesPerCellface,FE_maxNcellfaces) :: subvolume
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (.not. allocated(mesh_ipVolume)) then
allocate(mesh_ipVolume(mesh_maxNips,mesh_NcpElems))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_ipVolume = 0.0_pReal
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!$OMP PARALLEL DO PRIVATE(t,g,c,m,subvolume)
do e = 1_pInt,mesh_NcpElems ! loop over cpElems
t = mesh_element(2_pInt,e) ! get element type
g = FE_geomtype(t) ! get geometry type
c = FE_celltype(g) ! get cell type
select case (c)
case (1_pInt) ! 2D 3node
forall (i = 1_pInt:FE_Nips(g)) & ! loop over ips=cells in this element
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_ipVolume(i,e) = math_areaTriangle(mesh_cellnode(1:3,mesh_cell(1,i,e)), &
mesh_cellnode(1:3,mesh_cell(2,i,e)), &
mesh_cellnode(1:3,mesh_cell(3,i,e)))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
case (2_pInt) ! 2D 4node
forall (i = 1_pInt:FE_Nips(g)) & ! loop over ips=cells in this element
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_ipVolume(i,e) = math_areaTriangle(mesh_cellnode(1:3,mesh_cell(1,i,e)), & ! here we assume a planar shape, so division in two triangles suffices
mesh_cellnode(1:3,mesh_cell(2,i,e)), &
mesh_cellnode(1:3,mesh_cell(3,i,e))) &
+ math_areaTriangle(mesh_cellnode(1:3,mesh_cell(3,i,e)), &
mesh_cellnode(1:3,mesh_cell(4,i,e)), &
mesh_cellnode(1:3,mesh_cell(1,i,e)))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
case (3_pInt) ! 3D 4node
forall (i = 1_pInt:FE_Nips(g)) & ! loop over ips=cells in this element
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_ipVolume(i,e) = math_volTetrahedron(mesh_cellnode(1:3,mesh_cell(1,i,e)), &
mesh_cellnode(1:3,mesh_cell(2,i,e)), &
mesh_cellnode(1:3,mesh_cell(3,i,e)), &
mesh_cellnode(1:3,mesh_cell(4,i,e)))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
case (4_pInt) ! 3D 8node
m = FE_NcellnodesPerCellface(c)
do i = 1_pInt,FE_Nips(g) ! loop over ips=cells in this element
subvolume = 0.0_pReal
forall(f = 1_pInt:FE_NipNeighbors(c), n = 1_pInt:FE_NcellnodesPerCellface(c)) &
subvolume(n,f) = math_volTetrahedron(&
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_cellnode(1:3,mesh_cell(FE_cellface( n ,f,c),i,e)), &
mesh_cellnode(1:3,mesh_cell(FE_cellface(1+mod(n ,m),f,c),i,e)), &
mesh_cellnode(1:3,mesh_cell(FE_cellface(1+mod(n+1,m),f,c),i,e)), &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_ipCoordinates(1:3,i,e))
mesh_ipVolume(i,e) = 0.5_pReal * sum(subvolume) ! each subvolume is based on four tetrahedrons, altough the face consists of only two triangles -> averaging factor two
2d elements now have a volume corresponding to a thickness of 1; used to have zero volume now also 2d elements can be used with nonlocal model
2013-04-09 23:37:30 +05:30
enddo
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
end select
enddo
!$OMP END PARALLEL DO
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_build_ipVolumes
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
!> @brief Calculates IP Coordinates. Allocates global array 'mesh_ipCoordinates'
! Called by all solvers in mesh_init in order to initialize the ip coordinates.
! Later on the current ip coordinates are directly prvided by the spectral solver and by Abaqus,
! so no need to use this subroutine anymore; Marc however only provides nodal displacements,
! so in this case the ip coordinates are always calculated on the basis of this subroutine.
! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
! FOR THE MOMENT THIS SUBROUTINE ACTUALLY CALCULATES THE CELL CENTER AND NOT THE IP COORDINATES,
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
! AS THE IP IS NOT (ALWAYS) LOCATED IN THE CENTER OF THE IP VOLUME.
! HAS TO BE CHANGED IN A LATER VERSION.
! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
subroutine mesh_build_ipCoordinates
implicit none
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt) :: e,t,g,c,i,n
real(pReal), dimension(3) :: myCoords
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
if (.not. allocated(mesh_ipCoordinates)) then
allocate(mesh_ipCoordinates(3,mesh_maxNips,mesh_NcpElems))
mesh_ipCoordinates = 0.0_pReal
endif
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!$OMP PARALLEL DO PRIVATE(t,g,c,myCoords)
do e = 1_pInt,mesh_NcpElems ! loop over cpElems
t = mesh_element(2_pInt,e) ! get element type
g = FE_geomtype(t) ! get geometry type
c = FE_celltype(g) ! get cell type
do i = 1_pInt,FE_Nips(g) ! loop over ips=cells in this element
myCoords = 0.0_pReal
fixed bug in cell generation: "FE_Ncellnodes" was not correct for elem type 17 and 136 too much memory allocated for mesh_cell
2013-04-18 23:26:14 +05:30
do n = 1_pInt,FE_NcellnodesPerCell(c) ! loop over cell nodes in this cell
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
myCoords = myCoords + mesh_cellnode(1:3,mesh_cell(n,i,e))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_ipCoordinates(1:3,i,e) = myCoords / FE_NcellnodesPerCell(c)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo
enddo
parallelized mesh_build_subNodeCoords and mesh_build_ipCoordinates
2012-11-09 13:17:14 +05:30
!$OMP END PARALLEL DO
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_build_ipCoordinates
added _build_ipNeighborhood and _faceMatch
2007-03-26 14:20:57 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Calculates cell center coordinates.
!--------------------------------------------------------------------------------------------------
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
pure function mesh_cellCenterCoordinates(ip,el)
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
implicit none
integer(pInt), intent(in) :: el, & !< element number
ip !< integration point number
real(pReal), dimension(3) :: mesh_cellCenterCoordinates !< x,y,z coordinates of the cell center of the requested IP cell
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
integer(pInt) :: t,g,c,n
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
t = mesh_element(2_pInt,el) ! get element type
g = FE_geomtype(t) ! get geometry type
c = FE_celltype(g) ! get cell type
mesh_cellCenterCoordinates = 0.0_pReal
do n = 1_pInt,FE_NcellnodesPerCell(c) ! loop over cell nodes in this cell
mesh_cellCenterCoordinates = mesh_cellCenterCoordinates + mesh_cellnode(1:3,mesh_cell(n,ip,el))
enddo
mesh_cellCenterCoordinates = mesh_cellCenterCoordinates / FE_NcellnodesPerCell(c)
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
end function mesh_cellCenterCoordinates
introduced distinction between ip coordinates and cell center coordinates; the former can be globally calculated by subroutine mesh_build_ipCoordinates, the latter locally by the function mesh_cellCenterCoordinates; renamed mesh_ipCenterOfGravity to mesh_ipCordinates
2012-11-06 20:07:13 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#ifdef Spectral
!--------------------------------------------------------------------------------------------------
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
!> @brief Reads grid information from geometry file. If fileUnit is given,
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!! assumes an opened file, otherwise tries to open the one specified in geometryFile
!--------------------------------------------------------------------------------------------------
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
function mesh_spectral_getGrid(fileUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
use IO, only: &
IO_checkAndRewind, &
IO_open_file, &
IO_stringPos, &
IO_lc, &
IO_stringValue, &
IO_intValue, &
IO_floatValue, &
IO_error
use DAMASK_interface, only: &
geometryFile
implicit none
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
integer(pInt), dimension(3) :: mesh_spectral_getGrid
integer(pInt), intent(in), optional :: fileUnit
integer(pInt), dimension(1_pInt + 7_pInt*2_pInt) :: positions ! for a,b,c + 3 values + keyword
integer(pInt) :: headerLength = 0_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
character(len=1024) :: line, &
keyword
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
integer(pInt) :: i, j, myUnit
logical :: gotGrid = .false.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
mesh_spectral_getGrid = -1_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if(.not. present(fileUnit)) then
myUnit = 289_pInt
call IO_open_file(myUnit,trim(geometryFile))
else
myUnit = fileUnit
endif
call IO_checkAndRewind(myUnit)
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read(myUnit,'(a1024)') line
fixed bug when reading in geometry for spectral solver
2012-11-06 22:46:01 +05:30
positions = IO_stringPos(line,7_pInt)
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
keyword = IO_lc(IO_StringValue(line,positions,2_pInt,.true.))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (keyword(1:4) == 'head') then
headerLength = IO_intValue(line,positions,1_pInt) + 1_pInt
else
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
call IO_error(error_ID=841_pInt, ext_msg='mesh_spectral_getGrid')
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
rewind(myUnit)
do i = 1_pInt, headerLength
read(myUnit,'(a1024)') line
positions = IO_stringPos(line,7_pInt)
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
select case ( IO_lc(IO_StringValue(line,positions,1_pInt,.true.)) )
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
case ('resolution','grid')
gotGrid = .true.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do j = 2_pInt,6_pInt,2_pInt
select case (IO_lc(IO_stringValue(line,positions,j)))
case('a')
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
mesh_spectral_getGrid(1) = IO_intValue(line,positions,j+1_pInt)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case('b')
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
mesh_spectral_getGrid(2) = IO_intValue(line,positions,j+1_pInt)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case('c')
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
mesh_spectral_getGrid(3) = IO_intValue(line,positions,j+1_pInt)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end select
enddo
end select
enddo
if(.not. present(fileUnit)) close(myUnit)
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
if (.not. gotGrid) &
call IO_error(error_ID = 845_pInt, ext_msg='grid')
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
if(any(mesh_spectral_getGrid < 1_pInt)) &
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
call IO_error(error_ID = 843_pInt, ext_msg='mesh_spectral_getGrid')
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
end function mesh_spectral_getGrid
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
!> @brief Reads size information from geometry file. If fileUnit is given,
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!! assumes an opened file, otherwise tries to open the one specified in geometryFile
!--------------------------------------------------------------------------------------------------
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
function mesh_spectral_getSize(fileUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
use IO, only: &
IO_checkAndRewind, &
IO_open_file, &
IO_stringPos, &
IO_lc, &
IO_stringValue, &
IO_intValue, &
IO_floatValue, &
IO_error
use DAMASK_interface, only: &
geometryFile
implicit none
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
real(pReal), dimension(3) :: mesh_spectral_getSize
integer(pInt), intent(in), optional :: fileUnit
integer(pInt), dimension(1_pInt + 7_pInt*2_pInt) :: positions ! for x,y,z + 3 values + keyword
integer(pInt) :: headerLength = 0_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
character(len=1024) :: line, &
keyword
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
integer(pInt) :: i, j, myUnit
logical :: gotSize = .false.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
mesh_spectral_getSize = -1.0_pReal
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if(.not. present(fileUnit)) then
myUnit = 289_pInt
call IO_open_file(myUnit,trim(geometryFile))
else
myUnit = fileUnit
endif
call IO_checkAndRewind(myUnit)
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read(myUnit,'(a1024)') line
fixed bug when reading in geometry for spectral solver
2012-11-06 22:46:01 +05:30
positions = IO_stringPos(line,7_pInt)
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
keyword = IO_lc(IO_StringValue(line,positions,2_pInt,.true.))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (keyword(1:4) == 'head') then
headerLength = IO_intValue(line,positions,1_pInt) + 1_pInt
else
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
call IO_error(error_ID=841_pInt, ext_msg='mesh_spectral_getSize')
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
rewind(myUnit)
do i = 1_pInt, headerLength
read(myUnit,'(a1024)') line
positions = IO_stringPos(line,7_pInt)
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
select case ( IO_lc(IO_StringValue(line,positions,1,.true.)) )
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
case ('dimension', 'size')
gotSize = .true.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do j = 2_pInt,6_pInt,2_pInt
select case (IO_lc(IO_stringValue(line,positions,j)))
case('x')
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
mesh_spectral_getSize(1) = IO_floatValue(line,positions,j+1_pInt)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case('y')
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
mesh_spectral_getSize(2) = IO_floatValue(line,positions,j+1_pInt)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case('z')
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
mesh_spectral_getSize(3) = IO_floatValue(line,positions,j+1_pInt)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end select
enddo
end select
enddo
if(.not. present(fileUnit)) close(myUnit)
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
if (.not. gotSize) &
call IO_error(error_ID = 845_pInt, ext_msg='size')
if (any(mesh_spectral_getSize<=0.0_pReal)) &
call IO_error(error_ID = 844_pInt, ext_msg='mesh_spectral_getSize')
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
end function mesh_spectral_getSize
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Reads homogenization information from geometry file. If fileUnit is given,
!! assumes an opened file, otherwise tries to open the one specified in geometryFile
!--------------------------------------------------------------------------------------------------
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
integer(pInt) function mesh_spectral_getHomogenization(fileUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
use IO, only: &
IO_checkAndRewind, &
IO_open_file, &
IO_stringPos, &
IO_lc, &
IO_stringValue, &
IO_intValue, &
IO_error
use DAMASK_interface, only: &
geometryFile
implicit none
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
integer(pInt), intent(in), optional :: fileUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), dimension(1_pInt + 7_pInt*2_pInt) :: positions ! for a, b, c + 3 values + keyword
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
integer(pInt) :: headerLength = 0_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
character(len=1024) :: line, &
keyword
enabled alternative (and soon standard) keywords grid (resolution) and size (dimension). removed errors for odd resolution
2013-04-08 19:52:32 +05:30
integer(pInt) :: i, myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
logical :: gotHomogenization = .false.
now with full functionality for element type 6 (6-node pentaeder) and with element data structure partly initialized as function variables and not parameters
2009-04-06 18:55:19 +05:30
removed unused variables, added pure statements, and declared external functions as external (using gfortrans debug option)
2013-02-11 16:13:45 +05:30
mesh_spectral_getHomogenization = -1_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if(.not. present(fileUnit)) then
myUnit = 289_pInt
call IO_open_file(myUnit,trim(geometryFile))
else
myUnit = fileUnit
endif
call IO_checkAndRewind(myUnit)
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read(myUnit,'(a1024)') line
fixed bug when reading in geometry for spectral solver
2012-11-06 22:46:01 +05:30
positions = IO_stringPos(line,7_pInt)
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
keyword = IO_lc(IO_StringValue(line,positions,2_pInt,.true.))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (keyword(1:4) == 'head') then
headerLength = IO_intValue(line,positions,1_pInt) + 1_pInt
else
call IO_error(error_ID=841_pInt, ext_msg='mesh_spectral_getHomogenization')
endif
rewind(myUnit)
do i = 1_pInt, headerLength
read(myUnit,'(a1024)') line
positions = IO_stringPos(line,7_pInt)
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
select case ( IO_lc(IO_StringValue(line,positions,1,.true.)) )
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case ('homogenization')
gotHomogenization = .true.
mesh_spectral_getHomogenization = IO_intValue(line,positions,2_pInt)
end select
enddo
if(.not. present(fileUnit)) close(myUnit)
if (.not. gotHomogenization ) &
call IO_error(error_ID = 845_pInt, ext_msg='homogenization')
if (mesh_spectral_getHomogenization<1_pInt) &
call IO_error(error_ID = 842_pInt, ext_msg='mesh_spectral_getHomogenization')
end function mesh_spectral_getHomogenization
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Count overall number of nodes and elements in mesh and stores them in
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
!! 'mesh_Nelems', 'mesh_Nnodes' and 'mesh_NcpElems'
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
!--------------------------------------------------------------------------------------------------
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
subroutine mesh_spectral_count(myUnit)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
implicit none
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
integer(pInt), intent(in) :: myUnit
integer(pInt), dimension(3) :: grid
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
grid = mesh_spectral_getGrid(myUnit)
mesh_Nelems = product(grid)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
mesh_NcpElems = mesh_Nelems
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mesh_Nnodes = product(grid+1_pInt)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
end subroutine mesh_spectral_count
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
!--------------------------------------------------------------------------------------------------
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
!> @brief fake map node from FE ID to internal (consecutive) representation for node and element
!! Allocates global array 'mesh_mapFEtoCPnode' and 'mesh_mapFEtoCPelem'
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
!--------------------------------------------------------------------------------------------------
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
subroutine mesh_spectral_mapNodesAndElems
use math, only: &
math_range
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
implicit none
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
allocate (mesh_mapFEtoCPnode(2_pInt,mesh_Nnodes), source = 0_pInt)
allocate (mesh_mapFEtoCPelem(2_pInt,mesh_NcpElems), source = 0_pInt)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mesh_mapFEtoCPnode = spread(math_range(mesh_Nnodes),1,2)
mesh_mapFEtoCPelem = spread(math_range(mesh_NcpElems),1,2)
end subroutine mesh_spectral_mapNodesAndElems
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Gets maximum count of nodes, IPs, IP neighbors, and subNodes among cpElements.
!! Allocates global arrays 'mesh_maxNnodes', 'mesh_maxNips', mesh_maxNipNeighbors',
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!! and mesh_maxNcellnodes
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
!--------------------------------------------------------------------------------------------------
subroutine mesh_spectral_count_cpSizes
implicit none
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt) :: t,g,c
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
t = FE_mapElemtype('C3D8R') ! fake 3D hexahedral 8 node 1 IP element
g = FE_geomtype(t)
c = FE_celltype(g)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
mesh_maxNnodes = FE_Nnodes(t)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_maxNips = FE_Nips(g)
mesh_maxNipNeighbors = FE_NipNeighbors(c)
mesh_maxNcellnodes = FE_Ncellnodes(g)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
end subroutine mesh_spectral_count_cpSizes
!--------------------------------------------------------------------------------------------------
!> @brief Store x,y,z coordinates of all nodes in mesh.
!! Allocates global arrays 'mesh_node0' and 'mesh_node'
!--------------------------------------------------------------------------------------------------
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
subroutine mesh_spectral_build_nodes(myUnit)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
implicit none
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
integer(pInt), intent(in) :: myUnit
integer(pInt) :: n
integer(pInt), dimension(3) :: grid
real(pReal), dimension(3) :: geomSize
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
allocate (mesh_node0 (3,mesh_Nnodes), source = 0.0_pReal)
allocate (mesh_node (3,mesh_Nnodes), source = 0.0_pReal)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
grid = mesh_spectral_getGrid(myUnit)
geomSize = mesh_spectral_getSize(myUnit)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
forall (n = 0_pInt:mesh_Nnodes-1_pInt)
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
mesh_node0(1,n+1_pInt) = mesh_unitlength * &
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
geomSize(1)*real(mod(n,(grid(1)+1_pInt) ),pReal) &
/ real(grid(1),pReal)
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
mesh_node0(2,n+1_pInt) = mesh_unitlength * &
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
geomSize(2)*real(mod(n/(grid(1)+1_pInt),(grid(2)+1_pInt)),pReal) &
/ real(grid(2),pReal)
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
mesh_node0(3,n+1_pInt) = mesh_unitlength * &
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
geomSize(3)*real(mod(n/(grid(1)+1_pInt)/(grid(2)+1_pInt),(grid(3)+1_pInt)),pReal) &
/ real(grid(3),pReal)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
end forall
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_node = mesh_node0
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
end subroutine mesh_spectral_build_nodes
!--------------------------------------------------------------------------------------------------
!> @brief Store FEid, type, material, texture, and node list per element.
!! Allocates global array 'mesh_element'
!--------------------------------------------------------------------------------------------------
subroutine mesh_spectral_build_elements(myUnit)
use IO, only: &
IO_checkAndRewind, &
IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_error, &
IO_continuousIntValues, &
IO_intValue, &
IO_countContinuousIntValues
implicit none
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
integer(pInt), intent(in) :: &
myUnit
integer(pInt), dimension(1_pInt+7_pInt*2_pInt) :: &
myPos
integer(pInt) :: &
e, i, &
headerLength = 0_pInt, &
maxIntCount, &
homog
integer(pInt), dimension(:), allocatable :: &
microstructures
integer(pInt), dimension(3) :: &
grid
integer(pInt), dimension(1,1) :: &
dummySet = 0_pInt
character(len=65536) :: &
line, &
keyword
character(len=64), dimension(1) :: &
dummyName = ''
grid = mesh_spectral_getGrid(myUnit)
homog = mesh_spectral_getHomogenization(myUnit)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
call IO_checkAndRewind(myUnit)
read(myUnit,'(a65536)') line
fixed bug when reading in geometry for spectral solver
2012-11-06 22:46:01 +05:30
myPos = IO_stringPos(line,7_pInt)
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
keyword = IO_lc(IO_StringValue(line,myPos,2_pInt,.true.))
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
if (keyword(1:4) == 'head') then
headerLength = IO_intValue(line,myPos,1_pInt) + 1_pInt
else
call IO_error(error_ID=841_pInt, ext_msg='mesh_spectral_build_elements')
endif
rewind(myUnit)
do i = 1_pInt, headerLength
read(myUnit,'(a65536)') line
enddo
maxIntCount = 0_pInt
i = 1_pInt
do while (i > 0_pInt)
i = IO_countContinuousIntValues(myUnit)
maxIntCount = max(maxIntCount, i)
enddo
rewind (myUnit)
do i=1_pInt,headerLength ! skip header
read(myUnit,'(a65536)') line
enddo
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
allocate (mesh_element (4_pInt+mesh_maxNnodes,mesh_NcpElems)); mesh_element = 0_pInt
allocate (microstructures (1_pInt+maxIntCount)); microstructures = 2_pInt
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
e = 0_pInt
do while (e < mesh_NcpElems .and. microstructures(1) > 0_pInt) ! fill expected number of elements, stop at end of data (or blank line!)
microstructures = IO_continuousIntValues(myUnit,maxIntCount,dummyName,dummySet,0_pInt) ! get affected elements
do i = 1_pInt,microstructures(1_pInt)
e = e+1_pInt ! valid element entry
mesh_element( 1,e) = e ! FE id
mesh_element( 2,e) = FE_mapElemtype('C3D8R') ! elem type
mesh_element( 3,e) = homog ! homogenization
mesh_element( 4,e) = microstructures(1_pInt+i) ! microstructure
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mesh_element( 5,e) = e + (e-1_pInt)/grid(1) + &
((e-1_pInt)/(grid(1)*grid(2)))*(grid(1)+1_pInt) ! base node
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
mesh_element( 6,e) = mesh_element(5,e) + 1_pInt
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mesh_element( 7,e) = mesh_element(5,e) + grid(1) + 2_pInt
mesh_element( 8,e) = mesh_element(5,e) + grid(1) + 1_pInt
mesh_element( 9,e) = mesh_element(5,e) +(grid(1) + 1_pInt) * (grid(2) + 1_pInt) ! second floor base node
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
mesh_element(10,e) = mesh_element(9,e) + 1_pInt
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mesh_element(11,e) = mesh_element(9,e) + grid(1) + 2_pInt
mesh_element(12,e) = mesh_element(9,e) + grid(1) + 1_pInt
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
mesh_maxValStateVar(1) = max(mesh_maxValStateVar(1),mesh_element(3,e)) !needed for statistics
mesh_maxValStateVar(2) = max(mesh_maxValStateVar(2),mesh_element(4,e))
enddo
enddo
deallocate(microstructures)
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
if (e /= mesh_NcpElems) call IO_error(880_pInt,e) !@ToDo does that make sense?
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
end subroutine mesh_spectral_build_elements
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
spectral_build_neighborhood now "hidden" in spectral compile part. (should not interfere with Marc/Abaqus compile anymore)
2013-04-22 14:31:58 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief build neighborhood relations for spectral
!> @details assign globals: mesh_ipNeighborhood
!--------------------------------------------------------------------------------------------------
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
subroutine mesh_spectral_build_ipNeighborhood(myUnit)
implicit none
integer(pInt), intent(in) :: &
myUnit
integer(pInt) :: &
x,y,z, &
e
integer(pInt), dimension(3) :: &
grid
allocate(mesh_ipNeighborhood(3,mesh_maxNipNeighbors,mesh_maxNips,mesh_NcpElems),source=0_pInt)
grid = mesh_spectral_getGrid(myUnit)
spectral_build_neighborhood now "hidden" in spectral compile part. (should not interfere with Marc/Abaqus compile anymore)
2013-04-22 14:31:58 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
e = 0_pInt
fixed bug in mesh_spectral_build_ipneighborhood that was introduced in rev2390: loop counter were mixed up
2013-05-14 15:15:22 +05:30
do z = 0_pInt,grid(3)-1_pInt
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
do y = 0_pInt,grid(2)-1_pInt
fixed bug in mesh_spectral_build_ipneighborhood that was introduced in rev2390: loop counter were mixed up
2013-05-14 15:15:22 +05:30
do x = 0_pInt,grid(1)-1_pInt
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
e = e + 1_pInt
mesh_ipNeighborhood(1,1,1,e) = z * grid(1) * grid(2) &
+ y * grid(1) &
+ modulo(x+1_pInt,grid(1)) &
+ 1_pInt
mesh_ipNeighborhood(1,2,1,e) = z * grid(1) * grid(2) &
+ y * grid(1) &
+ modulo(x-1_pInt,grid(1)) &
+ 1_pInt
mesh_ipNeighborhood(1,3,1,e) = z * grid(1) * grid(2) &
+ modulo(y+1_pInt,grid(2)) * grid(1) &
+ x &
+ 1_pInt
mesh_ipNeighborhood(1,4,1,e) = z * grid(1) * grid(2) &
+ modulo(y-1_pInt,grid(2)) * grid(1) &
+ x &
+ 1_pInt
mesh_ipNeighborhood(1,5,1,e) = modulo(z+1_pInt,grid(3)) * grid(1) * grid(2) &
+ y * grid(1) &
+ x &
+ 1_pInt
mesh_ipNeighborhood(1,6,1,e) = modulo(z-1_pInt,grid(3)) * grid(1) * grid(2) &
+ y * grid(1) &
+ x &
+ 1_pInt
mesh_ipNeighborhood(2,1:6,1,e) = 1_pInt
mesh_ipNeighborhood(3,1,1,e) = 2_pInt
mesh_ipNeighborhood(3,2,1,e) = 1_pInt
mesh_ipNeighborhood(3,3,1,e) = 4_pInt
mesh_ipNeighborhood(3,4,1,e) = 3_pInt
mesh_ipNeighborhood(3,5,1,e) = 6_pInt
mesh_ipNeighborhood(3,6,1,e) = 5_pInt
enddo
enddo
enddo
spectral_build_neighborhood now "hidden" in spectral compile part. (should not interfere with Marc/Abaqus compile anymore)
2013-04-22 14:31:58 +05:30
end subroutine mesh_spectral_build_ipNeighborhood
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Performes a regridding from saved restart information
!--------------------------------------------------------------------------------------------------
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
function mesh_regrid(adaptive,resNewInput,minRes)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
use prec, only: &
pInt, &
pReal
use DAMASK_interface, only: &
corrected some bugs concerning the regridding prevented FEsolving from potentially write to a none existing file started to introduce petsc into the make chain (nothing happens if PETSC_DIR is not set)
2012-06-20 18:19:46 +05:30
getSolverWorkingDirectoryName, &
getSolverJobName, &
GeometryFile
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
use IO, only: &
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
IO_open_file, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_read_jobBinaryFile ,&
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
IO_read_jobBinaryIntFile ,&
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
IO_write_jobBinaryFile, &
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
IO_write_jobBinaryIntFile, &
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
IO_write_jobFile, &
IO_error
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
use numerics, only: &
mySpectralSolver
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
use math, only: &
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
math_periodicNearestNeighbor, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
math_mul33x3
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
implicit none
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
logical, intent(in) :: adaptive ! if true, choose adaptive grid based on resNewInput, otherwise keep it constant
integer(pInt), dimension(3), optional, intent(in) :: resNewInput ! f2py cannot handle optional arguments correctly (they are always present)
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
integer(pInt), dimension(3), optional, intent(in) :: minRes
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
integer(pInt), dimension(3) :: mesh_regrid, ratio, grid
integer(pInt), parameter :: myUnit = 777_pInt
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
integer(pInt), dimension(3,2) :: possibleResNew
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
integer(pInt):: maxsize, i, j, k, ielem, NpointsNew, spatialDim, Nelems
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
integer(pInt), dimension(3) :: resNew
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), dimension(:), allocatable :: indices
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
real(pReal) :: wgt
real(pReal), dimension(3) :: geomSizeNew, geomSize
adaptive regrid now working, updates F and F_lastInc based on mapping.
2012-08-10 21:18:27 +05:30
real(pReal), dimension(3,3) :: Favg, Favg_LastInc, &
FavgNew, Favg_LastIncNew, &
deltaF, deltaF_lastInc
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
real(pReal), dimension(:,:), allocatable :: &
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
coordinates, coordinatesNew
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
real(pReal), dimension(:,:,:), allocatable :: &
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
stateHomog
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
real(pReal), dimension (:,:,:,:), allocatable :: &
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
spectralF9, spectralF9New, &
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
Tstar, TstarNew, &
stateConst
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
real(pReal), dimension(:,:,:,:,:), allocatable :: &
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
spectralF33, spectralF33New, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
F, FNew, &
Fp, FpNew, &
Lp, LpNew, &
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
dcsdE, dcsdENew, &
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
F_lastIncNew
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
real(pReal), dimension (:,:,:,:,:,:,:), allocatable :: &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
dPdF, dPdFNew
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
character(len=1024):: formatString, N_Digits
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
integer(pInt), dimension(:,:), allocatable :: &
sizeStateHomog
integer(pInt), dimension(:,:,:), allocatable :: &
material_phase, material_phaseNew, &
sizeStateConst
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
call IO_open_file(myUnit,trim(geometryFile))
grid = mesh_spectral_getGrid(myUnit)
geomSize = mesh_spectral_getsize(myUnit)
close(myUnit)
Nelems = product(grid)
wgt = 1.0_pReal/real(Nelems,pReal)
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
write(6,'(a)') 'Regridding geometry'
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
if (adaptive) then
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
write(6,'(a)') 'adaptive resolution determination'
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
if (present(minRes)) then
if (all(minRes /= -1_pInt)) & !the f2py way to tell it is present
write(6,'(a,3(i12))') ' given minimum resolution ', minRes
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
endif
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
if (present(resNewInput)) then
if (any (resNewInput<1)) call IO_error(890_pInt, ext_msg = 'resNewInput') !the f2py way to tell it is not present
write(6,'(a,3(i12))') ' target resolution ', resNewInput
else
call IO_error(890_pInt, ext_msg = 'resNewInput')
endif
endif
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
allocate(coordinates(3,Nelems))
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! read in deformation gradient to calculate coordinates, shape depend of selected solver
select case(myspectralsolver)
case('basic')
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
allocate(spectralF33(3,3,grid(1),grid(2),grid(3)))
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
call IO_read_jobBinaryFile(777,'F',trim(getSolverJobName()),size(spectralF33))
read (777,rec=1) spectralF33
close (777)
Favg = sum(sum(sum(spectralF33,dim=5),dim=4),dim=3) * wgt
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
coordinates = reshape(mesh_deformedCoordsFFT(geomSize,spectralF33),[3,mesh_NcpElems])
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
case('basicpetsc','al')
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
allocate(spectralF9(9,grid(1),grid(2),grid(3)))
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
call IO_read_jobBinaryFile(777,'F',trim(getSolverJobName()),size(spectralF9))
read (777,rec=1) spectralF9
close (777)
Favg = reshape(sum(sum(sum(spectralF9,dim=4),dim=3),dim=2) * wgt, [3,3])
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
coordinates = reshape(mesh_deformedCoordsFFT(geomSize,reshape(spectralF9, &
[3,3,grid(1),grid(2),grid(3)])),[3,mesh_NcpElems])
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
end select
!--------------------------------------------------------------------------------------------------
! sanity check 2D/3D case
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
if (grid(3)== 1_pInt) then
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
spatialDim = 2_pInt
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
if (present (minRes)) then
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
if (minRes(1) > 0_pInt .or. minRes(2) > 0_pInt) then
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
if (minRes(3) /= 1_pInt .or. &
mod(minRes(1),2_pInt) /= 0_pInt .or. &
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mod(minRes(2),2_pInt) /= 0_pInt) call IO_error(890_pInt, ext_msg = '2D minRes') ! as f2py has problems with present, use pyf file for initialization to -1
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
endif; endif
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
else
spatialDim = 3_pInt
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
if (present (minRes)) then
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
if (any(minRes > 0_pInt)) then
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
if (mod(minRes(1),2_pInt) /= 0_pInt .or. &
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
mod(minRes(2),2_pInt) /= 0_pInt .or. &
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
mod(minRes(3),2_pInt) /= 0_pInt) call IO_error(890_pInt, ext_msg = '3D minRes') ! as f2py has problems with present, use pyf file for initialization to -1
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
endif; endif
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! Automatic detection based on current geom
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
geomSizeNew = math_mul33x3(Favg,geomSize)
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
if (adaptive) then
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
ratio = floor(real(resNewInput,pReal) * (geomSizeNew/geomSize), pInt)
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
possibleResNew = 1_pInt
do i = 1_pInt, spatialDim
if (mod(ratio(i),2) == 0_pInt) then
possibleResNew(i,1:2) = [ratio(i),ratio(i) + 2_pInt]
else
possibleResNew(i,1:2) = [ratio(i)-1_pInt, ratio(i) + 1_pInt]
endif
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
if (.not.present(minRes)) then ! calling from fortran, optional argument not given
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
possibleResNew = possibleResNew
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
else ! optional argument is there
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
if (any(minRes<1_pInt)) then
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
possibleResNew = possibleResNew ! f2py calling, but without specification (or choosing invalid values), standard from pyf = -1
else ! given useful values
forall(k = 1_pInt:3_pInt, j = 1_pInt:3_pInt) &
possibleResNew(j,k) = max(possibleResNew(j,k), minRes(j))
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
endif
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
endif
enddo
adaptive regrid now working, updates F and F_lastInc based on mapping.
2012-08-10 21:18:27 +05:30
k = huge(1_pInt)
do i = 0_pInt, 2_pInt**spatialDim - 1
j = abs( possibleResNew(1,iand(i,1_pInt)/1_pInt + 1_pInt) &
* possibleResNew(2,iand(i,2_pInt)/2_pInt + 1_pInt) &
* possibleResNew(3,iand(i,4_pInt)/4_pInt + 1_pInt) &
- resNewInput(1)*resNewInput(2)*resNewInput(3))
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
adaptive regrid now working, updates F and F_lastInc based on mapping.
2012-08-10 21:18:27 +05:30
if (j < k) then
k = j
resNew =[ possibleResNew(1,iand(i,1_pInt)/1_pInt + 1_pInt), &
possibleResNew(2,iand(i,2_pInt)/2_pInt + 1_pInt), &
possibleResNew(3,iand(i,4_pInt)/4_pInt + 1_pInt) ]
endif
enddo
else
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
resNew = grid
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
endif
adaptive regrid now working, updates F and F_lastInc based on mapping.
2012-08-10 21:18:27 +05:30
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
mesh_regrid = resNew
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
NpointsNew = product(resNew)
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! Calculate regular new coordinates
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(coordinatesNew(3,NpointsNew))
ielem = 0_pInt
do k=1_pInt,resNew(3); do j=1_pInt, resNew(2); do i=1_pInt, resNew(1)
ielem = ielem + 1_pInt
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
coordinatesNew(1:3,ielem) = math_mul33x3(Favg, geomSize/real(resNew,pReal)*real([i,j,k],pReal) &
- geomSize/real(2_pInt*resNew,pReal))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo; enddo; enddo
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! Nearest neighbour search
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
allocate(indices(NpointsNew))
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
indices = math_periodicNearestNeighbor(geomSize, Favg, coordinatesNew, coordinates)
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
deallocate(coordinates)
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! write out indices periodic
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
write(N_Digits, '(I16.16)') 1_pInt + int(log10(real(maxval(indices),pReal)))
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
N_Digits = adjustl(N_Digits)
corrected some bugs concerning the regridding prevented FEsolving from potentially write to a none existing file started to introduce petsc into the make chain (nothing happens if PETSC_DIR is not set)
2012-06-20 18:19:46 +05:30
formatString = '(I'//trim(N_Digits)//'.'//trim(N_Digits)//',a)'
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
call IO_write_jobFile(777,'IDX') ! make it a general open-write file
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
write(777, '(A)') '1 header'
write(777, '(A)') 'Numbered indices as per the large set'
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
do i = 1_pInt, NpointsNew
corrected some bugs concerning the regridding prevented FEsolving from potentially write to a none existing file started to introduce petsc into the make chain (nothing happens if PETSC_DIR is not set)
2012-06-20 18:19:46 +05:30
write(777,trim(formatString),advance='no') indices(i), ' '
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
if(mod(i,resNew(1)) == 0_pInt) write(777,'(A)') ''
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
enddo
close(777)
added new 2D triangle elements added some sourcecode commenting on internal database formats corrected database to allow for ipVol and Area calc in 2D element cases
2011-04-06 14:05:37 +05:30
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! calculate and write out indices non periodic
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
do i = 1_pInt, NpointsNew
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
indices(i) = indices(i) / 3_pInt**spatialDim +1_pInt ! +1 b'coz index count starts from '0'
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
enddo
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
write(N_Digits, '(I16.16)') 1_pInt + int(log10(real(maxval(indices),pReal)))
N_Digits = adjustl(N_Digits)
formatString = '(I'//trim(N_Digits)//'.'//trim(N_Digits)//',a)'
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
call IO_write_jobFile(777,'idx') ! make it a general open-write file
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
write(777, '(A)') '1 header'
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
write(777, '(A)') 'Numbered indices as per the small set'
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
do i = 1_pInt, NpointsNew
write(777,trim(formatString),advance='no') indices(i), ' '
if(mod(i,resNew(1)) == 0_pInt) write(777,'(A)') ''
enddo
close(777)
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! write out new geom file
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
write(N_Digits, '(I16.16)') 1_pInt+int(log10(real(maxval(mesh_element(4,1:mesh_NcpElems)),pReal)),pInt)
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
N_Digits = adjustl(N_Digits)
corrected some bugs concerning the regridding prevented FEsolving from potentially write to a none existing file started to introduce petsc into the make chain (nothing happens if PETSC_DIR is not set)
2012-06-20 18:19:46 +05:30
formatString = '(I'//trim(N_Digits)//'.'//trim(N_Digits)//',a)'
open(777,file=trim(getSolverWorkingDirectoryName())//trim(GeometryFile),status='REPLACE')
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
write(777, '(A)') '3 header'
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
write(777, '(3(A, I8))') 'grid a ', resNew(1), ' b ', resNew(2), ' c ', resNew(3)
write(777, '(3(A, g17.10))') 'size x ', geomSize(1), ' y ', geomSize(2), ' z ', geomSize(3)
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
write(777, '(A)') 'homogenization 1'
do i = 1_pInt, NpointsNew
corrected some bugs concerning the regridding prevented FEsolving from potentially write to a none existing file started to introduce petsc into the make chain (nothing happens if PETSC_DIR is not set)
2012-06-20 18:19:46 +05:30
write(777,trim(formatString),advance='no') mesh_element(4,indices(i)), ' '
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
if(mod(i,resNew(1)) == 0_pInt) write(777,'(A)') ''
enddo
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
close(777)
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! set F to average values
select case(myspectralsolver)
case('basic')
allocate(spectralF33New(3,3,resNew(1),resNew(2),resNew(3)))
spectralF33New = spread(spread(spread(Favg,3,resNew(1)),4,resNew(2)),5,resNew(3))
call IO_write_jobBinaryFile(777,'F',size(spectralF33New))
write (777,rec=1) spectralF33New
close (777)
case('basicpetsc','al')
allocate(spectralF9New(9,resNew(1),resNew(2),resNew(3)))
spectralF9New = spread(spread(spread(reshape(Favg,[9]),2,resNew(1)),3,resNew(2)),4,resNew(3))
call IO_write_jobBinaryFile(777,'F',size(spectralF9New))
write (777,rec=1) spectralF9New
close (777)
end select
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!---------------------------------------------------------------------------------
allocate(F_lastIncNew(3,3,resNew(1),resNew(2),resNew(3)))
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
adaptive regrid now working, updates F and F_lastInc based on mapping.
2012-08-10 21:18:27 +05:30
call IO_read_jobBinaryFile(777,'F_aim_lastInc', &
trim(getSolverJobName()),size(Favg_LastInc))
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
read (777,rec=1) Favg_LastInc
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
close (777)
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
F_lastIncNew = spread(spread(spread(Favg_LastInc,3,resNew(1)),4,resNew(2)),5,resNew(3))
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
adaptive regrid now working, updates F and F_lastInc based on mapping.
2012-08-10 21:18:27 +05:30
call IO_write_jobBinaryFile(777,'convergedSpectralDefgrad_lastInc',size(F_LastIncNew))
write (777,rec=1) F_LastIncNew
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
close (777)
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
enabling regridding more than once by introducing deallocation of arrays added J2 test (stub from Taymour)
2012-07-31 21:07:49 +05:30
deallocate(F_lastIncNew)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
! relocating data of material subroutine ---------------------------------------------------------
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(material_phase (1,1, mesh_NcpElems))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(material_phaseNew (1,1, NpointsNew))
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
call IO_read_jobBinaryIntFile(777,'recordedPhase',trim(getSolverJobName()),size(material_phase))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (777,rec=1) material_phase
close (777)
do i = 1, NpointsNew
material_phaseNew(1,1,i) = material_phase(1,1,indices(i))
enddo
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
do i = 1, mesh_NcpElems
if (all(material_phaseNew(1,1,:) /= material_phase(1,1,i))) then
write(6,*) 'mismatch in regridding'
write(6,*) material_phase(1,1,i), 'not found in material_phaseNew'
endif
enddo
call IO_write_jobBinaryIntFile(777,'recordedPhase',size(material_phaseNew))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
write (777,rec=1) material_phaseNew
close (777)
deallocate(material_phase)
deallocate(material_phaseNew)
!---------------------------------------------------------------------------
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(F (3,3,1,1, mesh_NcpElems))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(FNew (3,3,1,1, NpointsNew))
call IO_read_jobBinaryFile(777,'convergedF',trim(getSolverJobName()),size(F))
read (777,rec=1) F
close (777)
do i = 1, NpointsNew
FNew(1:3,1:3,1,1,i) = F(1:3,1:3,1,1,indices(i))
enddo
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
call IO_write_jobBinaryFile(777,'convergedF',size(FNew))
write (777,rec=1) FNew
close (777)
deallocate(F)
deallocate(FNew)
!---------------------------------------------------------------------
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(Fp (3,3,1,1,mesh_NcpElems))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(FpNew (3,3,1,1,NpointsNew))
call IO_read_jobBinaryFile(777,'convergedFp',trim(getSolverJobName()),size(Fp))
read (777,rec=1) Fp
close (777)
do i = 1, NpointsNew
FpNew(1:3,1:3,1,1,i) = Fp(1:3,1:3,1,1,indices(i))
enddo
call IO_write_jobBinaryFile(777,'convergedFp',size(FpNew))
write (777,rec=1) FpNew
close (777)
deallocate(Fp)
deallocate(FpNew)
!------------------------------------------------------------------------
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(Lp (3,3,1,1,mesh_NcpElems))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(LpNew (3,3,1,1,NpointsNew))
call IO_read_jobBinaryFile(777,'convergedLp',trim(getSolverJobName()),size(Lp))
read (777,rec=1) Lp
close (777)
do i = 1, NpointsNew
LpNew(1:3,1:3,1,1,i) = Lp(1:3,1:3,1,1,indices(i))
enddo
call IO_write_jobBinaryFile(777,'convergedLp',size(LpNew))
write (777,rec=1) LpNew
close (777)
deallocate(Lp)
deallocate(LpNew)
!----------------------------------------------------------------------------
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(dcsdE (6,6,1,1,mesh_NcpElems))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(dcsdENew (6,6,1,1,NpointsNew))
call IO_read_jobBinaryFile(777,'convergeddcsdE',trim(getSolverJobName()),size(dcsdE))
read (777,rec=1) dcsdE
close (777)
do i = 1, NpointsNew
dcsdENew(1:6,1:6,1,1,i) = dcsdE(1:6,1:6,1,1,indices(i))
enddo
call IO_write_jobBinaryFile(777,'convergeddcsdE',size(dcsdENew))
write (777,rec=1) dcsdENew
close (777)
deallocate(dcsdE)
deallocate(dcsdENew)
!---------------------------------------------------------------------------
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(dPdF (3,3,3,3,1,1,mesh_NcpElems))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(dPdFNew (3,3,3,3,1,1,NpointsNew))
call IO_read_jobBinaryFile(777,'convergeddPdF',trim(getSolverJobName()),size(dPdF))
read (777,rec=1) dPdF
close (777)
do i = 1, NpointsNew
dPdFNew(1:3,1:3,1:3,1:3,1,1,i) = dPdF(1:3,1:3,1:3,1:3,1,1,indices(i))
enddo
call IO_write_jobBinaryFile(777,'convergeddPdF',size(dPdFNew))
write (777,rec=1) dPdFNew
close (777)
deallocate(dPdF)
deallocate(dPdFNew)
!---------------------------------------------------------------------------
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(Tstar (6,1,1,mesh_NcpElems))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(TstarNew (6,1,1,NpointsNew))
call IO_read_jobBinaryFile(777,'convergedTstar',trim(getSolverJobName()),size(Tstar))
read (777,rec=1) Tstar
close (777)
do i = 1, NpointsNew
TstarNew(1:6,1,1,i) = Tstar(1:6,1,1,indices(i))
enddo
call IO_write_jobBinaryFile(777,'convergedTstar',size(TstarNew))
write (777,rec=1) TstarNew
close (777)
deallocate(Tstar)
deallocate(TstarNew)
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
! for the state, we first have to know the size------------------------------------------------------------------
allocate(sizeStateConst(1,1,mesh_NcpElems))
call IO_read_jobBinaryIntFile(777,'sizeStateConst',trim(getSolverJobName()),size(sizeStateConst))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (777,rec=1) sizeStateConst
close (777)
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
maxsize = maxval(sizeStateConst(1,1,1:mesh_NcpElems))
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(StateConst (1,1,mesh_NcpElems,maxsize))
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
call IO_read_jobBinaryFile(777,'convergedStateConst',trim(getSolverJobName()))
k = 0_pInt
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
do i =1, mesh_NcpElems
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
do j = 1,sizeStateConst(1,1,i)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
k = k+1_pInt
read(777,rec=k) StateConst(1,1,i,j)
enddo
enddo
close(777)
call IO_write_jobBinaryFile(777,'convergedStateConst')
k = 0_pInt
do i = 1,NpointsNew
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
do j = 1,sizeStateConst(1,1,indices(i))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
k=k+1_pInt
write(777,rec=k) StateConst(1,1,indices(i),j)
enddo
enddo
close (777)
deallocate(sizeStateConst)
deallocate(StateConst)
!----------------------------------------------------------------------------
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
allocate(sizeStateHomog(1,mesh_NcpElems))
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
call IO_read_jobBinaryIntFile(777,'sizeStateHomog',trim(getSolverJobName()),size(sizeStateHomog))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (777,rec=1) sizeStateHomog
close (777)
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
maxsize = maxval(sizeStateHomog(1,1:mesh_NcpElems))
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
allocate(stateHomog (1,mesh_NcpElems,maxsize))
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
call IO_read_jobBinaryFile(777,'convergedStateHomog',trim(getSolverJobName()))
k = 0_pInt
moved calculation and output of geometry related data to mesh.f90, moved output of command line related information to DAMASK_spectral_interface.f90. improved new solver structure (use with make SOLVER=NEW)
2012-08-06 18:13:05 +05:30
do i =1, mesh_NcpElems
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do j = 1,sizeStateHomog(1,i)
k = k+1_pInt
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
read(777,rec=k) stateHomog(1,i,j)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo
enddo
close(777)
call IO_write_jobBinaryFile(777,'convergedStateHomog')
k = 0_pInt
do i = 1,NpointsNew
fixed some bugs in the regridding routine
2012-06-21 00:03:20 +05:30
do j = 1,sizeStateHomog(1,indices(i))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
k=k+1_pInt
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
write(777,rec=k) stateHomog(1,indices(i),j)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo
enddo
close (777)
deallocate(sizeStateHomog)
deallocate(stateHomog)
added minRes to regridding function and writing out of new geometry file updated f2py wrapper to enable the use of init functions. added 2 new error messages to io
2012-06-19 19:01:15 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
deallocate(indices)
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
write(6,*) 'finished regridding'
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end function mesh_regrid
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief builds mesh of (distorted) cubes for given coordinates (= center of the cubes)
!--------------------------------------------------------------------------------------------------
function mesh_nodesAroundCentres(gDim,Favg,centres) result(nodes)
use debug, only: &
debug_mesh, &
debug_level, &
debug_levelBasic
use math, only: &
math_mul33x3
added new 2D triangle elements added some sourcecode commenting on internal database formats corrected database to allow for ipVol and Area calc in 2D element cases
2011-04-06 14:05:37 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
implicit none
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
real(pReal), intent(in), dimension(:,:,:,:) :: &
centres
real(pReal), dimension(3,size(centres,2)+1,size(centres,3)+1,size(centres,4)+1) :: &
nodes
real(pReal), intent(in), dimension(3) :: &
gDim
real(pReal), intent(in), dimension(3,3) :: &
Favg
real(pReal), dimension(3,size(centres,2)+2,size(centres,3)+2,size(centres,4)+2) :: &
wrappedCentres
integer(pInt) :: &
i,j,k,n
integer(pInt), dimension(3), parameter :: &
diag = 1_pInt
integer(pInt), dimension(3) :: &
shift = 0_pInt, &
lookup = 0_pInt, &
me = 0_pInt, &
iRes = 0_pInt
integer(pInt), dimension(3,8) :: &
neighbor = reshape([ &
0_pInt, 0_pInt, 0_pInt, &
1_pInt, 0_pInt, 0_pInt, &
1_pInt, 1_pInt, 0_pInt, &
0_pInt, 1_pInt, 0_pInt, &
0_pInt, 0_pInt, 1_pInt, &
1_pInt, 0_pInt, 1_pInt, &
1_pInt, 1_pInt, 1_pInt, &
0_pInt, 1_pInt, 1_pInt ], [3,8])
!--------------------------------------------------------------------------------------------------
! initializing variables
iRes = [size(centres,2),size(centres,3),size(centres,4)]
nodes = 0.0_pReal
wrappedCentres = 0.0_pReal
!--------------------------------------------------------------------------------------------------
! report
if (iand(debug_level(debug_mesh),debug_levelBasic) /= 0_pInt) then
write(6,'(a)') ' Meshing cubes around centroids'
write(6,'(a,3(e12.5))') ' Dimension: ', gDim
write(6,'(a,3(i5))') ' Resolution:', iRes
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
endif
changed assumed arrays (:) to properly defined ones (gfortran 4.7 was complaining) and remove use statements that are not needed
2012-02-21 21:09:36 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! building wrappedCentres = centroids + ghosts
wrappedCentres(1:3,2_pInt:iRes(1)+1_pInt,2_pInt:iRes(2)+1_pInt,2_pInt:iRes(3)+1_pInt) = centres
do k = 0_pInt,iRes(3)+1_pInt
do j = 0_pInt,iRes(2)+1_pInt
do i = 0_pInt,iRes(1)+1_pInt
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
if (k==0_pInt .or. k==iRes(3)+1_pInt .or. & ! z skin
j==0_pInt .or. j==iRes(2)+1_pInt .or. & ! y skin
i==0_pInt .or. i==iRes(1)+1_pInt ) then ! x skin
me = [i,j,k] ! me on skin
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
shift = sign(abs(iRes+diag-2_pInt*me)/(iRes+diag),iRes+diag-2_pInt*me)
lookup = me-diag+shift*iRes
wrappedCentres(1:3,i+1_pInt, j+1_pInt, k+1_pInt) = &
centres(1:3,lookup(1)+1_pInt,lookup(2)+1_pInt,lookup(3)+1_pInt) - &
math_mul33x3(Favg, shift*gDim)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
endif
enddo; enddo; enddo
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! averaging
do k = 0_pInt,iRes(3); do j = 0_pInt,iRes(2); do i = 0_pInt,iRes(1)
do n = 1_pInt,8_pInt
nodes(1:3,i+1_pInt,j+1_pInt,k+1_pInt) = &
nodes(1:3,i+1_pInt,j+1_pInt,k+1_pInt) + wrappedCentres(1:3,i+1_pInt+neighbor(1,n), &
j+1_pInt+neighbor(2,n), &
k+1_pInt+neighbor(3,n) )
enddo
enddo; enddo; enddo
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
nodes = nodes/8.0_pReal
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
end function mesh_nodesAroundCentres
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief calculate coordinates in current configuration for given defgrad using linear
! interpolation
!--------------------------------------------------------------------------------------------------
function mesh_deformedCoordsLinear(gDim,F,FavgIn) result(coords)
use debug, only: &
debug_mesh, &
debug_level, &
debug_levelBasic
use math, only: &
math_mul33x3
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
implicit none
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
real(pReal), intent(in), dimension(:,:,:,:,:) :: &
F
real(pReal), dimension(3,size(F,3),size(F,4),size(F,5)) :: &
coords
real(pReal), intent(in), dimension(3) :: &
gDim
real(pReal), intent(in), dimension(3,3), optional :: &
FavgIn
real(pReal), dimension(3,0:size(F,3)-1,0:size(F,4)-1,0:size(F,5)-1,0:7) :: &
coordsAvgOrder
integer(pInt), parameter, dimension(3) :: &
improved reading in of values, now only warnings in case of problematic entries in material.config divergence calculation sqrt scaling optionally introduced for basic scheme spectral solver
2013-02-08 21:25:53 +05:30
iOnes = 1_pInt
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
real(pReal), parameter, dimension(3) :: &
fOnes = 1.0_pReal
real(pReal), dimension(3) :: &
myStep, &
negative, &
positive, &
offsetCoords, &
parameterCoords, &
stepLength, &
fRes
real(pReal), dimension(3,3) :: &
Favg
integer(pInt), dimension(3) :: &
rear, &
init, &
oppo, &
me, &
smallRes, &
iRes
integer(pInt) :: &
i, j, k, s, o
integer(pInt), parameter, dimension(3,0:7) :: &
corner = reshape([ &
spectral_build_neighborhood now "hidden" in spectral compile part. (should not interfere with Marc/Abaqus compile anymore)
2013-04-22 14:31:58 +05:30
0_pInt, 0_pInt, 0_pInt,&
1_pInt, 0_pInt, 0_pInt,&
1_pInt, 1_pInt, 0_pInt,&
0_pInt, 1_pInt, 0_pInt,&
1_pInt, 1_pInt, 1_pInt,&
0_pInt, 1_pInt, 1_pInt,&
0_pInt, 0_pInt, 1_pInt,&
1_pInt, 0_pInt, 1_pInt &
],[3,8]), &
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
step = reshape([&
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
1_pInt, 1_pInt, 1_pInt,&
-1_pInt, 1_pInt, 1_pInt,&
-1_pInt,-1_pInt, 1_pInt,&
1_pInt,-1_pInt, 1_pInt,&
-1_pInt,-1_pInt,-1_pInt,&
1_pInt,-1_pInt,-1_pInt,&
1_pInt, 1_pInt,-1_pInt,&
-1_pInt, 1_pInt,-1_pInt &
], [3,8])
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
integer(pInt), parameter, dimension(3,6) :: &
order = reshape([ &
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
1_pInt, 2_pInt, 3_pInt,&
1_pInt, 3_pInt, 2_pInt,&
2_pInt, 1_pInt, 3_pInt,&
2_pInt, 3_pInt, 1_pInt,&
3_pInt, 1_pInt, 2_pInt,&
3_pInt, 2_pInt, 1_pInt &
], [3,6])
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! initializing variables
iRes = [size(F,3),size(F,4),size(F,5)]
fRes = real(iRes,pReal)
smallRes = iRes - 1_pInt
coordsAvgOrder = 0.0_pReal
stepLength = gDim/fRes
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! report
if (iand(debug_level(debug_mesh),debug_levelBasic) /= 0_pInt) then
write(6,'(a)') ' Restore geometry using linear integration'
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
write(6,'(a,3(i12 ))') ' grid a b c: ', iRes
write(6,'(a,3(f12.5))') ' size x y z: ', gDim
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
endif
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! determine average deformation gradient
if (present(FavgIn)) then
if (all(FavgIn < 0.0_pReal)) then ! the f2py way to tell it is not present
Favg = sum(sum(sum(F,dim=5),dim=4),dim=3) / product(fRes)
else
Favg = FavgIn
endif
else
Favg = sum(sum(sum(F,dim=5),dim=4),dim=3) / product(fRes)
endif
!--------------------------------------------------------------------------------------------------
! loop over starting corners (from 0 to 7)
cornerLooping: do s = 0_pInt, 7_pInt
init = corner(1:3,s) * smallRes + iOnes
oppo = corner(1:3,mod((s+4_pInt),8_pInt)) * smallRes + iOnes
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! permutation of ways on each corner
permutationLooping: do o = 1_pInt,6_pInt
coords = 0.0_pReal
do k = init(order(3,o)), oppo(order(3,o)), step(order(3,o),s)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
rear(order(2,o)) = init(order(2,o))
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
do j = init(order(2,o)), oppo(order(2,o)), step(order(2,o),s)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
rear(order(1,o)) = init(order(1,o))
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
do i = init(order(1,o)), oppo(order(1,o)), step(order(1,o),s)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
me(order(1:3,o)) = [i,j,k]
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
if ( all(me == init)) then
coords(1:3,me(1),me(2),me(3)) = gDim*( math_mul33x3(Favg,real(corner(1:3,s),pReal)) &
+math_mul33x3(F(1:3,1:3,me(1),me(2),me(3)), &
0.5_pReal*real(step(1:3,s)/iRes,pReal)))
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
else
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
myStep = (me-rear)*stepLength
coords(1:3,me(1),me(2),me(3)) = coords(1:3,rear(1),rear(2),rear(3)) + &
0.5_pReal*math_mul33x3(F(1:3,1:3,me(1),me(2),me(3)) + &
F(1:3,1:3,rear(1),rear(2),rear(3)),myStep)
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
endif
rear = me
enddo; enddo; enddo
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
coordsAvgOrder(1:3,0:smallRes(1),0:smallRes(2),0:smallRes(3),s) = &
coordsAvgOrder(1:3,0:smallRes(1),0:smallRes(2),0:smallRes(3),s) + coords/6.0_pReal
enddo permutationLooping
offsetCoords = coordsAvgOrder(1:3,0,0,0,s)
do k = 0_pInt, smallRes(3); do j = 0_pInt, smallRes(2); do i = 0_pInt, smallRes(1)
coordsAvgOrder(1:3,i,j,k,s) = coordsAvgOrder(1:3,i,j,k,s) - offsetCoords
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
enddo; enddo; enddo
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
enddo cornerLooping
!--------------------------------------------------------------------------------------------------
! linear interpolation starting at each corner (comparable to linear shape function FEM)
do k = 0_pInt, smallRes(3); do j = 0_pInt, smallRes(2); do i = 0_pInt, smallRes(1)
parameterCoords = (2.0_pReal*real([i,j,k]+1,pReal)-fRes)/fRes
positive = fones + parameterCoords
negative = fones - parameterCoords
coords(1:3,i+1_pInt,j+1_pInt,k+1_pInt) &
=(coordsAvgOrder(1:3,i,j,k,0) *negative(1)*negative(2)*negative(3)&
+ coordsAvgOrder(1:3,i,j,k,1) *positive(1)*negative(2)*negative(3)&
+ coordsAvgOrder(1:3,i,j,k,2) *positive(1)*positive(2)*negative(3)&
+ coordsAvgOrder(1:3,i,j,k,3) *negative(1)*positive(2)*negative(3)&
+ coordsAvgOrder(1:3,i,j,k,4) *positive(1)*positive(2)*positive(3)&
+ coordsAvgOrder(1:3,i,j,k,5) *negative(1)*positive(2)*positive(3)&
+ coordsAvgOrder(1:3,i,j,k,6) *negative(1)*negative(2)*positive(3)&
+ coordsAvgOrder(1:3,i,j,k,7) *positive(1)*negative(2)*positive(3))*0.125_pReal
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
enddo; enddo; enddo
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! setting base node to (0,0,0)
offsetCoords = math_mul33x3(F(1:3,1:3,1,1,1),stepLength/2.0_pReal) - coords(1:3,1,1,1)
do k = 1_pInt, iRes(3); do j = 1_pInt, iRes(2); do i = 1_pInt, iRes(1)
coords(1:3,i,j,k) = coords(1:3,i,j,k) + offsetCoords
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
enddo; enddo; enddo
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
end function mesh_deformedCoordsLinear
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief calculate coordinates in current configuration for given defgrad
! using integration in Fourier space
!--------------------------------------------------------------------------------------------------
scaling of deformation gradient fluctuations now possible in x,y,z independently (give array).
2013-03-22 20:39:55 +05:30
function mesh_deformedCoordsFFT(gDim,F,FavgIn,scalingIn) result(coords)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
use IO, only: &
IO_error
use numerics, only: &
fftw_timelimit, &
fftw_planner_flag
use debug, only: &
debug_mesh, &
debug_level, &
debug_levelBasic
use math, only: &
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
PI, &
math_mul33x3
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
implicit none
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
real(pReal), intent(in), dimension(:,:,:,:,:) :: F
real(pReal), dimension(3,size(F,3),size(F,4),size(F,5)) :: coords
real(pReal), intent(in), dimension(3) :: gDim
real(pReal), intent(in), dimension(3,3), optional :: FavgIn
scaling of deformation gradient fluctuations now possible in x,y,z independently (give array).
2013-03-22 20:39:55 +05:30
real(pReal), intent(in), dimension(3), optional :: scalingIn
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
! allocatable arrays for fftw c routines
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
type(C_PTR) :: planForth, planBack
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
type(C_PTR) :: coords_fftw, defgrad_fftw
real(pReal), dimension(:,:,:,:,:), pointer :: F_real
complex(pReal), dimension(:,:,:,:,:), pointer :: F_fourier
real(pReal), dimension(:,:,:,:), pointer :: coords_real
complex(pReal), dimension(:,:,:,:), pointer :: coords_fourier
! other variables
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
integer(pInt) :: i, j, k, m, res1Red
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
integer(pInt), dimension(3) :: k_s, iRes
scaling of deformation gradient fluctuations now possible in x,y,z independently (give array).
2013-03-22 20:39:55 +05:30
real(pReal), dimension(3) :: scaling, step, offset_coords, integrator
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
real(pReal), dimension(3,3) :: Favg
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
integer(pInt), dimension(2:3,2) :: Nyquist ! highest frequencies to be removed (1 if even, 2 if odd)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
if (present(scalingIn)) then
fixed bug causing no fluctuation in case of using with 3Dvisualize
2013-04-25 20:13:20 +05:30
where (scalingIn < 0.0_pReal) ! invalid values. in case of f2py -1 if not present
scaling of deformation gradient fluctuations now possible in x,y,z independently (give array).
2013-03-22 20:39:55 +05:30
scaling = [1.0_pReal,1.0_pReal,1.0_pReal]
elsewhere
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
scaling = scalingIn
fixed bug causing no fluctuation in case of using with 3Dvisualize
2013-04-25 20:13:20 +05:30
end where
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
else
scaling = 1.0_pReal
endif
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
iRes = [size(F,3),size(F,4),size(F,5)]
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
integrator = gDim / 2.0_pReal / PI ! see notes where it is used
fixed bug causing no fluctuation in case of using with 3Dvisualize
2013-04-25 20:13:20 +05:30
res1Red = iRes(1)/2_pInt + 1_pInt ! size of complex array in first dimension (c2r, r2c)
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
step = gDim/real(iRes, pReal)
fixed bug causing no fluctuation in case of using with 3Dvisualize
2013-04-25 20:13:20 +05:30
Nyquist(2,1:2) = [iRes(2)/2_pInt + 1_pInt, iRes(2)/2_pInt + 1_pInt + mod(iRes(2),2_pInt)]
Nyquist(3,1:2) = [iRes(3)/2_pInt + 1_pInt, iRes(3)/2_pInt + 1_pInt + mod(iRes(3),2_pInt)]
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! report
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
if (iand(debug_level(debug_mesh),debug_levelBasic) /= 0_pInt) then
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
write(6,'(a)') ' Restore geometry using FFT-based integration'
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
write(6,'(a,3(i12 ))') ' grid a b c: ', iRes
write(6,'(a,3(f12.5))') ' size x y z: ', gDim
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
endif
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!--------------------------------------------------------------------------------------------------
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
! sanity check
made restart working for all solvers and added corresponding test corrected error handling for FFT based function in case of odd numbers etc.
2012-12-14 23:00:22 +05:30
if (pReal /= C_DOUBLE .or. pInt /= C_INT) &
call IO_error(0_pInt,ext_msg='Fortran to C in mesh_deformedCoordsFFT')
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!--------------------------------------------------------------------------------------------------
! allocation and FFTW initialization
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
defgrad_fftw = fftw_alloc_complex(int(res1Red *iRes(2)*iRes(3)*9_pInt,C_SIZE_T)) ! C_SIZE_T is of type integer(8)
coords_fftw = fftw_alloc_complex(int(res1Red *iRes(2)*iRes(3)*3_pInt,C_SIZE_T)) ! C_SIZE_T is of type integer(8)
call c_f_pointer(defgrad_fftw, F_real, &
[iRes(1)+2_pInt-mod(iRes(1),2_pInt),iRes(2),iRes(3),3_pInt,3_pInt])
call c_f_pointer(defgrad_fftw, F_fourier, &
[res1Red, iRes(2),iRes(3),3_pInt,3_pInt])
call c_f_pointer(coords_fftw, coords_real, &
[iRes(1)+2_pInt-mod(iRes(1),2_pInt),iRes(2),iRes(3),3_pInt])
call c_f_pointer(coords_fftw, coords_fourier, &
[res1Red, iRes(2),iRes(3),3_pInt])
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
call fftw_set_timelimit(fftw_timelimit)
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
planForth = fftw_plan_many_dft_r2c(3_pInt,[iRes(3),iRes(2) ,iRes(1)],9_pInt,& ! dimensions , length in each dimension in reversed order
F_real,[iRes(3),iRes(2) ,iRes(1)+2_pInt-mod(iRes(1),2_pInt)],& ! input data , physical length in each dimension in reversed order
1_pInt, iRes(3)*iRes(2)*(iRes(1)+2_pInt-mod(iRes(1),2_pInt)),& ! striding , product of physical lenght in the 3 dimensions
F_fourier,[iRes(3),iRes(2) ,res1Red],&
1_pInt, iRes(3)*iRes(2)* res1Red,fftw_planner_flag)
planBack = fftw_plan_many_dft_c2r(3_pInt,[iRes(3),iRes(2) ,iRes(1)],3_pInt,&
coords_fourier,[iRes(3),iRes(2) ,res1Red],&
1_pInt, iRes(3)*iRes(2)* res1Red,&
coords_real,[iRes(3),iRes(2) ,iRes(1)+2_pInt-mod(iRes(1),2_pInt)],&
1_pInt, iRes(3)*iRes(2)*(iRes(1)+2_pInt-mod(iRes(1),2_pInt)),&
fftw_planner_flag)
F_real(1:iRes(1),1:iRes(2),1:iRes(3),1:3,1:3) = &
fixed bug causing no fluctuation in case of using with 3Dvisualize
2013-04-25 20:13:20 +05:30
reshape(F,[iRes(1),iRes(2),iRes(3),3,3], order = [4,5,1,2,3]) ! F_real is overwritten during plan creatio, is larger (padding) and has different order
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!--------------------------------------------------------------------------------------------------
! FFT
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
call fftw_execute_dft_r2c(planForth, F_real, F_fourier)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!--------------------------------------------------------------------------------------------------
! if no average F is given, compute it in Fourier space
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
if (present(FavgIn)) then
if (all(FavgIn < 0.0_pReal)) then
fixed bug causing no fluctuation in case of using with 3Dvisualize
2013-04-25 20:13:20 +05:30
Favg = real(F_fourier(1,1,1,1:3,1:3),pReal)/real(product(iRes),pReal) !the f2py way to tell it is not present
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
else
Favg = FavgIn
endif
else
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
Favg = real(F_fourier(1,1,1,1:3,1:3),pReal)/real(product(iRes),pReal)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
endif
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!--------------------------------------------------------------------------------------------------
! remove highest frequency in each direction, in third direction only if not 2D
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
if(iRes(1)/=1_pInt) & ! do not delete the whole slice in case of 2D calculation
F_fourier (res1Red, 1:iRes(2), 1:iRes(3), 1:3,1:3) &
= cmplx(0.0_pReal,0.0_pReal,pReal)
if(iRes(2)/=1_pInt) & ! do not delete the whole slice in case of 2D calculation
F_fourier (1:res1Red,Nyquist(2,1):Nyquist(2,2),1:iRes(3), 1:3,1:3) &
= cmplx(0.0_pReal,0.0_pReal,pReal)
if(iRes(3)/=1_pInt) & ! do not delete the whole slice in case of 2D calculation
F_fourier (1:res1Red,1:iRes(2), Nyquist(3,1):Nyquist(3,2),1:3,1:3) &
= cmplx(0.0_pReal,0.0_pReal,pReal)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!--------------------------------------------------------------------------------------------------
! integration in Fourier space
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
coords_fourier = cmplx(0.0_pReal,0.0_pReal,pReal)
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
do k = 1_pInt, iRes(3)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
k_s(3) = k-1_pInt
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
if(k > iRes(3)/2_pInt+1_pInt) k_s(3) = k_s(3)-iRes(3)
do j = 1_pInt, iRes(2)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
k_s(2) = j-1_pInt
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
if(j > iRes(2)/2_pInt+1_pInt) k_s(2) = k_s(2)-iRes(2)
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
do i = 1_pInt, res1Red
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
k_s(1) = i-1_pInt
do m = 1_pInt,3_pInt
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
coords_fourier(i,j,k,m) = sum(F_fourier(i,j,k,m,1:3)*&
cmplx(0.0_pReal,real(k_s,pReal)*integrator,pReal))
enddo
if (any(k_s /= 0_pInt)) coords_fourier(i,j,k,1:3) = &
coords_fourier(i,j,k,1:3) / cmplx(-sum(k_s*k_s),0.0_pReal,pReal)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
enddo; enddo; enddo
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!--------------------------------------------------------------------------------------------------
! iFFT and freeing memory
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
call fftw_execute_dft_c2r(planBack,coords_fourier,coords_real)
coords = reshape(coords_real(1:iRes(1),1:iRes(2),1:iRes(3),1:3), [3,iRes(1),iRes(2),iRes(3)], &
order = [2,3,4,1])/real(product(iRes),pReal) ! weight and change order
call fftw_destroy_plan(planForth)
call fftw_destroy_plan(planBack)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
call fftw_free(defgrad_fftw)
call fftw_free(coords_fftw)
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!--------------------------------------------------------------------------------------------------
! add average to scaled fluctuation and put (0,0,0) on (0,0,0)
offset_coords = math_mul33x3(F(1:3,1:3,1,1,1),step/2.0_pReal) - scaling*coords(1:3,1,1,1)
forall(k = 1_pInt:iRes(3), j = 1_pInt:iRes(2), i = 1_pInt:iRes(1)) &
coords(1:3,i,j,k) = scaling(1:3)*coords(1:3,i,j,k) &
+ offset_coords &
+ math_mul33x3(Favg,step*real([i,j,k]-1_pInt,pReal))
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
end function mesh_deformedCoordsFFT
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
!> @brief calculates the mismatch between volume of reconstructed (compatible) cube and
! determinant of defgrad at the FP
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
function mesh_volumeMismatch(gDim,F,nodes) result(vMismatch)
use IO, only: &
IO_error
use debug, only: &
debug_mesh, &
debug_level, &
debug_levelBasic
use math, only: &
PI, &
math_det33, &
math_volTetrahedron
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
implicit none
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
real(pReal), intent(in), dimension(:,:,:,:,:) :: &
F
real(pReal), dimension(size(F,3),size(F,4),size(F,5)) :: &
vMismatch
real(pReal), intent(in), dimension(:,:,:,:) :: &
nodes
real(pReal), dimension(3) :: &
gDim
integer(pInt), dimension(3) :: &
iRes
real(pReal), dimension(3,8) :: coords
integer(pInt) :: i,j,k
real(pReal) :: volInitial
iRes = [size(F,3),size(F,4),size(F,5)]
volInitial = product(gDim)/real(product(iRes), pReal)
!--------------------------------------------------------------------------------------------------
! report and check
if (iand(debug_level(debug_mesh),debug_levelBasic) /= 0_pInt) then
write(6,'(a)') ' Calculating volume mismatch'
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
write(6,'(a,3(i12 ))') ' grid a b c: ', iRes
write(6,'(a,3(f12.5))') ' size x y z: ', gDim
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
endif
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
if (any([iRes/=size(nodes,2)-1_pInt,iRes/=size(nodes,3)-1_pInt,iRes/=size(nodes,4)-1_pInt]))&
call IO_error(0_pInt,ext_msg='Arrays F and nodes in mesh_volumeMismatch')
!--------------------------------------------------------------------------------------------------
! calculate actual volume and volume resulting from deformation gradient
do k = 1_pInt,iRes(3)
do j = 1_pInt,iRes(2)
do i = 1_pInt,iRes(1)
coords(1:3,1) = nodes(1:3,i, j, k )
coords(1:3,2) = nodes(1:3,i+1_pInt,j, k )
coords(1:3,3) = nodes(1:3,i+1_pInt,j+1_pInt,k )
coords(1:3,4) = nodes(1:3,i, j+1_pInt,k )
coords(1:3,5) = nodes(1:3,i, j, k+1_pInt)
coords(1:3,6) = nodes(1:3,i+1_pInt,j, k+1_pInt)
coords(1:3,7) = nodes(1:3,i+1_pInt,j+1_pInt,k+1_pInt)
coords(1:3,8) = nodes(1:3,i, j+1_pInt,k+1_pInt)
vMismatch(i,j,k) = &
abs(math_volTetrahedron(coords(1:3,7),coords(1:3,1),coords(1:3,8),coords(1:3,4))) &
+ abs(math_volTetrahedron(coords(1:3,7),coords(1:3,1),coords(1:3,8),coords(1:3,5))) &
+ abs(math_volTetrahedron(coords(1:3,7),coords(1:3,1),coords(1:3,3),coords(1:3,4))) &
+ abs(math_volTetrahedron(coords(1:3,7),coords(1:3,1),coords(1:3,3),coords(1:3,2))) &
+ abs(math_volTetrahedron(coords(1:3,7),coords(1:3,5),coords(1:3,2),coords(1:3,6))) &
+ abs(math_volTetrahedron(coords(1:3,7),coords(1:3,5),coords(1:3,2),coords(1:3,1)))
vMismatch(i,j,k) = vMismatch(i,j,k)/math_det33(F(1:3,1:3,i,j,k))
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
enddo; enddo; enddo
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
vMismatch = vMismatch/volInitial
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
end function mesh_volumeMismatch
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Routine to calculate the mismatch between the vectors from the central point to
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
! the corners of reconstructed (combatible) volume element and the vectors calculated by deforming
! the initial volume element with the current deformation gradient
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
function mesh_shapeMismatch(gDim,F,nodes,centres) result(sMismatch)
use IO, only: &
IO_error
use debug, only: &
debug_mesh, &
debug_level, &
debug_levelBasic
use math, only: &
math_mul33x3
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
implicit none
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
real(pReal), intent(in), dimension(:,:,:,:,:) :: &
F
real(pReal), dimension(size(F,3),size(F,4),size(F,5)) :: &
sMismatch
real(pReal), intent(in), dimension(:,:,:,:) :: &
nodes, &
centres
real(pReal), dimension(3) :: &
gDim, &
fRes
integer(pInt), dimension(3) :: &
iRes
real(pReal), dimension(3,8) :: coordsInitial
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
integer(pInt) i,j,k
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
iRes = [size(F,3),size(F,4),size(F,5)]
fRes = real(iRes,pReal)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
!--------------------------------------------------------------------------------------------------
! report and check
if (iand(debug_level(debug_mesh),debug_levelBasic) /= 0_pInt) then
write(6,'(a)') ' Calculating shape mismatch'
allowed arbitrary grid (formerly know as resolution), i.e. any integer>1 in each dimension is possible. removed square root correction for divergence (doesn't make sense) fixed FFT debug of spectral solver
2013-04-10 15:49:16 +05:30
write(6,'(a,3(i12 ))') ' grid a b c: ', iRes
write(6,'(a,3(f12.5))') ' size x y z: ', gDim
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
endif
if(any([iRes/=size(nodes,2)-1_pInt,iRes/=size(nodes,3)-1_pInt,iRes/=size(nodes,4)-1_pInt]) .or.&
any([iRes/=size(centres,2), iRes/=size(centres,3), iRes/=size(centres,4)]))&
call IO_error(0_pInt,ext_msg='Arrays F and nodes/centres in mesh_shapeMismatch')
!--------------------------------------------------------------------------------------------------
! initial positions
coordsInitial(1:3,1) = [-gDim(1)/fRes(1),-gDim(2)/fRes(2),-gDim(3)/fRes(3)]
coordsInitial(1:3,2) = [+gDim(1)/fRes(1),-gDim(2)/fRes(2),-gDim(3)/fRes(3)]
coordsInitial(1:3,3) = [+gDim(1)/fRes(1),+gDim(2)/fRes(2),-gDim(3)/fRes(3)]
coordsInitial(1:3,4) = [-gDim(1)/fRes(1),+gDim(2)/fRes(2),-gDim(3)/fRes(3)]
coordsInitial(1:3,5) = [-gDim(1)/fRes(1),-gDim(2)/fRes(2),+gDim(3)/fRes(3)]
coordsInitial(1:3,6) = [+gDim(1)/fRes(1),-gDim(2)/fRes(2),+gDim(3)/fRes(3)]
coordsInitial(1:3,7) = [+gDim(1)/fRes(1),+gDim(2)/fRes(2),+gDim(3)/fRes(3)]
coordsInitial(1:3,8) = [-gDim(1)/fRes(1),+gDim(2)/fRes(2),+gDim(3)/fRes(3)]
coordsInitial = coordsInitial/2.0_pReal
!--------------------------------------------------------------------------------------------------
! compare deformed original and deformed positions to actual positions
do k = 1_pInt,iRes(3)
do j = 1_pInt,iRes(2)
do i = 1_pInt,iRes(1)
sMismatch(i,j,k) = &
sqrt(sum((nodes(1:3,i, j, k ) - centres(1:3,i,j,k)&
- math_mul33x3(F(1:3,1:3,i,j,k), coordsInitial(1:3,1)))**2.0_pReal))&
+ sqrt(sum((nodes(1:3,i+1_pInt,j, k ) - centres(1:3,i,j,k)&
- math_mul33x3(F(1:3,1:3,i,j,k), coordsInitial(1:3,2)))**2.0_pReal))&
+ sqrt(sum((nodes(1:3,i+1_pInt,j+1_pInt,k ) - centres(1:3,i,j,k)&
- math_mul33x3(F(1:3,1:3,i,j,k), coordsInitial(1:3,3)))**2.0_pReal))&
+ sqrt(sum((nodes(1:3,i, j+1_pInt,k ) - centres(1:3,i,j,k)&
- math_mul33x3(F(1:3,1:3,i,j,k), coordsInitial(1:3,4)))**2.0_pReal))&
+ sqrt(sum((nodes(1:3,i, j, k+1_pInt) - centres(1:3,i,j,k)&
- math_mul33x3(F(1:3,1:3,i,j,k), coordsInitial(1:3,5)))**2.0_pReal))&
+ sqrt(sum((nodes(1:3,i+1_pInt,j, k+1_pInt) - centres(1:3,i,j,k)&
- math_mul33x3(F(1:3,1:3,i,j,k), coordsInitial(1:3,6)))**2.0_pReal))&
+ sqrt(sum((nodes(1:3,i+1_pInt,j+1_pInt,k+1_pInt) - centres(1:3,i,j,k)&
- math_mul33x3(F(1:3,1:3,i,j,k), coordsInitial(1:3,7)))**2.0_pReal))&
+ sqrt(sum((nodes(1:3,i, j+1_pInt,k+1_pInt) - centres(1:3,i,j,k)&
- math_mul33x3(F(1:3,1:3,i,j,k), coordsInitial(1:3,8)))**2.0_pReal))
moved some more 'mesh related' functions for post processing from math.f90 to mesh.f90 f2py functions remaining in math.f90 now uses assumed size arrays in order to have simpler interfaces. This is only working with python 2.7! changed python pre- and postprocessing scripts. If you encounter any problems whith core modules, try to remove the old core.so in the lib/damask
2012-08-27 13:34:47 +05:30
enddo; enddo; enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
end function mesh_shapeMismatch
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#endif
removed #define Marc and substituted #ifdef Marc by #ifdef Marc4DAMASK which is defined by either the submit scripts or DAMASK_marcXXXX.f90
2013-04-30 15:19:30 +05:30
#ifdef Marc4DAMASK
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Figures out table styles (Marc only) and stores to 'initialcondTableStyle' and
!! 'hypoelasticTableStyle'
!--------------------------------------------------------------------------------------------------
subroutine mesh_marc_get_tableStyles(myUnit)
use IO, only: &
IO_lc, &
IO_intValue, &
IO_stringValue, &
IO_stringPos
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
added function mesh_FEtoCPelement(FEid)
2007-03-21 21:48:33 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 6_pInt
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
integer(pInt), dimension (1+2*maxNchunks) :: myPos
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
character(len=300) line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
initialcondTableStyle = 0_pInt
hypoelasticTableStyle = 0_pInt
added function calls in init()
2007-04-10 16:52:53 +05:30
610 FORMAT(A300)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
fixed bug in coordinate calculation when no average F was given (scaled wrong by ncp_elems**2) other changes: just polishing + some more comments
2013-03-27 17:58:55 +05:30
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'table' .and. myPos(1_pInt) > 5) then
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
initialcondTableStyle = IO_intValue(line,myPos,4_pInt)
hypoelasticTableStyle = IO_intValue(line,myPos,5_pInt)
keyword "table" appeared more than once -- restricted interpretation to header
2007-09-28 20:26:26 +05:30
exit
added a new module called numerics.f90 which reads in all numerical "parameters" from the file numerics.config (also being added). From now on this file has to be located in the working directory of the FEM-model and has to contain all necessary parameters.
2009-06-15 18:41:21 +05:30
endif
enddo
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
620 end subroutine mesh_marc_get_tableStyles
Simplified 'mesh_get_nodeElemDimensions', mesh_build_elements' and 'mesh_build_sharedElems'. Corrected entries in 'FE_subNodeOnIPFace'
2009-04-03 12:33:25 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!> @brief Count overall number of nodes and elements in mesh and stores the numbers in
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!! 'mesh_Nelems' and 'mesh_Nnodes'
!--------------------------------------------------------------------------------------------------
subroutine mesh_marc_count_nodesAndElements(myUnit)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_IntValue
added awareness of set definitions (new function build_elemSetMapping) added new style tables for initial conditions
2007-10-23 18:39:46 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 4_pInt
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
integer(pInt), dimension (1+2*maxNchunks) :: myPos
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
character(len=300) line
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
mesh_Nnodes = 0_pInt
mesh_Nelems = 0_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
if ( IO_lc(IO_StringValue(line,myPos,1_pInt)) == 'sizing') &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_Nelems = IO_IntValue (line,myPos,3_pInt)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
if ( IO_lc(IO_StringValue(line,myPos,1_pInt)) == 'coordinates') then
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
mesh_Nnodes = IO_IntValue (line,myPos,2_pInt)
exit ! assumes that "coordinates" comes later in file
added a new module called numerics.f90 which reads in all numerical "parameters" from the file numerics.config (also being added). From now on this file has to be located in the working directory of the FEM-model and has to contain all necessary parameters.
2009-06-15 18:41:21 +05:30
endif
enddo
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
620 end subroutine mesh_marc_count_nodesAndElements
!--------------------------------------------------------------------------------------------------
!> @brief Count overall number of element sets in mesh. Stores to 'mesh_NelemSets', and
!! 'mesh_maxNelemInSet'
!--------------------------------------------------------------------------------------------------
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
subroutine mesh_marc_count_elementSets(myUnit)
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
IO_countContinuousIntValues
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
corrected mesh_build_elements (unit) so that state variables are read correctly from input file
2007-04-25 13:03:24 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), parameter :: maxNchunks = 2_pInt
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
integer(pInt), dimension (1+2*maxNchunks) :: myPos
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
character(len=300) line
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
mesh_NelemSets = 0_pInt
mesh_maxNelemInSet = 0_pInt
610 FORMAT(A300)
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
if ( IO_lc(IO_StringValue(line,myPos,1_pInt)) == 'define' .and. &
IO_lc(IO_StringValue(line,myPos,2_pInt)) == 'element' ) then
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
mesh_NelemSets = mesh_NelemSets + 1_pInt
mesh_maxNelemInSet = max(mesh_maxNelemInSet, &
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
IO_countContinuousIntValues(myUnit))
added a new module called numerics.f90 which reads in all numerical "parameters" from the file numerics.config (also being added). From now on this file has to be located in the working directory of the FEM-model and has to contain all necessary parameters.
2009-06-15 18:41:21 +05:30
endif
enddo
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
620 end subroutine mesh_marc_count_elementSets
corrected mesh_build_elements (unit) so that state variables are read correctly from input file
2007-04-25 13:03:24 +05:30
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added function calls in init()
2007-04-10 16:52:53 +05:30
!********************************************************************
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! map element sets
added function calls in init()
2007-04-10 16:52:53 +05:30
!
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! allocate globals: mesh_nameElemSet, mesh_mapElemSet
added function calls in init()
2007-04-10 16:52:53 +05:30
!********************************************************************
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_marc_map_elementSets(myUnit)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_continuousIntValues
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added function calls in init()
2007-04-10 16:52:53 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 4_pInt
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
integer(pInt), dimension (1+2*maxNchunks) :: myPos
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
character(len=300) :: line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt) :: elemSet = 0_pInt
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate (mesh_nameElemSet(mesh_NelemSets)) ; mesh_nameElemSet = ''
allocate (mesh_mapElemSet(1_pInt+mesh_maxNelemInSet,mesh_NelemSets)) ; mesh_mapElemSet = 0_pInt
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
corrected mesh_build_elements (unit) so that state variables are read correctly from input file
2007-04-25 13:03:24 +05:30
610 FORMAT(A300)
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do
read (myUnit,610,END=640) line
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if( (IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'define' ) .and. &
(IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'element' ) ) then
elemSet = elemSet+1_pInt
mesh_nameElemSet(elemSet) = trim(IO_stringValue(line,myPos,4_pInt))
mesh_mapElemSet(:,elemSet) = IO_continuousIntValues(myUnit,mesh_maxNelemInSet,mesh_nameElemSet,mesh_mapElemSet,mesh_NelemSets)
endif
added a new module called numerics.f90 which reads in all numerical "parameters" from the file numerics.config (also being added). From now on this file has to be located in the working directory of the FEM-model and has to contain all necessary parameters.
2009-06-15 18:41:21 +05:30
enddo
adopted improvements done by Arun Prakash. mesh: elemType identification based on lower case Abaqus now reports more errors IO: new function to inquire whether inputfile contains "parts" new function to assemble multiply included inputfile into a flat one awareness of range generation in element numbers error reporting
2010-07-13 15:56:07 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
640 end subroutine mesh_marc_map_elementSets
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Count overall number of CP elements in mesh and stores them in 'mesh_NcpElems'
!--------------------------------------------------------------------------------------------------
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
subroutine mesh_marc_count_cpElements(myUnit)
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
IO_countContinuousIntValues
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
corrected mesh_build_elements (unit) so that state variables are read correctly from input file
2007-04-25 13:03:24 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), parameter :: maxNchunks = 1_pInt
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
integer(pInt), dimension (1+2*maxNchunks) :: myPos
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt) :: i
character(len=300):: line
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
mesh_NcpElems = 0_pInt
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
corrected mesh_build_elements (unit) so that state variables are read correctly from input file
2007-04-25 13:03:24 +05:30
610 FORMAT(A300)
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
Simplified 'mesh_get_nodeElemDimensions', mesh_build_elements' and 'mesh_build_sharedElems'. Corrected entries in 'FE_subNodeOnIPFace'
2009-04-03 12:33:25 +05:30
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'hypoelastic') then
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do i=1_pInt,3_pInt+hypoelasticTableStyle ! Skip 3 or 4 lines
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
read (myUnit,610,END=620) line
Simplified 'mesh_get_nodeElemDimensions', mesh_build_elements' and 'mesh_build_sharedElems'. Corrected entries in 'FE_subNodeOnIPFace'
2009-04-03 12:33:25 +05:30
enddo
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
mesh_NcpElems = mesh_NcpElems + IO_countContinuousIntValues(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
exit
Simplified 'mesh_get_nodeElemDimensions', mesh_build_elements' and 'mesh_build_sharedElems'. Corrected entries in 'FE_subNodeOnIPFace'
2009-04-03 12:33:25 +05:30
endif
enddo
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
620 end subroutine mesh_marc_count_cpElements
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Maps elements from FE ID to internal (consecutive) representation.
!! Allocates global array 'mesh_mapFEtoCPelem'
!--------------------------------------------------------------------------------------------------
subroutine mesh_marc_map_elements(myUnit)
changed misplaced private to public statement and renamed qsort to math_qsort
2013-02-01 12:56:21 +05:30
use math, only: math_qsort
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_continuousIntValues
added function calls in init()
2007-04-10 16:52:53 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 1_pInt
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
character(len=300) line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), dimension (1_pInt+mesh_NcpElems) :: contInts
integer(pInt) :: i,cpElem = 0_pInt
allocate (mesh_mapFEtoCPelem(2,mesh_NcpElems)) ; mesh_mapFEtoCPelem = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do
read (myUnit,610,END=660) line
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if( IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'hypoelastic' ) then
do i=1_pInt,3_pInt+hypoelasticTableStyle ! skip three (or four if new table style!) lines
read (myUnit,610,END=660) line
enddo
contInts = IO_continuousIntValues(myUnit,mesh_NcpElems,mesh_nameElemSet,&
mesh_mapElemSet,mesh_NelemSets)
do i = 1_pInt,contInts(1)
cpElem = cpElem+1_pInt
mesh_mapFEtoCPelem(1,cpElem) = contInts(1_pInt+i)
mesh_mapFEtoCPelem(2,cpElem) = cpElem
enddo
endif
added a new module called numerics.f90 which reads in all numerical "parameters" from the file numerics.config (also being added). From now on this file has to be located in the working directory of the FEM-model and has to contain all necessary parameters.
2009-06-15 18:41:21 +05:30
enddo
adopted improvements done by Arun Prakash. mesh: elemType identification based on lower case Abaqus now reports more errors IO: new function to inquire whether inputfile contains "parts" new function to assemble multiply included inputfile into a flat one awareness of range generation in element numbers error reporting
2010-07-13 15:56:07 +05:30
changed misplaced private to public statement and renamed qsort to math_qsort
2013-02-01 12:56:21 +05:30
660 call math_qsort(mesh_mapFEtoCPelem,1_pInt,int(size(mesh_mapFEtoCPelem,2_pInt),pInt)) ! should be mesh_NcpElems
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_marc_map_elements
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Maps node from FE ID to internal (consecutive) representation.
!! Allocates global array 'mesh_mapFEtoCPnode'
!--------------------------------------------------------------------------------------------------
subroutine mesh_marc_map_nodes(myUnit)
changed misplaced private to public statement and renamed qsort to math_qsort
2013-02-01 12:56:21 +05:30
use math, only: math_qsort
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_fixedIntValue
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 1_pInt
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
character(len=300) line
integer(pInt), dimension (mesh_Nnodes) :: node_count
integer(pInt) :: i
allocate (mesh_mapFEtoCPnode(2_pInt,mesh_Nnodes)) ; mesh_mapFEtoCPnode = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
node_count = 0_pInt
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (myUnit,610,END=650) line
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if( IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'coordinates' ) then
read (myUnit,610,END=650) line ! skip crap line
do i = 1_pInt,mesh_Nnodes
read (myUnit,610,END=650) line
mesh_mapFEtoCPnode(1_pInt,i) = IO_fixedIntValue (line,[ 0_pInt,10_pInt],1_pInt)
mesh_mapFEtoCPnode(2_pInt,i) = i
enddo
exit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
changed misplaced private to public statement and renamed qsort to math_qsort
2013-02-01 12:56:21 +05:30
650 call math_qsort(mesh_mapFEtoCPnode,1_pInt,int(size(mesh_mapFEtoCPnode,2_pInt),pInt))
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_marc_map_nodes
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief store x,y,z coordinates of all nodes in mesh.
!! Allocates global arrays 'mesh_node0' and 'mesh_node'
!--------------------------------------------------------------------------------------------------
subroutine mesh_marc_build_nodes(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
use IO, only: &
IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_fixedIntValue, &
IO_fixedNoEFloatValue
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), dimension(5), parameter :: node_ends = int([0,10,30,50,70],pInt)
integer(pInt), parameter :: maxNchunks = 1_pInt
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
character(len=300) :: line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt) :: i,j,m
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate ( mesh_node0 (3,mesh_Nnodes) ); mesh_node0 = 0.0_pReal
allocate ( mesh_node (3,mesh_Nnodes) ); mesh_node = 0.0_pReal
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (myUnit,610,END=670) line
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if( IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'coordinates' ) then
read (myUnit,610,END=670) line ! skip crap line
do i=1_pInt,mesh_Nnodes
read (myUnit,610,END=670) line
m = mesh_FEasCP('node',IO_fixedIntValue(line,node_ends,1_pInt))
as some IO functions are no longer pure I changed some forall statements to ordinary do loops
2013-02-07 14:22:47 +05:30
do j = 1_pInt,3_pInt
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
mesh_node0(j,m) = mesh_unitlength * IO_fixedNoEFloatValue(line,node_ends,j+1_pInt)
as some IO functions are no longer pure I changed some forall statements to ordinary do loops
2013-02-07 14:22:47 +05:30
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo
exit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
670 mesh_node = mesh_node0
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_marc_build_nodes
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!> @brief Gets maximum count of nodes, IPs, IP neighbors, and cellnodes among cpElements.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!! Allocates global arrays 'mesh_maxNnodes', 'mesh_maxNips', mesh_maxNipNeighbors',
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!! and mesh_maxNcellnodes
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
subroutine mesh_marc_count_cpSizes(myUnit)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_intValue, &
IO_skipChunks
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 2_pInt
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
character(len=300) :: line
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt) :: i,t,g,e,c
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_maxNnodes = 0_pInt
mesh_maxNips = 0_pInt
mesh_maxNipNeighbors = 0_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_maxNcellnodes = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do
read (myUnit,610,END=630) line
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if( IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'connectivity' ) then
read (myUnit,610,END=630) line ! Garbage line
do i=1_pInt,mesh_Nelems ! read all elements
read (myUnit,610,END=630) line
myPos = IO_stringPos(line,maxNchunks) ! limit to id and type
e = mesh_FEasCP('elem',IO_intValue(line,myPos,1_pInt))
if (e /= 0_pInt) then
t = FE_mapElemtype(IO_stringValue(line,myPos,2_pInt))
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
g = FE_geomtype(t)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
c = FE_celltype(g)
mesh_maxNnodes = max(mesh_maxNnodes,FE_Nnodes(t))
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
mesh_maxNips = max(mesh_maxNips,FE_Nips(g))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_maxNipNeighbors = max(mesh_maxNipNeighbors,FE_NipNeighbors(c))
mesh_maxNcellnodes = max(mesh_maxNcellnodes,FE_Ncellnodes(g))
call IO_skipChunks(myUnit,FE_Nnodes(t)-(myPos(1_pInt)-2_pInt)) ! read on if FE_Nnodes exceeds node count present on current line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
exit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
endif
enddo
adopted improvements done by Arun Prakash. mesh: elemType identification based on lower case Abaqus now reports more errors IO: new function to inquire whether inputfile contains "parts" new function to assemble multiply included inputfile into a flat one awareness of range generation in element numbers error reporting
2010-07-13 15:56:07 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
630 end subroutine mesh_marc_count_cpSizes
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Store FEid, type, mat, tex, and node list per elemen.
!! Allocates global array 'mesh_element'
!--------------------------------------------------------------------------------------------------
subroutine mesh_marc_build_elements(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_fixedNoEFloatValue, &
IO_skipChunks, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_intValue, &
IO_continuousIntValues
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
integer(pInt), parameter :: maxNchunks = 66_pInt ! limit to 64 nodes max (plus ID, type)
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
character(len=300) line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), dimension(1_pInt+mesh_NcpElems) :: contInts
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt) :: i,j,t,sv,myVal,e,nNodesAlreadyRead
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
allocate (mesh_element(4_pInt+mesh_maxNnodes,mesh_NcpElems)) ; mesh_element = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (myUnit,610,END=620) line
myPos(1:1+2*1) = IO_stringPos(line,1_pInt)
if( IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'connectivity' ) then
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
read (myUnit,610,END=620) line ! garbage line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do i = 1_pInt,mesh_Nelems
read (myUnit,610,END=620) line
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
e = mesh_FEasCP('elem',IO_intValue(line,myPos,1_pInt))
if (e /= 0_pInt) then ! disregard non CP elems
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
t = FE_mapElemtype(IO_StringValue(line,myPos,2_pInt)) ! elem type
mesh_element(2,e) = t
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_element(1,e) = IO_IntValue (line,myPos,1_pInt) ! FE id
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
nNodesAlreadyRead = 0_pInt
do j = 1_pInt,myPos(1)-2_pInt
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_element(4_pInt+j,e) = mesh_FEasCP('node',IO_IntValue(line,myPos,j+2_pInt)) ! CP ids of nodes
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
enddo
nNodesAlreadyRead = myPos(1) - 2_pInt
do while(nNodesAlreadyRead < FE_Nnodes(t)) ! read on if not all nodes in one line
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
do j = 1_pInt,myPos(1)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_element(4_pInt+nNodesAlreadyRead+j,e) &
= mesh_FEasCP('node',IO_IntValue(line,myPos,j)) ! CP ids of nodes
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
enddo
nNodesAlreadyRead = nNodesAlreadyRead + myPos(1)
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
enddo
exit
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
620 rewind(myUnit) ! just in case "initial state" apears before "connectivity"
read (myUnit,610,END=620) line
do
myPos(1:1+2*2) = IO_stringPos(line,2_pInt)
if( (IO_lc(IO_stringValue(line,myPos,1_pInt)) == 'initial') .and. &
(IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'state') ) then
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
if (initialcondTableStyle == 2_pInt) read (myUnit,610,END=620) line ! read extra line for new style
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (myUnit,610,END=630) line ! read line with index of state var
myPos(1:1+2*1) = IO_stringPos(line,1_pInt)
sv = IO_IntValue(line,myPos,1_pInt) ! figure state variable index
if( (sv == 2_pInt).or.(sv == 3_pInt) ) then ! only state vars 2 and 3 of interest
read (myUnit,610,END=620) line ! read line with value of state var
myPos(1:1+2*1) = IO_stringPos(line,1_pInt)
do while (scan(IO_stringValue(line,myPos,1_pInt),'+-',back=.true.)>1) ! is noEfloat value?
myVal = nint(IO_fixedNoEFloatValue(line,[0_pInt,20_pInt],1_pInt),pInt) ! state var's value
mesh_maxValStateVar(sv-1_pInt) = max(myVal,mesh_maxValStateVar(sv-1_pInt)) ! remember max val of homogenization and microstructure index
if (initialcondTableStyle == 2_pInt) then
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
read (myUnit,610,END=630) line ! read extra line
read (myUnit,610,END=630) line ! read extra line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
contInts = IO_continuousIntValues& ! get affected elements
(myUnit,mesh_Nelems,mesh_nameElemSet,mesh_mapElemSet,mesh_NelemSets)
do i = 1_pInt,contInts(1)
e = mesh_FEasCP('elem',contInts(1_pInt+i))
mesh_element(1_pInt+sv,e) = myVal
enddo
if (initialcondTableStyle == 0_pInt) read (myUnit,610,END=620) line ! ignore IP range for old table style
read (myUnit,610,END=630) line
myPos(1:1+2*1) = IO_stringPos(line,1_pInt)
enddo
endif
else
read (myUnit,610,END=630) line
endif
enddo
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
630 end subroutine mesh_marc_build_elements
#endif
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#ifdef Abaqus
!--------------------------------------------------------------------------------------------------
!> @brief Count overall number of nodes and elements in mesh and stores them in
!! 'mesh_Nelems' and 'mesh_Nnodes'
!--------------------------------------------------------------------------------------------------
subroutine mesh_abaqus_count_nodesAndElements(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_countDataLines, &
IO_error
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), parameter :: maxNchunks = 2_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), dimension (1+2*maxNchunks) :: myPos
character(len=300) :: line
logical :: inPart
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
mesh_Nnodes = 0_pInt
mesh_Nelems = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
inPart = .false.
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
do
read (myUnit,610,END=620) line
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
myPos = IO_stringPos(line,maxNchunks)
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (inPart .or. noPart) then
select case ( IO_lc(IO_stringValue(line,myPos,1_pInt)))
case('*node')
if( &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'output' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'print' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'file' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'response' &
) &
mesh_Nnodes = mesh_Nnodes + IO_countDataLines(myUnit)
case('*element')
if( &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'output' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'matrix' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'response' &
) then
mesh_Nelems = mesh_Nelems + IO_countDataLines(myUnit)
endif
endselect
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
620 if (mesh_Nnodes < 2_pInt) call IO_error(error_ID=900_pInt)
if (mesh_Nelems == 0_pInt) call IO_error(error_ID=901_pInt)
end subroutine mesh_abaqus_count_nodesAndElements
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief count overall number of element sets in mesh and write 'mesh_NelemSets' and
!! 'mesh_maxNelemInSet'
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_abaqus_count_elementSets(myUnit)
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_error
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
implicit none
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), intent(in) :: myUnit
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 2_pInt
integer(pInt), dimension (1+2*maxNchunks) :: myPos
character(len=300) :: line
logical :: inPart
mesh_NelemSets = 0_pInt
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
mesh_maxNelemInSet = mesh_Nelems ! have to be conservative, since Abaqus allows for recursive definitons
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
610 FORMAT(A300)
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
inPart = .false.
rewind(myUnit)
do
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
if ( (inPart .or. noPart) .and. IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*elset' ) &
mesh_NelemSets = mesh_NelemSets + 1_pInt
enddo
corrected some problems in building mesh
2010-07-01 20:50:39 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
620 continue
if (mesh_NelemSets == 0) call IO_error(error_ID=902_pInt)
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_abaqus_count_elementSets
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! count overall number of solid sections sets in mesh (Abaqus only)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
!
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! mesh_Nmaterials
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_abaqus_count_materials(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_error
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 2_pInt
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
character(len=300) :: line
logical inPart
mesh_Nmaterials = 0_pInt
610 FORMAT(A300)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
inPart = .false.
rewind(myUnit)
do
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if ( (inPart .or. noPart) .and. &
IO_lc(IO_StringValue(line,myPos,1_pInt)) == '*solid' .and. &
IO_lc(IO_StringValue(line,myPos,2_pInt)) == 'section' ) &
mesh_Nmaterials = mesh_Nmaterials + 1_pInt
enddo
620 if (mesh_Nmaterials == 0_pInt) call IO_error(error_ID=903_pInt)
end subroutine mesh_abaqus_count_materials
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! Build element set mapping
!
! allocate globals: mesh_nameElemSet, mesh_mapElemSet
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_abaqus_map_elementSets(myUnit)
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_extractValue, &
IO_continuousIntValues, &
IO_error
implicit none
integer(pInt), intent(in) :: myUnit
integer(pInt), parameter :: maxNchunks = 4_pInt
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
character(len=300) :: line
integer(pInt) :: elemSet = 0_pInt,i
logical :: inPart = .false.
allocate (mesh_nameElemSet(mesh_NelemSets)) ; mesh_nameElemSet = ''
allocate (mesh_mapElemSet(1_pInt+mesh_maxNelemInSet,mesh_NelemSets)) ; mesh_mapElemSet = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (myUnit,610,END=640) line
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
if ( (inPart .or. noPart) .and. IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*elset' ) then
elemSet = elemSet + 1_pInt
mesh_nameElemSet(elemSet) = trim(IO_extractValue(IO_lc(IO_stringValue(line,myPos,2_pInt)),'elset'))
mesh_mapElemSet(:,elemSet) = IO_continuousIntValues(myUnit,mesh_Nelems,mesh_nameElemSet,&
mesh_mapElemSet,elemSet-1_pInt)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
640 do i = 1_pInt,elemSet
if (mesh_mapElemSet(1,i) == 0_pInt) call IO_error(error_ID=904_pInt,ext_msg=mesh_nameElemSet(i))
enddo
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_abaqus_map_elementSets
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! map solid section (Abaqus only)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
!
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! allocate globals: mesh_nameMaterial, mesh_mapMaterial
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_abaqus_map_materials(myUnit)
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_extractValue, &
IO_error
implicit none
integer(pInt), intent(in) :: myUnit
integer(pInt), parameter :: maxNchunks = 20_pInt
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
character(len=300) line
integer(pInt) :: i,c = 0_pInt
logical :: inPart = .false.
character(len=64) :: elemSetName,materialName
allocate (mesh_nameMaterial(mesh_Nmaterials)) ; mesh_nameMaterial = ''
allocate (mesh_mapMaterial(mesh_Nmaterials)) ; mesh_mapMaterial = ''
610 FORMAT(A300)
rewind(myUnit)
do
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
if ( (inPart .or. noPart) .and. &
IO_lc(IO_StringValue(line,myPos,1_pInt)) == '*solid' .and. &
IO_lc(IO_StringValue(line,myPos,2_pInt)) == 'section' ) then
elemSetName = ''
materialName = ''
do i = 3_pInt,myPos(1_pInt)
if (IO_extractValue(IO_lc(IO_stringValue(line,myPos,i)),'elset') /= '') &
elemSetName = trim(IO_extractValue(IO_lc(IO_stringValue(line,myPos,i)),'elset'))
if (IO_extractValue(IO_lc(IO_stringValue(line,myPos,i)),'material') /= '') &
materialName = trim(IO_extractValue(IO_lc(IO_stringValue(line,myPos,i)),'material'))
enddo
if (elemSetName /= '' .and. materialName /= '') then
c = c + 1_pInt
mesh_nameMaterial(c) = materialName ! name of material used for this section
mesh_mapMaterial(c) = elemSetName ! mapped to respective element set
endif
endif
enddo
620 if (c==0_pInt) call IO_error(error_ID=905_pInt)
do i=1_pInt,c
if (mesh_nameMaterial(i)=='' .or. mesh_mapMaterial(i)=='') call IO_error(error_ID=905_pInt)
enddo
end subroutine mesh_abaqus_map_materials
!--------------------------------------------------------------------------------------------------
!> @brief Count overall number of CP elements in mesh and stores them in 'mesh_NcpElems'
!--------------------------------------------------------------------------------------------------
subroutine mesh_abaqus_count_cpElements(myUnit)
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_error, &
IO_extractValue
implicit none
integer(pInt), intent(in) :: myUnit
integer(pInt), parameter :: maxNchunks = 2_pInt
integer(pInt), dimension (1+2*maxNchunks) :: myPos
character(len=300) line
integer(pInt) :: i,k
logical :: materialFound = .false.
character(len=64) ::materialName,elemSetName
mesh_NcpElems = 0_pInt
610 FORMAT(A300)
rewind(myUnit)
do
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
select case ( IO_lc(IO_stringValue(line,myPos,1_pInt)) )
case('*material')
materialName = trim(IO_extractValue(IO_lc(IO_stringValue(line,myPos,2_pInt)),'name')) ! extract name=value
materialFound = materialName /= '' ! valid name?
case('*user')
if (IO_lc(IO_StringValue(line,myPos,2_pInt)) == 'material' .and. materialFound) then
do i = 1_pInt,mesh_Nmaterials ! look thru material names
if (materialName == mesh_nameMaterial(i)) then ! found one
elemSetName = mesh_mapMaterial(i) ! take corresponding elemSet
do k = 1_pInt,mesh_NelemSets ! look thru all elemSet definitions
if (elemSetName == mesh_nameElemSet(k)) & ! matched?
mesh_NcpElems = mesh_NcpElems + mesh_mapElemSet(1,k) ! add those elem count
enddo
endif
enddo
materialFound = .false.
endif
endselect
enddo
620 if (mesh_NcpElems == 0_pInt) call IO_error(error_ID=906_pInt)
end subroutine mesh_abaqus_count_cpElements
!--------------------------------------------------------------------------------------------------
!> @brief Maps elements from FE ID to internal (consecutive) representation.
!! Allocates global array 'mesh_mapFEtoCPelem'
!--------------------------------------------------------------------------------------------------
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
subroutine mesh_abaqus_map_elements(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
changed misplaced private to public statement and renamed qsort to math_qsort
2013-02-01 12:56:21 +05:30
use math, only: math_qsort
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_extractValue, &
IO_error
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), parameter :: maxNchunks = 2_pInt
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
character(len=300) :: line
integer(pInt) ::i,j,k,cpElem = 0_pInt
logical :: materialFound = .false.
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
character (len=64) materialName,elemSetName ! why limited to 64? ABAQUS?
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
allocate (mesh_mapFEtoCPelem(2,mesh_NcpElems)) ; mesh_mapFEtoCPelem = 0_pInt
610 FORMAT(A300)
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
read (myUnit,610,END=660) line
myPos = IO_stringPos(line,maxNchunks)
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
select case ( IO_lc(IO_stringValue(line,myPos,1_pInt)) )
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
case('*material')
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
materialName = trim(IO_extractValue(IO_lc(IO_stringValue(line,myPos,2_pInt)),'name')) ! extract name=value
materialFound = materialName /= '' ! valid name?
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
case('*user')
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
if (IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'material' .and. materialFound) then
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
do i = 1_pInt,mesh_Nmaterials ! look thru material names
if (materialName == mesh_nameMaterial(i)) then ! found one
elemSetName = mesh_mapMaterial(i) ! take corresponding elemSet
do k = 1_pInt,mesh_NelemSets ! look thru all elemSet definitions
if (elemSetName == mesh_nameElemSet(k)) then ! matched?
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do j = 1_pInt,mesh_mapElemSet(1,k)
mpie_cpfem_abaqus_std.f: first call is with kinc==1 mesh: allow multiple element sets with identical material
2011-03-23 21:50:12 +05:30
cpElem = cpElem + 1_pInt
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
mesh_mapFEtoCPelem(1,cpElem) = mesh_mapElemSet(1_pInt+j,k) ! store FE id
mesh_mapFEtoCPelem(2,cpElem) = cpElem ! store our id
mpie_cpfem_abaqus_std.f: first call is with kinc==1 mesh: allow multiple element sets with identical material
2011-03-23 21:50:12 +05:30
enddo
endif
enddo
endif
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
enddo
materialFound = .false.
endif
endselect
enddo
changed misplaced private to public statement and renamed qsort to math_qsort
2013-02-01 12:56:21 +05:30
660 call math_qsort(mesh_mapFEtoCPelem,1_pInt,int(size(mesh_mapFEtoCPelem,2_pInt),pInt)) ! should be mesh_NcpElems
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
if (int(size(mesh_mapFEtoCPelem),pInt) < 2_pInt) call IO_error(error_ID=907_pInt)
did a lot of polishing: - removed unnecessary "return" before end of subroutine or function: - changed undetermined array length (:) to (1:3) To prevent problems with some code analysing tools: - "3D oneliner loops" (with ";) only for "do" and "enddo" at the same time - removed line continuation in OMP statements made the makefile more flexible, removed heap-arrays switch
2011-09-13 21:24:06 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
end subroutine mesh_abaqus_map_elements
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Maps node from FE ID to internal (consecutive) representation.
!! Allocates global array 'mesh_mapFEtoCPnode'
!--------------------------------------------------------------------------------------------------
subroutine mesh_abaqus_map_nodes(myUnit)
new errors regarding Spectral method input file parsing mesh setup from Spectral input file $Id$ in FEsolving activated
2010-05-06 22:10:47 +05:30
changed misplaced private to public statement and renamed qsort to math_qsort
2013-02-01 12:56:21 +05:30
use math, only: math_qsort
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_countDataLines, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
IO_intValue, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_error
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), parameter :: maxNchunks = 2_pInt
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
character(len=300) line
integer(pInt) :: i,c,cpNode = 0_pInt
logical :: inPart = .false.
allocate (mesh_mapFEtoCPnode(2_pInt,mesh_Nnodes)) ; mesh_mapFEtoCPnode = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (myUnit,610,END=650) line
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
myPos = IO_stringPos(line,maxNchunks)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
if( (inPart .or. noPart) .and. &
IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*node' .and. &
( IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'output' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'print' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'file' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'response' ) &
) then
c = IO_countDataLines(myUnit)
do i = 1_pInt,c
backspace(myUnit)
enddo
do i = 1_pInt,c
read (myUnit,610,END=650) line
myPos = IO_stringPos(line,maxNchunks)
cpNode = cpNode + 1_pInt
mesh_mapFEtoCPnode(1_pInt,cpNode) = IO_intValue(line,myPos,1_pInt)
mesh_mapFEtoCPnode(2_pInt,cpNode) = cpNode
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
enddo
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
changed misplaced private to public statement and renamed qsort to math_qsort
2013-02-01 12:56:21 +05:30
650 call math_qsort(mesh_mapFEtoCPnode,1_pInt,int(size(mesh_mapFEtoCPnode,2_pInt),pInt))
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (int(size(mesh_mapFEtoCPnode),pInt) == 0_pInt) call IO_error(error_ID=908_pInt)
end subroutine mesh_abaqus_map_nodes
!--------------------------------------------------------------------------------------------------
!> @brief store x,y,z coordinates of all nodes in mesh.
!! Allocates global arrays 'mesh_node0' and 'mesh_node'
!--------------------------------------------------------------------------------------------------
subroutine mesh_abaqus_build_nodes(myUnit)
replaced numerics_unitlength with mesh_unitlength
2013-05-17 22:22:19 +05:30
use IO, only: &
IO_lc, &
IO_stringValue, &
IO_floatValue, &
IO_stringPos, &
IO_error, &
IO_countDataLines, &
IO_intValue
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
implicit none
integer(pInt), intent(in) :: myUnit
integer(pInt), parameter :: maxNchunks = 4_pInt
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
character(len=300) :: line
integer(pInt) :: i,j,m,c
logical :: inPart
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate ( mesh_node0 (3,mesh_Nnodes) ); mesh_node0 = 0.0_pReal
allocate ( mesh_node (3,mesh_Nnodes) ); mesh_node = 0.0_pReal
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
610 FORMAT(A300)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
inPart = .false.
rewind(myUnit)
do
read (myUnit,610,END=670) line
myPos = IO_stringPos(line,maxNchunks)
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
if( (inPart .or. noPart) .and. &
IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*node' .and. &
( IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'output' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'print' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'file' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'response' ) &
) then
c = IO_countDataLines(myUnit) ! how many nodes are defined here?
do i = 1_pInt,c
backspace(myUnit) ! rewind to first entry
enddo
do i = 1_pInt,c
read (myUnit,610,END=670) line
myPos = IO_stringPos(line,maxNchunks)
m = mesh_FEasCP('node',IO_intValue(line,myPos,1_pInt))
added missing bracket and substituted one more forall by do loop
2013-02-09 13:53:47 +05:30
do j=1_pInt, 3_pInt
replaced numerics_unitlength with mesh_unitlength
2013-05-17 22:22:19 +05:30
mesh_node0(j,m) = mesh_unitlength * IO_floatValue(line,myPos,j+1_pInt)
added missing bracket and substituted one more forall by do loop
2013-02-09 13:53:47 +05:30
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo
endif
enddo
670 if (int(size(mesh_node0,2_pInt),pInt) /= mesh_Nnodes) call IO_error(error_ID=909_pInt)
mesh_node = mesh_node0
end subroutine mesh_abaqus_build_nodes
!--------------------------------------------------------------------------------------------------
!> @brief Gets maximum count of nodes, IPs, IP neighbors, and subNodes among cpElements.
!! Allocates global arrays 'mesh_maxNnodes', 'mesh_maxNips', mesh_maxNipNeighbors',
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!! and mesh_maxNcellnodes
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
subroutine mesh_abaqus_count_cpSizes(myUnit)
use IO, only: IO_lc, &
IO_stringValue, &
IO_stringPos, &
IO_extractValue ,&
IO_error, &
IO_countDataLines, &
IO_intValue
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), parameter :: maxNchunks = 2_pInt
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
character(len=300) :: line
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
integer(pInt) :: i,c,t,g
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
logical :: inPart
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
mesh_maxNnodes = 0_pInt
mesh_maxNips = 0_pInt
mesh_maxNipNeighbors = 0_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_maxNcellnodes = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
inPart = .false.
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
adopted improvements done by Arun Prakash. mesh: elemType identification based on lower case Abaqus now reports more errors IO: new function to inquire whether inputfile contains "parts" new function to assemble multiply included inputfile into a flat one awareness of range generation in element numbers error reporting
2010-07-13 15:56:07 +05:30
if( (inPart .or. noPart) .and. &
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*element' .and. &
( IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'output' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'matrix' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'response' ) &
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
) then
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
t = FE_mapElemtype(IO_extractValue(IO_lc(IO_stringValue(line,myPos,2_pInt)),'type')) ! remember elem type
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
g = FE_geomtype(t)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
c = FE_celltype(g)
mesh_maxNnodes = max(mesh_maxNnodes,FE_Nnodes(t))
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
mesh_maxNips = max(mesh_maxNips,FE_Nips(g))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
mesh_maxNipNeighbors = max(mesh_maxNipNeighbors,FE_NipNeighbors(c))
mesh_maxNcellnodes = max(mesh_maxNcellnodes,FE_Ncellnodes(g))
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
endif
enddo
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
620 end subroutine mesh_abaqus_count_cpSizes
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief Store FEid, type, mat, tex, and node list per elemen.
!! Allocates global array 'mesh_element'
!--------------------------------------------------------------------------------------------------
subroutine mesh_abaqus_build_elements(myUnit)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
use IO, only: IO_lc, &
IO_stringValue, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_skipChunks, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
IO_stringPos, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_intValue, &
IO_extractValue, &
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
IO_floatValue, &
IO_error, &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_countDataLines
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
implicit none
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
integer(pInt), intent(in) :: myUnit
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), parameter :: maxNchunks = 65_pInt
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
integer(pInt), dimension (1_pInt+2_pInt*maxNchunks) :: myPos
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
integer(pInt) :: i,j,k,c,e,t,homog,micro, nNodesAlreadyRead
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
logical inPart,materialFound
character (len=64) :: materialName,elemSetName
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
character(len=300) :: line
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate (mesh_element (4_pInt+mesh_maxNnodes,mesh_NcpElems)) ; mesh_element = 0_pInt
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
610 FORMAT(A300)
inPart = .false.
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
rewind(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
do
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (myUnit,610,END=620) line
myPos(1:1+2*2) = IO_stringPos(line,2_pInt)
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*part' ) inPart = .true.
if ( IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*end' .and. &
IO_lc(IO_stringValue(line,myPos,2_pInt)) == 'part' ) inPart = .false.
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
adopted improvements done by Arun Prakash. mesh: elemType identification based on lower case Abaqus now reports more errors IO: new function to inquire whether inputfile contains "parts" new function to assemble multiply included inputfile into a flat one awareness of range generation in element numbers error reporting
2010-07-13 15:56:07 +05:30
if( (inPart .or. noPart) .and. &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_lc(IO_stringValue(line,myPos,1_pInt)) == '*element' .and. &
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
( IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'output' .and. &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'matrix' .and. &
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
IO_lc(IO_stringValue(line,myPos,2_pInt)) /= 'response' ) &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
) then
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
t = FE_mapElemtype(IO_extractValue(IO_lc(IO_stringValue(line,myPos,2_pInt)),'type')) ! remember elem type
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
c = IO_countDataLines(myUnit)
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do i = 1_pInt,c
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
backspace(myUnit)
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
enddo
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do i = 1_pInt,c
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
read (myUnit,610,END=620) line
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
myPos = IO_stringPos(line,maxNchunks) ! limit to 64 nodes max
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
e = mesh_FEasCP('elem',IO_intValue(line,myPos,1_pInt))
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
if (e /= 0_pInt) then ! disregard non CP elems
mesh_element(1,e) = IO_intValue(line,myPos,1_pInt) ! FE id
mesh_element(2,e) = t ! elem type
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
nNodesAlreadyRead = 0_pInt
do j = 1_pInt,myPos(1)-1_pInt
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_element(4_pInt+j,e) = mesh_FEasCP('node',IO_intValue(line,myPos,1_pInt+j)) ! put CP ids of nodes to position 5:
now also fixed the abaqus part
2013-02-07 16:15:10 +05:30
enddo
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
nNodesAlreadyRead = myPos(1) - 1_pInt
do while(nNodesAlreadyRead < FE_Nnodes(t)) ! read on if not all nodes in one line
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
do j = 1_pInt,myPos(1)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_element(4_pInt+nNodesAlreadyRead+j,e) &
= mesh_FEasCP('node',IO_IntValue(line,myPos,j)) ! CP ids of nodes
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
enddo
nNodesAlreadyRead = nNodesAlreadyRead + myPos(1)
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
enddo
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
620 rewind(myUnit) ! just in case "*material" definitions apear before "*element"
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
materialFound = .false.
do
added 'spectralPictureMode' for reading in geometries with 1to1 mapping (e.g. EBSD pattern) polishing !incremental update, wait for commit of damask_spectral.f90 before checking out
2011-10-18 14:50:29 +05:30
read (myUnit,610,END=630) line
myPos = IO_stringPos(line,maxNchunks)
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
select case ( IO_lc(IO_StringValue(line,myPos,1_pInt)))
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
case('*material')
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
materialName = trim(IO_extractValue(IO_lc(IO_StringValue(line,myPos,2_pInt)),'name')) ! extract name=value
materialFound = materialName /= '' ! valid name?
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
case('*user')
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
if ( IO_lc(IO_StringValue(line,myPos,2_pInt)) == 'material' .and. &
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
materialFound ) then
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
read (myUnit,610,END=630) line ! read homogenization and microstructure
fixing inconsistencies in variable assignments
2012-02-10 17:26:05 +05:30
myPos(1:1+2*2) = IO_stringPos(line,2_pInt)
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
homog = nint(IO_floatValue(line,myPos,1_pInt),pInt)
micro = nint(IO_floatValue(line,myPos,2_pInt),pInt)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
do i = 1_pInt,mesh_Nmaterials ! look thru material names
if (materialName == mesh_nameMaterial(i)) then ! found one
elemSetName = mesh_mapMaterial(i) ! take corresponding elemSet
do k = 1_pInt,mesh_NelemSets ! look thru all elemSet definitions
if (elemSetName == mesh_nameElemSet(k)) then ! matched?
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do j = 1_pInt,mesh_mapElemSet(1,k)
mpie_cpfem_abaqus_std.f: first call is with kinc==1 mesh: allow multiple element sets with identical material
2011-03-23 21:50:12 +05:30
e = mesh_FEasCP('elem',mesh_mapElemSet(1+j,k))
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
mesh_element(3,e) = homog ! store homogenization
mesh_element(4,e) = micro ! store microstructure
mpie_cpfem_abaqus_std.f: first call is with kinc==1 mesh: allow multiple element sets with identical material
2011-03-23 21:50:12 +05:30
mesh_maxValStateVar(1) = max(mesh_maxValStateVar(1),homog)
mesh_maxValStateVar(2) = max(mesh_maxValStateVar(2),micro)
enddo
endif
enddo
endif
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
enddo
materialFound = .false.
endif
endselect
enddo
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
630 end subroutine mesh_abaqus_build_elements
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
#endif
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief get any additional damask options from input file, sets mesh_periodicSurface
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_get_damaskOptions(myUnit)
use IO, only: &
IO_lc, &
IO_stringValue, &
IO_stringPos
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
fixed closing of file (unit 222) statement in FEsolving.f90 added missing keyword in mesh.f90 with ifndef Spectral statement (needed for Marc and Abaqus)
2012-06-19 20:03:24 +05:30
implicit none
integer(pInt), intent(in) :: myUnit
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
fluxes now always periodic for spectral solver, no need to explicitly state this in the geom file by use of "periodic" keyword
2013-04-22 14:02:51 +05:30
#ifndef Spectral
fixed closing of file (unit 222) statement in FEsolving.f90 added missing keyword in mesh.f90 with ifndef Spectral statement (needed for Marc and Abaqus)
2012-06-19 20:03:24 +05:30
integer(pInt), parameter :: maxNchunks = 5_pInt
integer(pInt), dimension (1+2*maxNchunks) :: myPos
integer(pInt) chunk, Nchunks
character(len=300) :: line, damaskOption, v
character(len=300) :: keyword
#endif
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
fluxes now always periodic for spectral solver, no need to explicitly state this in the geom file by use of "periodic" keyword
2013-04-22 14:02:51 +05:30
#ifdef Spectral
mesh_periodicSurface = .true.
#else
mesh_periodicSurface = .false.
removed #define Marc and substituted #ifdef Marc by #ifdef Marc4DAMASK which is defined by either the submit scripts or DAMASK_marcXXXX.f90
2013-04-30 15:19:30 +05:30
#ifdef Marc4DAMASK
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
keyword = '$damask'
#endif
#ifdef Abaqus
keyword = '**damask'
#endif
fixed closing of file (unit 222) statement in FEsolving.f90 added missing keyword in mesh.f90 with ifndef Spectral statement (needed for Marc and Abaqus)
2012-06-19 20:03:24 +05:30
rewind(myUnit)
do
read (myUnit,610,END=620) line
myPos = IO_stringPos(line,maxNchunks)
Nchunks = myPos(1)
fluxes now always periodic for spectral solver, no need to explicitly state this in the geom file by use of "periodic" keyword
2013-04-22 14:02:51 +05:30
if (IO_lc(IO_stringValue(line,myPos,1_pInt)) == keyword .and. Nchunks > 1_pInt) then ! found keyword for damask option and there is at least one more chunk to read
fixed closing of file (unit 222) statement in FEsolving.f90 added missing keyword in mesh.f90 with ifndef Spectral statement (needed for Marc and Abaqus)
2012-06-19 20:03:24 +05:30
damaskOption = IO_lc(IO_stringValue(line,myPos,2_pInt))
select case(damaskOption)
fluxes now always periodic for spectral solver, no need to explicitly state this in the geom file by use of "periodic" keyword
2013-04-22 14:02:51 +05:30
case('periodic') ! damask Option that allows to specify periodic fluxes
do chunk = 3_pInt,Nchunks ! loop through chunks (skipping the keyword)
v = IO_lc(IO_stringValue(line,myPos,chunk)) ! chunk matches keyvalues x,y, or z?
fixed closing of file (unit 222) statement in FEsolving.f90 added missing keyword in mesh.f90 with ifndef Spectral statement (needed for Marc and Abaqus)
2012-06-19 20:03:24 +05:30
mesh_periodicSurface(1) = mesh_periodicSurface(1) .or. v == 'x'
mesh_periodicSurface(2) = mesh_periodicSurface(2) .or. v == 'y'
mesh_periodicSurface(3) = mesh_periodicSurface(3) .or. v == 'z'
enddo
endselect
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
fluxes now always periodic for spectral solver, no need to explicitly state this in the geom file by use of "periodic" keyword
2013-04-22 14:02:51 +05:30
610 FORMAT(A300)
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
#endif
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
620 end subroutine mesh_get_damaskOptions
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief calculation of IP interface areas, allocate globals '_ipArea', and '_ipAreaNormal'
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_build_ipAreas
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
use math, only: &
math_norm3, &
math_vectorproduct
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
implicit none
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt) :: e,t,g,c,i,f,n,m
real(pReal), dimension (3,FE_maxNcellnodesPerCellface) :: nodePos, normals
real(pReal), dimension(3) :: normal
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
moved public data res,size and homog from mesh to DAMASK_spectral_utilities (as grid and geomSize)
2013-05-08 21:22:29 +05:30
allocate(mesh_ipArea(mesh_maxNipNeighbors,mesh_maxNips,mesh_NcpElems)); mesh_ipArea = 0.0_pReal
allocate(mesh_ipAreaNormal(3_pInt,mesh_maxNipNeighbors,mesh_maxNips,mesh_NcpElems)); mesh_ipAreaNormal = 0.0_pReal
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!$OMP PARALLEL DO PRIVATE(t,g,c,nodePos,normal,normals)
do e = 1_pInt,mesh_NcpElems ! loop over cpElems
t = mesh_element(2_pInt,e) ! get element type
g = FE_geomtype(t) ! get geometry type
c = FE_celltype(g) ! get cell type
select case (c)
case (1_pInt,2_pInt) ! 2D 3 or 4 node
do i = 1_pInt,FE_Nips(g) ! loop over ips=cells in this element
do f = 1_pInt,FE_NipNeighbors(c) ! loop over cell faces
forall(n = 1_pInt:FE_NcellnodesPerCellface(c)) &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
nodePos(1:3,n) = mesh_cellnode(1:3,mesh_cell(FE_cellface(n,f,c),i,e))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
normal(1) = nodePos(2,2) - nodePos(2,1) ! x_normal = y_connectingVector
normal(2) = -(nodePos(1,2) - nodePos(1,1)) ! y_normal = -x_connectingVector
normal(3) = 0.0_pReal
mesh_ipArea(f,i,e) = math_norm3(normal)
mesh_ipAreaNormal(1:3,f,i,e) = normal / math_norm3(normal) ! ensure unit length of area normal
enddo
2d elements now have a volume corresponding to a thickness of 1; used to have zero volume now also 2d elements can be used with nonlocal model
2013-04-09 23:37:30 +05:30
enddo
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
case (3_pInt) ! 3D 4node
do i = 1_pInt,FE_Nips(g) ! loop over ips=cells in this element
do f = 1_pInt,FE_NipNeighbors(c) ! loop over cell faces
forall(n = 1_pInt:FE_NcellnodesPerCellface(c)) &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
nodePos(1:3,n) = mesh_cellnode(1:3,mesh_cell(FE_cellface(n,f,c),i,e))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
normal = math_vectorproduct(nodePos(1:3,2) - nodePos(1:3,1), &
nodePos(1:3,3) - nodePos(1:3,1))
mesh_ipArea(f,i,e) = math_norm3(normal)
mesh_ipAreaNormal(1:3,f,i,e) = normal / math_norm3(normal) ! ensure unit length of area normal
enddo
2d elements now have a volume corresponding to a thickness of 1; used to have zero volume now also 2d elements can be used with nonlocal model
2013-04-09 23:37:30 +05:30
enddo
introduced dedicated calculation of interface normals in IP neighborhood. (2D elements did not work so far..!) fixed definitions of some (2D) interface planes to have all normals pointing outward. removed "int, private :: i" defined for spectral compilation.
2013-02-05 18:57:37 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
case (4_pInt) ! 3D 8node
! for this cell type we get the normal of the quadrilateral face as an average of
! four normals of triangular subfaces; since the face consists only of two triangles,
! the sum has to be divided by two; this whole prcedure tries to compensate for
! probable non-planar cell surfaces
m = FE_NcellnodesPerCellface(c)
do i = 1_pInt,FE_Nips(g) ! loop over ips=cells in this element
do f = 1_pInt,FE_NipNeighbors(c) ! loop over cell faces
forall(n = 1_pInt:FE_NcellnodesPerCellface(c)) &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
nodePos(1:3,n) = mesh_cellnode(1:3,mesh_cell(FE_cellface(n,f,c),i,e))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
forall(n = 1_pInt:FE_NcellnodesPerCellface(c)) &
normals(1:3,n) = 0.5_pReal &
* math_vectorproduct(nodePos(1:3,1+mod(n ,m)) - nodePos(1:3,n), &
nodePos(1:3,1+mod(n+1,m)) - nodePos(1:3,n))
normal = 0.5_pReal * sum(normals,2)
mesh_ipArea(f,i,e) = math_norm3(normal)
mesh_ipAreaNormal(1:3,f,i,e) = normal / math_norm3(normal)
enddo
enddo
end select
enddo
!$OMP END PARALLEL DO
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_build_ipAreas
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief assignment of twin nodes for each cp node, allocate globals '_nodeTwins'
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_build_nodeTwins
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
implicit none
integer(pInt) dir, & ! direction of periodicity
node, &
minimumNode, &
maximumNode, &
n1, &
n2
integer(pInt), dimension(mesh_Nnodes+1) :: minimumNodes, maximumNodes ! list of surface nodes (minimum and maximum coordinate value) with first entry giving the number of nodes
real(pReal) minCoord, maxCoord, & ! extreme positions in one dimension
tolerance ! tolerance below which positions are assumed identical
real(pReal), dimension(3) :: distance ! distance between two nodes in all three coordinates
logical, dimension(mesh_Nnodes) :: unpaired
allocate(mesh_nodeTwins(3,mesh_Nnodes))
mesh_nodeTwins = 0_pInt
tolerance = 0.001_pReal * minval(mesh_ipVolume) ** 0.333_pReal
do dir = 1_pInt,3_pInt ! check periodicity in directions of x,y,z
if (mesh_periodicSurface(dir)) then ! only if periodicity is requested
!*** find out which nodes sit on the surface
!*** and have a minimum or maximum position in this dimension
minimumNodes = 0_pInt
maximumNodes = 0_pInt
minCoord = minval(mesh_node0(dir,:))
maxCoord = maxval(mesh_node0(dir,:))
do node = 1_pInt,mesh_Nnodes ! loop through all nodes and find surface nodes
if (abs(mesh_node0(dir,node) - minCoord) <= tolerance) then
minimumNodes(1) = minimumNodes(1) + 1_pInt
minimumNodes(minimumNodes(1)+1_pInt) = node
elseif (abs(mesh_node0(dir,node) - maxCoord) <= tolerance) then
maximumNodes(1) = maximumNodes(1) + 1_pInt
maximumNodes(maximumNodes(1)+1_pInt) = node
endif
enddo
!*** find the corresponding node on the other side with the same position in this dimension
unpaired = .true.
do n1 = 1_pInt,minimumNodes(1)
minimumNode = minimumNodes(n1+1_pInt)
if (unpaired(minimumNode)) then
do n2 = 1_pInt,maximumNodes(1)
maximumNode = maximumNodes(n2+1_pInt)
distance = abs(mesh_node0(:,minimumNode) - mesh_node0(:,maximumNode))
if (sum(distance) - distance(dir) <= tolerance) then ! minimum possible distance (within tolerance)
mesh_nodeTwins(dir,minimumNode) = maximumNode
mesh_nodeTwins(dir,maximumNode) = minimumNode
unpaired(maximumNode) = .false. ! remember this node, we don't have to look for his partner again
exit
endif
enddo
endif
enddo
endif
enddo
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_build_nodeTwins
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief get maximum count of shared elements among cpElements and build list of elements shared
!! by each node in mesh. Allocate globals '_maxNsharedElems' and '_sharedElem'
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_build_sharedElems
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
implicit none
integer(pint) e, & ! element index
g, & ! element type
node, & ! CP node index
n, & ! node index per element
myDim, & ! dimension index
nodeTwin ! node twin in the specified dimension
integer(pInt), dimension (mesh_Nnodes) :: node_count
integer(pInt), dimension (:), allocatable :: node_seen
allocate(node_seen(maxval(FE_NmatchingNodes)))
node_count = 0_pInt
do e = 1_pInt,mesh_NcpElems
g = FE_geomtype(mesh_element(2,e)) ! get elemGeomType
node_seen = 0_pInt ! reset node duplicates
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
do n = 1_pInt,FE_NmatchingNodes(g) ! check each node of element
node = mesh_element(4+n,e)
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
if (all(node_seen /= node)) then
node_count(node) = node_count(node) + 1_pInt ! if FE node not yet encountered -> count it
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
do myDim = 1_pInt,3_pInt ! check in each dimension...
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
nodeTwin = mesh_nodeTwins(myDim,node)
if (nodeTwin > 0_pInt) & ! if I am a twin of some node...
node_count(nodeTwin) = node_count(nodeTwin) + 1_pInt ! -> count me again for the twin node
enddo
endif
node_seen(n) = node ! remember this node to be counted already
enddo
enddo
mesh_maxNsharedElems = int(maxval(node_count),pInt) ! most shared node
allocate(mesh_sharedElem(1+mesh_maxNsharedElems,mesh_Nnodes))
mesh_sharedElem = 0_pInt
do e = 1_pInt,mesh_NcpElems
g = FE_geomtype(mesh_element(2,e)) ! get elemGeomType
node_seen = 0_pInt
do n = 1_pInt,FE_NmatchingNodes(g)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
node = mesh_element(4_pInt+n,e)
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
if (all(node_seen /= node)) then
mesh_sharedElem(1,node) = mesh_sharedElem(1,node) + 1_pInt ! count for each node the connected elements
mesh_sharedElem(mesh_sharedElem(1,node)+1_pInt,node) = e ! store the respective element id
do myDim = 1_pInt,3_pInt ! check in each dimension...
nodeTwin = mesh_nodeTwins(myDim,node)
if (nodeTwin > 0_pInt) then ! if i am a twin of some node...
mesh_sharedElem(1,nodeTwin) = mesh_sharedElem(1,nodeTwin) + 1_pInt ! ...count me again for the twin
mesh_sharedElem(mesh_sharedElem(1,nodeTwin)+1,nodeTwin) = e ! store the respective element id
endif
enddo
endif
node_seen(n) = node
enddo
enddo
deallocate(node_seen)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
end subroutine mesh_build_sharedElems
# new interface for Abaqus # IO has some additional functionality for Abaqus parsing # ping pong scheme in FE interface now similar (and more human understandable) in both versions # mesh has better splitting of different tasks, plus operation on database whenever possible # FEsolver as new global var to indicate FEM solver type # computation mode reshuffling: 6 is now Marc special case of recycling...
2009-10-12 21:31:49 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief build up of IP neighborhood, allocate globals '_ipNeighborhood'
!--------------------------------------------------------------------------------------------------
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
subroutine mesh_build_ipNeighborhood
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
use math, only: &
math_mul3x3
implicit none
integer(pInt) :: myElem, & ! my CP element index
myIP, &
myType, & ! my element type
myFace, &
neighbor, & ! neighor index
neighboringIPkey, & ! positive integer indicating the neighboring IP (for intra-element) and negative integer indicating the face towards neighbor (for neighboring element)
candidateIP, &
neighboringType, & ! element type of neighbor
NlinkedNodes, & ! number of linked nodes
twin_of_linkedNode, & ! node twin of a specific linkedNode
NmatchingNodes, & ! number of matching nodes
dir, & ! direction of periodicity
matchingElem, & ! CP elem number of matching element
matchingFace, & ! face ID of matching element
a, anchor, &
neighboringIP, &
neighboringElem, &
pointingToMe
integer(pInt), dimension(FE_maxmaxNnodesAtIP) :: &
linkedNodes = 0_pInt, &
matchingNodes
logical checkTwins
allocate(mesh_ipNeighborhood(3,mesh_maxNipNeighbors,mesh_maxNips,mesh_NcpElems))
mesh_ipNeighborhood = 0_pInt
do myElem = 1_pInt,mesh_NcpElems ! loop over cpElems
myType = FE_geomtype(mesh_element(2,myElem)) ! get elemGeomType
do myIP = 1_pInt,FE_Nips(myType) ! loop over IPs of elem
do neighbor = 1_pInt,FE_NipNeighbors(FE_celltype(myType)) ! loop over neighbors of IP
neighboringIPkey = FE_ipNeighbor(neighbor,myIP,myType)
!*** if the key is positive, the neighbor is inside the element
!*** that means, we have already found our neighboring IP
if (neighboringIPkey > 0_pInt) then
mesh_ipNeighborhood(1,neighbor,myIP,myElem) = myElem
mesh_ipNeighborhood(2,neighbor,myIP,myElem) = neighboringIPkey
!*** if the key is negative, the neighbor resides in a neighboring element
!*** that means, we have to look through the face indicated by the key and see which element is behind that face
elseif (neighboringIPkey < 0_pInt) then ! neighboring element's IP
myFace = -neighboringIPkey
call mesh_faceMatch(myElem, myFace, matchingElem, matchingFace) ! get face and CP elem id of face match
if (matchingElem > 0_pInt) then ! found match?
neighboringType = FE_geomtype(mesh_element(2,matchingElem))
!*** trivial solution if neighbor has only one IP
if (FE_Nips(neighboringType) == 1_pInt) then
mesh_ipNeighborhood(1,neighbor,myIP,myElem) = matchingElem
mesh_ipNeighborhood(2,neighbor,myIP,myElem) = 1_pInt
cycle
endif
!*** find those nodes which build the link to the neighbor
NlinkedNodes = 0_pInt
linkedNodes = 0_pInt
do a = 1_pInt,FE_maxNnodesAtIP(myType) ! figure my anchor nodes on connecting face
anchor = FE_nodesAtIP(a,myIP,myType)
if (anchor /= 0_pInt) then ! valid anchor node
if (any(FE_face(:,myFace,myType) == anchor)) then ! ip anchor sits on face?
NlinkedNodes = NlinkedNodes + 1_pInt
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
linkedNodes(NlinkedNodes) = mesh_element(4_pInt+anchor,myElem) ! CP id of anchor node
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
else ! something went wrong with the linkage, since not all anchors sit on my face
NlinkedNodes = 0_pInt
linkedNodes = 0_pInt
exit
endif
endif
enddo
!*** loop through the ips of my neighbor
!*** and try to find an ip with matching nodes
!*** also try to match with node twins
checkCandidateIP: do candidateIP = 1_pInt,FE_Nips(neighboringType)
NmatchingNodes = 0_pInt
matchingNodes = 0_pInt
do a = 1_pInt,FE_maxNnodesAtIP(neighboringType) ! check each anchor node of that ip
anchor = FE_nodesAtIP(a,candidateIP,neighboringType)
if (anchor /= 0_pInt) then ! valid anchor node
if (any(FE_face(:,matchingFace,neighboringType) == anchor)) then ! sits on matching face?
NmatchingNodes = NmatchingNodes + 1_pInt
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
matchingNodes(NmatchingNodes) = mesh_element(4+anchor,matchingElem) ! CP id of neighbor's anchor node
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
else ! no matching, because not all nodes sit on the matching face
NmatchingNodes = 0_pInt
matchingNodes = 0_pInt
exit
endif
endif
enddo
if (NmatchingNodes /= NlinkedNodes) & ! this ip has wrong count of anchors on face
cycle checkCandidateIP
!*** check "normal" nodes whether they match or not
checkTwins = .false.
do a = 1_pInt,NlinkedNodes
if (all(matchingNodes /= linkedNodes(a))) then ! this linkedNode does not match any matchingNode
checkTwins = .true.
exit ! no need to search further
endif
enddo
!*** if no match found, then also check node twins
if(checkTwins) then
dir = int(maxloc(abs(mesh_ipAreaNormal(1:3,neighbor,myIP,myElem)),1),pInt) ! check for twins only in direction of the surface normal
do a = 1_pInt,NlinkedNodes
twin_of_linkedNode = mesh_nodeTwins(dir,linkedNodes(a))
if (twin_of_linkedNode == 0_pInt .or. & ! twin of linkedNode does not exist...
all(matchingNodes /= twin_of_linkedNode)) then ! ... or it does not match any matchingNode
cycle checkCandidateIP ! ... then check next candidateIP
endif
enddo
endif
!*** we found a match !!!
mesh_ipNeighborhood(1,neighbor,myIP,myElem) = matchingElem
mesh_ipNeighborhood(2,neighbor,myIP,myElem) = candidateIP
exit checkCandidateIP
enddo checkCandidateIP
endif ! end of valid external matching
endif ! end of internal/external matching
enddo
enddo
enddo
do myElem = 1_pInt,mesh_NcpElems ! loop over cpElems
myType = FE_geomtype(mesh_element(2,myElem)) ! get elemGeomType
do myIP = 1_pInt,FE_Nips(myType) ! loop over IPs of elem
do neighbor = 1_pInt,FE_NipNeighbors(FE_celltype(myType)) ! loop over neighbors of IP
neighboringElem = mesh_ipNeighborhood(1,neighbor,myIP,myElem)
neighboringIP = mesh_ipNeighborhood(2,neighbor,myIP,myElem)
if (neighboringElem > 0_pInt .and. neighboringIP > 0_pInt) then ! if neighbor exists ...
neighboringType = FE_geomtype(mesh_element(2,neighboringElem))
do pointingToMe = 1_pInt,FE_NipNeighbors(FE_celltype(neighboringType)) ! find neighboring index that points from my neighbor to myself
if ( myElem == mesh_ipNeighborhood(1,pointingToMe,neighboringIP,neighboringElem) &
.and. myIP == mesh_ipNeighborhood(2,pointingToMe,neighboringIP,neighboringElem)) then ! possible candidate
if (math_mul3x3(mesh_ipAreaNormal(1:3,neighbor,myIP,myElem),&
mesh_ipAreaNormal(1:3,pointingToMe,neighboringIP,neighboringElem)) < 0.0_pReal) then ! area normals have opposite orientation (we have to check that because of special case for single element with two ips and periodicity. In this case the neighbor is identical in two different directions.)
mesh_ipNeighborhood(3,neighbor,myIP,myElem) = pointingToMe ! found match
exit ! so no need to search further
endif
endif
enddo
endif
enddo
enddo
enddo
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
end subroutine mesh_build_ipNeighborhood
major debugging and streamlining of the calculation of ipVolumes and ipAreas
2009-01-20 00:12:31 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief write statistics regarding input file parsing to the output file
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_tell_statistics
fixed small bug when calculating mesh_Ncells
2013-04-19 18:11:06 +05:30
use math, only: &
math_range
use IO, only: &
IO_error
use debug, only: &
debug_level, &
debug_MESH, &
debug_LEVELBASIC, &
debug_LEVELEXTENSIVE, &
debug_LEVELSELECTIVE, &
debug_e, &
debug_i
Calculation of current ip volumes now working. Crystallite output also reflects current grain volume, not reference volume. However, this is only available for Marc. Abaqus and spectral method still return the reference ip volume. The ip coordinates though are correctly updated for all solver types.
2011-08-10 22:07:17 +05:30
implicit none
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt), dimension (:,:), allocatable :: mesh_HomogMicro
character(len=64) :: myFmt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt) :: i,e,n,f,t,g,c, myDebug
added full description of element type 2
2009-01-16 23:06:37 +05:30
changed variable name "debug_what" to "debug_level"
2012-07-05 15:24:50 +05:30
myDebug = debug_level(debug_mesh)
Calculation of current ip volumes now working. Crystallite output also reflects current grain volume, not reference volume. However, this is only available for Marc. Abaqus and spectral method still return the reference ip volume. The ip coordinates though are correctly updated for all solver types.
2011-08-10 22:07:17 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (mesh_maxValStateVar(1) < 1_pInt) call IO_error(error_ID=170_pInt) ! no homogenization specified
if (mesh_maxValStateVar(2) < 1_pInt) call IO_error(error_ID=180_pInt) ! no microstructure specified
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
allocate (mesh_HomogMicro(mesh_maxValStateVar(1),mesh_maxValStateVar(2))); mesh_HomogMicro = 0_pInt
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do e = 1_pInt,mesh_NcpElems
reordered (and partly redistributed) error message identifiers, deleted those which are not in use anymore. all constitutive as well numerics now raises an error, if an unknown keyword is found in the respective config file
2012-02-13 23:11:27 +05:30
if (mesh_element(3,e) < 1_pInt) call IO_error(error_ID=170_pInt,e=e) ! no homogenization specified
if (mesh_element(4,e) < 1_pInt) call IO_error(error_ID=180_pInt,e=e) ! no microstructure specified
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
mesh_HomogMicro(mesh_element(3,e),mesh_element(4,e)) = &
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
mesh_HomogMicro(mesh_element(3,e),mesh_element(4,e)) + 1_pInt ! count combinations of homogenization and microstructure
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
enddo
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
!$OMP CRITICAL (write2out)
if (iand(myDebug,debug_levelBasic) /= 0_pInt) then
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
write(6,'(/,a,/)') ' Input Parser: STATISTICS'
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,*) mesh_Nelems, ' : total number of elements in mesh'
write(6,*) mesh_NcpElems, ' : total number of CP elements in mesh'
write(6,*) mesh_Nnodes, ' : total number of nodes in mesh'
write(6,*) mesh_maxNnodes, ' : max number of nodes in any CP element'
write(6,*) mesh_maxNips, ' : max number of IPs in any CP element'
write(6,*) mesh_maxNipNeighbors, ' : max number of IP neighbors in any CP element'
write(6,*) mesh_maxNsharedElems, ' : max number of CP elements sharing a node'
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
write(6,'(/,a,/)') ' Input Parser: HOMOGENIZATION/MICROSTRUCTURE'
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,*) mesh_maxValStateVar(1), ' : maximum homogenization index'
write(6,*) mesh_maxValStateVar(2), ' : maximum microstructure index'
write(6,*)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
write (myFmt,'(a,i32.32,a)') '(9x,a2,1x,',mesh_maxValStateVar(2),'(i8))'
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,myFmt) '+-',math_range(mesh_maxValStateVar(2))
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
write (myFmt,'(a,i32.32,a)') '(i8,1x,a2,1x,',mesh_maxValStateVar(2),'(i8))'
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do i=1_pInt,mesh_maxValStateVar(1) ! loop over all (possibly assigned) homogenizations
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,myFmt) i,'| ',mesh_HomogMicro(i,:) ! loop over all (possibly assigned) microstructures
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
enddo
changed all remaining routines to fortran-fast arrays (geometry reconstruction etc.) changed all remaining routines in f2py to more clever determination of array size (requires f2py >= 2.0) enabled 3D visualize to work with odd resolution by switching to linear reconstruction PLEASE NOTE: Redefinition of routines for f2py might cause trouble -> DELETE DAMASK_ROOT/lib/damask/core.so in this case further changes: added pure statement where possible, polished, unified use of "Q" for "Quaternion" and reordered math to have similar routines together
2013-01-31 21:58:08 +05:30
write(6,'(/,a,/)') ' Input Parser: ADDITIONAL MPIE OPTIONS'
made code standard conform to Fortran 2008 (ignoring warning concerning comments beyond character 132). Basically, changing "x" format specifier to "Nx" ("1x") plus removing $ format specifier added compiler switches for gfortran and ifort to check for standard conformity old gnu compilers <4.4 are not longer supported because they don't provide the c binding for fftw
2012-02-01 00:48:55 +05:30
write(6,*) 'periodic surface : ', mesh_periodicSurface
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
write(6,*)
gfortran now detects "call flush" statements, removed the last of them in mesh.f90
2013-01-24 03:20:01 +05:30
flush(6)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
endif
debugging output is now controlled by the "verbosity" parameter in the debug.config ranging from 0 (=almost no output) to 8 (=very detailed output) 0 : only version infos and all from "hypela2"/"umat" 1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info 2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines 3 : basic outputs from "homogenization.f90" 4 : extensive outputs from "homogenization.f90" 5 : basic outputs from "crystallite.f90" 6 : extensive outputs from "crystallite.f90" 7 : basic outputs from the constitutive files 8 : extensive outputs from the constitutive files If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
2011-03-21 16:01:17 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
if (iand(myDebug,debug_levelExtensive) /= 0_pInt) then
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,*)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
write(6,*) 'Input Parser: ELEMENT TYPE'
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,*)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
write(6,'(a8,3(1x,a8))') 'elem','elemtype','geomtype','celltype'
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do e = 1_pInt,mesh_NcpElems
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_e /= e) cycle
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
t = mesh_element(2,e) ! get elemType
g = FE_geomtype(t) ! get elemGeomType
c = FE_celltype(g) ! get cellType
write(6,'(i8,3(1x,i8))') e,t,g,c
enddo
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,*)
write(6,*) 'Input Parser: ELEMENT VOLUME'
write(6,*)
write(6,'(a13,1x,e15.8)') 'total volume', sum(mesh_ipVolume)
write(6,*)
write(6,'(a8,1x,a5,1x,a15,1x,a5,1x,a15,1x,a16)') 'elem','IP','volume','face','area','-- normal --'
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do e = 1_pInt,mesh_NcpElems
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_e /= e) cycle
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
t = mesh_element(2,e) ! get element type
g = FE_geomtype(t) ! get geometry type
c = FE_celltype(g) ! get cell type
do i = 1_pInt,FE_Nips(g)
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_i /= i) cycle
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,'(i8,1x,i5,1x,e15.8)') e,i,mesh_IPvolume(i,e)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
do f = 1_pInt,FE_NipNeighbors(c)
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,'(i33,1x,e15.8,1x,3(f6.3,1x))') f,mesh_ipArea(f,i,e),mesh_ipAreaNormal(:,f,i,e)
corrected bug in debug: stressloop info was erroneous mesh init now showing much more ip statistics in verbose mode
2010-09-07 14:36:02 +05:30
enddo
enddo
enddo
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,*)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
write(6,*) 'Input Parser: CELLNODE COORDINATES'
write(6,*)
write(6,'(a8,1x,a2,1x,a8,3(1x,a12))') 'elem','IP','cellnode','x','y','z'
do e = 1_pInt,mesh_NcpElems ! loop over cpElems
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_e /= e) cycle
t = mesh_element(2,e) ! get element type
g = FE_geomtype(t) ! get geometry type
c = FE_celltype(g) ! get cell type
do i = 1_pInt,FE_Nips(g) ! loop over IPs of elem
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_i /= i) cycle
write(6,'(i8,1x,i2)') e,i
do n = 1_pInt,FE_NcellnodesPerCell(c) ! loop over cell nodes in the cell
write(6,'(12x,i8,3(1x,f12.8))') mesh_cell(n,i,e), &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
mesh_cellnode(1:3,mesh_cell(n,i,e))
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
enddo
enddo
enddo
write(6,*)
write(6,*) 'Input Parser: IP COORDINATES'
write(6,'(a8,1x,a5,3(1x,a12))') 'elem','IP','x','y','z'
do e = 1_pInt,mesh_NcpElems
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_e /= e) cycle
do i = 1_pInt,FE_Nips(FE_geomtype(mesh_element(2,e)))
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_i /= i) cycle
write(6,'(i8,1x,i5,3(1x,f12.8))') e, i, mesh_ipCoordinates(:,i,e)
enddo
enddo
write(6,*)
ip neighborhood was screwed up for spectral solver as of rev 2332 due to new "mesh_spectral_ipNeighborhood": used to have fastest index in z direction although it should be in x direction! also removed output of node twins in "mesh_tell_statistics" when using spectral solver, since not allocated
2013-05-07 23:20:58 +05:30
#ifndef Spectral
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,*) 'Input Parser: NODE TWINS'
write(6,*)
made code standard conform to Fortran 2008 (ignoring warning concerning comments beyond character 132). Basically, changing "x" format specifier to "Nx" ("1x") plus removing $ format specifier added compiler switches for gfortran and ifort to check for standard conformity old gnu compilers <4.4 are not longer supported because they don't provide the c binding for fftw
2012-02-01 00:48:55 +05:30
write(6,'(a6,3(3x,a6))') ' node','twin_x','twin_y','twin_z'
polishing: removed variable names like 'unit' and 'data' that are keywords of fortran and ensured that integer and real precision matches independent of machine standard. removed cut_off parameter for damask_spectral removed outpot of derived divergence measures and added RMS output in brackets added comments and options to the makefile
2012-02-16 00:28:38 +05:30
do n = 1_pInt,mesh_Nnodes ! loop over cpNodes
corrected debugging output of node twins
2013-03-27 15:55:37 +05:30
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. .not. any(mesh_element(5:,debug_e) == n)) cycle
write(6,'(i6,3(3x,i6))') n, mesh_nodeTwins(1:3,n)
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
enddo
write(6,*)
ip neighborhood was screwed up for spectral solver as of rev 2332 due to new "mesh_spectral_ipNeighborhood": used to have fastest index in z direction although it should be in x direction! also removed output of node twins in "mesh_tell_statistics" when using spectral solver, since not allocated
2013-05-07 23:20:58 +05:30
#endif
made code standard conform to Fortran 2008 (ignoring warning concerning comments beyond character 132). Basically, changing "x" format specifier to "Nx" ("1x") plus removing $ format specifier added compiler switches for gfortran and ifort to check for standard conformity old gnu compilers <4.4 are not longer supported because they don't provide the c binding for fftw
2012-02-01 00:48:55 +05:30
write(6,*) 'Input Parser: IP NEIGHBORHOOD'
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
write(6,*)
made code standard conform to Fortran 2008 (ignoring warning concerning comments beyond character 132). Basically, changing "x" format specifier to "Nx" ("1x") plus removing $ format specifier added compiler switches for gfortran and ifort to check for standard conformity old gnu compilers <4.4 are not longer supported because they don't provide the c binding for fftw
2012-02-01 00:48:55 +05:30
write(6,'(a8,1x,a10,1x,a10,1x,a3,1x,a13,1x,a13)') 'elem','IP','neighbor','','elemNeighbor','ipNeighbor'
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
do e = 1_pInt,mesh_NcpElems ! loop over cpElems
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_e /= e) cycle
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
t = mesh_element(2,e) ! get element type
g = FE_geomtype(t) ! get geometry type
c = FE_celltype(g) ! get cell type
do i = 1_pInt,FE_Nips(g) ! loop over IPs of elem
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
if (iand(myDebug,debug_levelSelective) /= 0_pInt .and. debug_i /= i) cycle
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
do n = 1_pInt,FE_NipNeighbors(c) ! loop over neighbors of IP
removed coordinates from call to CPFEM_general and introduced direct storage of coordinates to mesh_ipCoordinates
2012-11-14 20:08:10 +05:30
write(6,'(i8,1x,i10,1x,i10,1x,a3,1x,i13,1x,i13)') e,i,n,'-->',mesh_ipNeighborhood(1,n,i,e),mesh_ipNeighborhood(2,n,i,e)
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
enddo
enddo
enddo
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
endif
!$OMP END CRITICAL (write2out)
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
deallocate(mesh_HomogMicro)
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
added new, flexible debugging scheme. now all modules have their own debug specification. compiles and runs, I hope nothing is broken did a lot of polishing
2012-03-09 01:55:28 +05:30
end subroutine mesh_tell_statistics
Now able to have real periodicity for fluxes (fluxes leaving the model on one side enter on the other side). To enable this feature one has to add the following somewhere in the marc input file: $mpie periodic x y z for having periodicity in all directions $mpie periodic z x for having periodicity in x and z direction etc. Note that this only works for regular meshes!!!
2011-02-16 21:53:08 +05:30
improved regridding, especially interfacing to python (also for nearest neighbor search)
2012-04-24 22:32:27 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief mapping of FE element types to internal representation
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
integer(pInt) function FE_mapElemtype(what)
- added element type 54 for macro (2d 8node 4ip reduced integration) - automatically raising error for unknown element type inside mapping function
2013-04-10 15:08:40 +05:30
use IO, only: IO_lc, IO_error
improved regridding, especially interfacing to python (also for nearest neighbor search)
2012-04-24 22:32:27 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
implicit none
character(len=*), intent(in) :: what
select case (IO_lc(what))
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
case ( '6')
FE_mapElemtype = 1_pInt ! Two-dimensional Plane Strain Triangle
case ( '155', &
'125', &
'128')
FE_mapElemtype = 2_pInt ! Two-dimensional Plane Strain triangle (155: cubic shape function, 125/128: second order isoparametric)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case ( '11', &
'cpe4')
FE_mapElemtype = 3_pInt ! Arbitrary Quadrilateral Plane-strain
case ( '27', &
'cpe8')
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
FE_mapElemtype = 4_pInt ! Plane Strain, Eight-node Distorted Quadrilateral
- added element type 54 for macro (2d 8node 4ip reduced integration) - automatically raising error for unknown element type inside mapping function
2013-04-10 15:08:40 +05:30
case ( '54')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 5_pInt ! Plane Strain, Eight-node Distorted Quadrilateral with reduced integration
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
case ('134', &
'c3d4')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 6_pInt ! Three-dimensional Four-node Tetrahedron
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case ('157')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 7_pInt ! Three-dimensional, Low-order, Tetrahedron, Herrmann Formulations
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
case ('127')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 8_pInt ! Three-dimensional Ten-node Tetrahedron
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case ('136', &
'c3d6')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 9_pInt ! Three-dimensional Arbitrarily Distorted Pentahedral
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case ( '117', &
'123', &
'c3d8r')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 10_pInt ! Three-dimensional Arbitrarily Distorted linear hexahedral with reduced integration
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
case ( '7', &
'c3d8')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 11_pInt ! Three-dimensional Arbitrarily Distorted Brick
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case ( '57', &
'c3d20r')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 12_pInt ! Three-dimensional Arbitrarily Distorted quad hexahedral with reduced integration
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
case ( '21', &
'c3d20')
fixed bug that was introduced in the element type mapping in version 2305
2013-04-10 18:52:02 +05:30
FE_mapElemtype = 13_pInt ! Three-dimensional Arbitrarily Distorted quadratic hexahedral
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
case default
- added element type 54 for macro (2d 8node 4ip reduced integration) - automatically raising error for unknown element type inside mapping function
2013-04-10 15:08:40 +05:30
call IO_error(error_ID=190_pInt,ext_msg=IO_lc(what))
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
end select
improved regridding, especially interfacing to python (also for nearest neighbor search)
2012-04-24 22:32:27 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end function FE_mapElemtype
regridding is now working, changed the subroutine into a function changed order of arrays in nearest neighbor search to make it fortran fast constitutive.f90 and homogenization.f90 write state size out during initialization setup/setup_processing.py is using byterecl to be compatible with binary files written out by solver
2012-05-08 20:27:06 +05:30
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief find face-matching element of same type
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_faceMatch(elem, face ,matchingElem, matchingFace)
regridding is now working, changed the subroutine into a function changed order of arrays in nearest neighbor search to make it fortran fast constitutive.f90 and homogenization.f90 write state size out during initialization setup/setup_processing.py is using byterecl to be compatible with binary files written out by solver
2012-05-08 20:27:06 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
implicit none
!*** output variables
integer(pInt), intent(out) :: matchingElem, & ! matching CP element ID
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
matchingFace ! matching face ID
reworked neares neigbor search subroutine, now also works for voronoi tesselation (general nn search in periodic cell)
2012-05-04 18:37:37 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!*** input variables
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
integer(pInt), intent(in) :: face, & ! face ID
elem ! CP elem ID
improved regridding, especially interfacing to python (also for nearest neighbor search)
2012-04-24 22:32:27 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
!*** local variables
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
integer(pInt), dimension(FE_NmatchingNodesPerFace(face,FE_geomtype(mesh_element(2,elem)))) :: &
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
myFaceNodes ! global node ids on my face
integer(pInt) :: myType, &
candidateType, &
candidateElem, &
candidateFace, &
candidateFaceNode, &
minNsharedElems, &
NsharedElems, &
lonelyNode = 0_pInt, &
i, &
n, &
dir ! periodicity direction
integer(pInt), dimension(:), allocatable :: element_seen
logical checkTwins
improved regridding, especially interfacing to python (also for nearest neighbor search)
2012-04-24 22:32:27 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
matchingElem = 0_pInt
matchingFace = 0_pInt
minNsharedElems = mesh_maxNsharedElems + 1_pInt ! init to worst case
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
myType = FE_geomtype(mesh_element(2_pInt,elem)) ! figure elemGeomType
regridding is now working, changed the subroutine into a function changed order of arrays in nearest neighbor search to make it fortran fast constitutive.f90 and homogenization.f90 write state size out during initialization setup/setup_processing.py is using byterecl to be compatible with binary files written out by solver
2012-05-08 20:27:06 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
do n = 1_pInt,FE_NmatchingNodesPerFace(face,myType) ! loop over nodes on face
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
myFaceNodes(n) = mesh_element(4_pInt+FE_face(n,face,myType),elem) ! CP id of face node
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
NsharedElems = mesh_sharedElem(1_pInt,myFaceNodes(n)) ! figure # shared elements for this node
if (NsharedElems < minNsharedElems) then
minNsharedElems = NsharedElems ! remember min # shared elems
lonelyNode = n ! remember most lonely node
endif
enddo
regridding is now working, changed the subroutine into a function changed order of arrays in nearest neighbor search to make it fortran fast constitutive.f90 and homogenization.f90 write state size out during initialization setup/setup_processing.py is using byterecl to be compatible with binary files written out by solver
2012-05-08 20:27:06 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
allocate(element_seen(minNsharedElems))
element_seen = 0_pInt
regridding is now working, changed the subroutine into a function changed order of arrays in nearest neighbor search to make it fortran fast constitutive.f90 and homogenization.f90 write state size out during initialization setup/setup_processing.py is using byterecl to be compatible with binary files written out by solver
2012-05-08 20:27:06 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
checkCandidate: do i = 1_pInt,minNsharedElems ! iterate over lonelyNode's shared elements
candidateElem = mesh_sharedElem(1_pInt+i,myFaceNodes(lonelyNode)) ! present candidate elem
if (all(element_seen /= candidateElem)) then ! element seen for the first time?
element_seen(i) = candidateElem
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
candidateType = FE_geomtype(mesh_element(2_pInt,candidateElem)) ! figure elemGeomType of candidate
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
checkCandidateFace: do candidateFace = 1_pInt,FE_maxNipNeighbors ! check each face of candidate
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
if (FE_NmatchingNodesPerFace(candidateFace,candidateType) &
/= FE_NmatchingNodesPerFace(face,myType) & ! incompatible face
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
.or. (candidateElem == elem .and. candidateFace == face)) then ! this is my face
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
cycle checkCandidateFace
endif
checkTwins = .false.
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
do n = 1_pInt,FE_NmatchingNodesPerFace(candidateFace,candidateType) ! loop through nodes on face
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
candidateFaceNode = mesh_element(4_pInt+FE_face(n,candidateFace,candidateType),candidateElem)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
if (all(myFaceNodes /= candidateFaceNode)) then ! candidate node does not match any of my face nodes
checkTwins = .true. ! perhaps the twin nodes do match
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
exit
endif
enddo
if(checkTwins) then
checkCandidateFaceTwins: do dir = 1_pInt,3_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
do n = 1_pInt,FE_NmatchingNodesPerFace(candidateFace,candidateType) ! loop through nodes on face
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
candidateFaceNode = mesh_element(4+FE_face(n,candidateFace,candidateType),candidateElem)
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
if (all(myFaceNodes /= mesh_nodeTwins(dir,candidateFaceNode))) then ! node twin does not match either
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
if (dir == 3_pInt) then
cycle checkCandidateFace
else
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
cycle checkCandidateFaceTwins ! try twins in next dimension
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
endif
endif
enddo
exit checkCandidateFaceTwins
enddo checkCandidateFaceTwins
endif
matchingFace = candidateFace
matchingElem = candidateElem
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
exit checkCandidate ! found my matching candidate
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
enddo checkCandidateFace
endif
enddo checkCandidate
regridding is now working, changed the subroutine into a function changed order of arrays in nearest neighbor search to make it fortran fast constitutive.f90 and homogenization.f90 write state size out during initialization setup/setup_processing.py is using byterecl to be compatible with binary files written out by solver
2012-05-08 20:27:06 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
deallocate(element_seen)
regridding is now working, changed the subroutine into a function changed order of arrays in nearest neighbor search to make it fortran fast constitutive.f90 and homogenization.f90 write state size out during initialization setup/setup_processing.py is using byterecl to be compatible with binary files written out by solver
2012-05-08 20:27:06 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_faceMatch
regridding is now working, changed the subroutine into a function changed order of arrays in nearest neighbor search to make it fortran fast constitutive.f90 and homogenization.f90 write state size out during initialization setup/setup_processing.py is using byterecl to be compatible with binary files written out by solver
2012-05-08 20:27:06 +05:30
integrated subroutine for regridding into mesh.f90 and made it available for python
2012-04-10 20:45:46 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief get properties of different types of finite elements
!> @details assign globals: FE_nodesAtIP, FE_ipNeighbor, FE_cellnodeParentnodeWeights, FE_subNodeOnIPFace
!--------------------------------------------------------------------------------------------------
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
subroutine mesh_build_FEdata
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
implicit none
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
integer(pInt) :: me
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
allocate(FE_nodesAtIP(FE_maxmaxNnodesAtIP,FE_maxNips,FE_Ngeomtypes)); FE_nodesAtIP = 0_pInt
allocate(FE_ipNeighbor(FE_maxNipNeighbors,FE_maxNips,FE_Ngeomtypes)); FE_ipNeighbor = 0_pInt
allocate(FE_cell(FE_maxNcellnodesPerCell,FE_maxNips,FE_Ngeomtypes)); FE_cell = 0_pInt
allocate(FE_cellnodeParentnodeWeights(FE_maxNnodes,FE_maxNcellnodes,FE_Nelemtypes)); FE_cellnodeParentnodeWeights = 0.0_pReal
allocate(FE_cellface(FE_maxNcellnodesPerCellface,FE_maxNcellfaces,FE_Ncelltypes)); FE_cellface = 0_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
!*** fill FE_nodesAtIP with data ***
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = 0_pInt
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 6 (2D 3node 1ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1,2,3 &
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
several small corrections/polishing
2012-04-20 17:28:41 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 125 (2D 6node 3ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
1, &
2, &
3 &
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
fixed bug concerning file handling.
2012-04-12 00:16:36 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 11 (2D 4node 4ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
1, &
2, &
4, &
3 &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 27 (2D 8node 9ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
1,0, &
1,2, &
2,0, &
1,4, &
0,0, &
2,3, &
4,0, &
3,4, &
3,0 &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 134 (3D 4node 1ip)
reshape(int([&
1,2,3,4 &
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 127 (3D 10node 4ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
1, &
2, &
3, &
4 &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 136 (3D 6node 6ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
1, &
2, &
3, &
4, &
5, &
6 &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
several small corrections/polishing
2012-04-20 17:28:41 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 117 (3D 8node 1ip)
reshape(int([&
1,2,3,4,5,6,7,8 &
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 7 (3D 8node 8ip)
reshape(int([&
1, &
2, &
4, &
3, &
5, &
6, &
8, &
7 &
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
me = me + 1_pInt
FE_nodesAtIP(1:FE_maxNnodesAtIP(me),1:FE_Nips(me),me) = & ! element 21 (3D 20node 27ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
1,0, 0,0, &
1,2, 0,0, &
2,0, 0,0, &
1,4, 0,0, &
1,3, 2,4, &
2,3, 0,0, &
4,0, 0,0, &
3,4, 0,0, &
3,0, 0,0, &
1,5, 0,0, &
1,6, 2,5, &
2,6, 0,0, &
1,8, 4,5, &
0,0, 0,0, &
2,7, 3,6, &
4,8, 0,0, &
3,8, 4,7, &
3,7, 0,0, &
5,0, 0,0, &
5,6, 0,0, &
6,0, 0,0, &
5,8, 0,0, &
5,7, 6,8, &
6,7, 0,0, &
8,0, 0,0, &
7,8, 0,0, &
7,0, 0,0 &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
],pInt),[FE_maxNnodesAtIP(me),FE_Nips(me)])
merged code for python and spectral solver interfacing (shared most of it anyway). put functionality for getting header information (dimension, homogenization, resolution) in functions in mesh.f90
2012-04-11 22:58:08 +05:30
fixed bug concerning file handling.
2012-04-12 00:16:36 +05:30
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
! *** FE_ipNeighbor ***
! is a list of the neighborhood of each IP.
! It is sorted in (local) +x,-x, +y,-y, +z,-z direction.
! Positive integers denote an intra-FE IP identifier.
! Negative integers denote the interface behind which the neighboring (extra-FE) IP will be located.
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = 0_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 6 (2D 3node 1ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
-2,-3,-1 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 125 (2D 6node 3ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
2,-3, 3,-1, &
-2, 1, 3,-1, &
2,-3,-2, 1 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 11 (2D 4node 4ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
2,-4, 3,-1, &
-2, 1, 4,-1, &
4,-4,-3, 1, &
-2, 3,-3, 2 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 27 (2D 8node 9ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
2,-4, 4,-1, &
3, 1, 5,-1, &
-2, 2, 6,-1, &
5,-4, 7, 1, &
6, 4, 8, 2, &
-2, 5, 9, 3, &
8,-4,-3, 4, &
9, 7,-3, 5, &
-2, 8,-3, 6 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 134 (3D 4node 1ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([&
-1,-2,-3,-4 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 127 (3D 10node 4ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
2,-4, 3,-2, 4,-1, &
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
-2, 1, 3,-2, 4,-1, &
2,-4,-3, 1, 4,-1, &
2,-4, 3,-2,-3, 1 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 136 (3D 6node 6ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
2,-4, 3,-2, 4,-1, &
-3, 1, 3,-2, 5,-1, &
2,-4,-3, 1, 6,-1, &
5,-4, 6,-2,-5, 1, &
-3, 4, 6,-2,-5, 2, &
5,-4,-3, 4,-5, 3 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 117 (3D 8node 1ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([&
-3,-5,-4,-2,-6,-1 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 7 (3D 8node 8ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([&
2,-5, 3,-2, 5,-1, &
-3, 1, 4,-2, 6,-1, &
4,-5,-4, 1, 7,-1, &
-3, 3,-4, 2, 8,-1, &
6,-5, 7,-2,-6, 1, &
-3, 5, 8,-2,-6, 2, &
8,-5,-4, 5,-6, 3, &
-3, 7,-4, 6,-6, 4 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_ipNeighbor(1:FE_NipNeighbors(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 21 (3D 20node 27ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
2,-5, 4,-2,10,-1, &
3, 1, 5,-2,11,-1, &
-3, 2, 6,-2,12,-1, &
5,-5, 7, 1,13,-1, &
6, 4, 8, 2,14,-1, &
-3, 5, 9, 3,15,-1, &
8,-5,-4, 4,16,-1, &
9, 7,-4, 5,17,-1, &
-3, 8,-4, 6,18,-1, &
11,-5,13,-2,19, 1, &
12,10,14,-2,20, 2, &
-3,11,15,-2,21, 3, &
14,-5,16,10,22, 4, &
15,13,17,11,23, 5, &
-3,14,18,12,24, 6, &
17,-5,-4,13,25, 7, &
18,16,-4,14,26, 8, &
-3,17,-4,15,27, 9, &
20,-5,22,-2,-6,10, &
21,19,23,-2,-6,11, &
-3,20,24,-2,-6,12, &
23,-5,25,19,-6,13, &
24,22,26,20,-6,14, &
-3,23,27,21,-6,15, &
26,-5,-4,22,-6,16, &
27,25,-4,23,-6,17, &
-3,26,-4,24,-6,18 &
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
],pInt),[FE_NipNeighbors(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
! *** FE_cell ***
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = 0_pInt
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 6 (2D 3node 1ip)
reshape(int([&
1,2,3 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 125 (2D 6node 3ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1, 4, 7, 6, &
2, 5, 7, 4, &
3, 6, 7, 5 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 11 (2D 4node 4ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1, 5, 9, 8, &
5, 2, 6, 9, &
8, 9, 7, 4, &
9, 6, 3, 7 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 27 (2D 8node 9ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1, 5,13,12, &
5, 6,14,13, &
6, 2, 7,14, &
12,13,16,11, &
13,14,15,16, &
14, 7, 8,15, &
11,16,10, 4, &
16,15, 9,10, &
15, 8, 3, 9 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 134 (3D 4node 1ip)
reshape(int([&
1, 2, 3, 4 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 127 (3D 10node 4ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1, 5,11, 7, 8,12,15,14, &
5, 2, 6,11,12, 9,13,15, &
7,11, 6, 3,14,15,13,10, &
8,12,15, 4, 4, 9,13,10 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 136 (3D 6node 6ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1, 7,16, 9,10,17,21,19, &
7, 2, 8,16,17,11,18,21, &
9,16, 8, 3,19,21,18,12, &
10,17,21,19, 4,13,20,15, &
17,11,18,21,13, 5,14,20, &
19,21,18,12,15,20,14, 6 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 117 (3D 8node 1ip)
reshape(int([&
1, 2, 3, 4, 5, 6, 7, 8 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 7 (3D 8node 8ip)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1, 9,21,12,13,22,27,25, &
9, 2,10,21,22,14,23,27, &
12,21,11, 4,25,27,24,16, &
21,10, 3,11,27,23,15,24, &
13,22,27,25, 5,17,26,20, &
22,14,23,27,17, 6,18,26, &
25,27,24,16,20,26,19, 8, &
27,23,15,24,26,18, 7,19 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cell(1:FE_NcellnodesPerCell(FE_celltype(me)),1:FE_Nips(me),me) = & ! element 21 (3D 20node 27ip)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1, 9,33,16,17,37,57,44, &
9,10,34,33,37,38,58,57, &
10, 2,11,34,38,18,39,58, &
16,33,36,15,44,57,60,43, &
33,34,35,36,57,58,59,60, &
34,11,12,35,58,39,40,59, &
15,36,14, 4,43,60,42,20, &
36,35,13,14,60,59,41,42, &
35,12, 3,13,59,40,19,41, &
17,37,57,44,21,45,61,52, &
37,38,58,57,45,46,62,61, &
38,18,39,58,46,22,47,62, &
44,57,60,43,52,61,64,51, &
57,58,59,60,61,62,63,64, &
58,39,40,59,62,47,48,63, &
43,60,42,20,51,64,50,24, &
60,59,41,42,64,63,49,50, &
59,40,19,41,63,48,23,49, &
21,45,61,52, 5,25,53,32, &
45,46,62,61,25,26,54,53, &
46,22,47,62,26, 6,27,54, &
52,61,64,51,32,53,56,31, &
61,62,63,64,53,54,55,56, &
62,47,48,63,54,27,28,55, &
51,64,50,24,31,56,30, 8, &
64,63,49,50,56,55,29,30, &
63,48,23,49,55,28, 7,29 &
],pInt),[FE_NcellnodesPerCell(FE_celltype(me)),FE_Nips(me)])
! *** FE_cellnodeParentnodeWeights ***
! center of gravity of the weighted nodes gives the position of the cell node.
! fill with 0.
! example: face-centered cell node with face nodes 1,2,5,6 to be used in,
! e.g., an 8 node element, would be encoded:
! 1, 1, 0, 0, 1, 1, 0, 0
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = 0_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
me = me + 1_pInt
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 6 (2D 3node 1ip)
reshape(real([&
1, 0, 0, &
0, 1, 0, &
0, 0, 1 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
me = me + 1_pInt
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 125 (2D 6node 3ip)
reshape(real([&
1, 0, 0, 0, 0, 0, &
0, 1, 0, 0, 0, 0, &
0, 0, 1, 0, 0, 0, &
0, 0, 0, 1, 0, 0, &
0, 0, 0, 0, 1, 0, &
0, 0, 0, 0, 0, 1, &
1, 1, 1, 2, 2, 2 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 11 (2D 4node 4ip)
reshape(real([&
1, 0, 0, 0, &
0, 1, 0, 0, &
0, 0, 1, 0, &
0, 0, 0, 1, &
1, 1, 0, 0, &
0, 1, 1, 0, &
0, 0, 1, 1, &
1, 0, 0, 1, &
1, 1, 1, 1 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 27 (2D 8node 9ip)
reshape(real([&
1, 0, 0, 0, 0, 0, 0, 0, &
0, 1, 0, 0, 0, 0, 0, 0, &
0, 0, 1, 0, 0, 0, 0, 0, &
0, 0, 0, 1, 0, 0, 0, 0, &
1, 0, 0, 0, 2, 0, 0, 0, &
0, 1, 0, 0, 2, 0, 0, 0, &
0, 1, 0, 0, 0, 2, 0, 0, &
0, 0, 1, 0, 0, 2, 0, 0, &
0, 0, 1, 0, 0, 0, 2, 0, &
0, 0, 0, 1, 0, 0, 2, 0, &
0, 0, 0, 1, 0, 0, 0, 2, &
1, 0, 0, 0, 0, 0, 0, 2, &
4, 1, 1, 1, 8, 2, 2, 8, &
1, 4, 1, 1, 8, 8, 2, 2, &
1, 1, 4, 1, 2, 8, 8, 2, &
1, 1, 1, 4, 2, 2, 8, 8 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 54 (2D 8node 4ip)
reshape(real([&
1, 0, 0, 0, 0, 0, 0, 0, &
0, 1, 0, 0, 0, 0, 0, 0, &
0, 0, 1, 0, 0, 0, 0, 0, &
0, 0, 0, 1, 0, 0, 0, 0, &
0, 0, 0, 0, 1, 0, 0, 0, &
0, 0, 0, 0, 0, 1, 0, 0, &
0, 0, 0, 0, 0, 0, 1, 0, &
0, 0, 0, 0, 0, 0, 0, 1, &
1, 1, 1, 1, 2, 2, 2, 2 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 134 (3D 4node 1ip)
reshape(real([&
1, 0, 0, 0, &
0, 1, 0, 0, &
0, 0, 1, 0, &
0, 0, 0, 1 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 157 (3D 5node 4ip)
reshape(real([&
1, 0, 0, 0, 0, &
0, 1, 0, 0, 0, &
0, 0, 1, 0, 0, &
0, 0, 0, 1, 0, &
1, 1, 0, 0, 0, &
0, 1, 1, 0, 0, &
1, 0, 1, 0, 0, &
1, 0, 0, 1, 0, &
0, 1, 0, 1, 0, &
0, 0, 1, 1, 0, &
1, 1, 1, 0, 0, &
1, 1, 0, 1, 0, &
0, 1, 1, 1, 0, &
1, 0, 1, 1, 0, &
0, 0, 0, 0, 1 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 127 (3D 10node 4ip)
reshape(real([&
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, &
0, 1, 0, 0, 0, 0, 0, 0, 0, 0, &
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, &
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, &
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, &
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, &
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, &
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, &
0, 0, 0, 0, 0, 0, 0, 0, 1, 0, &
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, &
1, 1, 1, 0, 2, 2, 2, 0, 0, 0, &
1, 1, 0, 1, 2, 0, 0, 2, 2, 0, &
0, 1, 1, 1, 0, 2, 0, 0, 2, 2, &
1, 0, 1, 1, 0, 0, 2, 2, 0, 2, &
3, 3, 3, 3, 4, 4, 4, 4, 4, 4 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
me = me + 1_pInt
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 136 (3D 6node 6ip)
reshape(real([&
1, 0, 0, 0, 0, 0, &
0, 1, 0, 0, 0, 0, &
0, 0, 1, 0, 0, 0, &
0, 0, 0, 1, 0, 0, &
0, 0, 0, 0, 1, 0, &
0, 0, 0, 0, 0, 1, &
1, 1, 0, 0, 0, 0, &
0, 1, 1, 0, 0, 0, &
1, 0, 1, 0, 0, 0, &
1, 0, 0, 1, 0, 0, &
0, 1, 0, 0, 1, 0, &
0, 0, 1, 0, 0, 1, &
0, 0, 0, 1, 1, 0, &
0, 0, 0, 0, 1, 1, &
0, 0, 0, 1, 0, 1, &
1, 1, 1, 0, 0, 0, &
1, 1, 0, 1, 1, 0, &
0, 1, 1, 0, 1, 1, &
1, 0, 1, 1, 0, 1, &
0, 0, 0, 1, 1, 1, &
1, 1, 1, 1, 1, 1 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
me = me + 1_pInt
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 117 (3D 8node 1ip)
reshape(real([&
1, 0, 0, 0, 0, 0, 0, 0, &
0, 1, 0, 0, 0, 0, 0, 0, &
0, 0, 1, 0, 0, 0, 0, 0, &
0, 0, 0, 1, 0, 0, 0, 0, &
0, 0, 0, 0, 1, 0, 0, 0, &
0, 0, 0, 0, 0, 1, 0, 0, &
0, 0, 0, 0, 0, 0, 1, 0, &
0, 0, 0, 0, 0, 0, 0, 1 &
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
me = me + 1_pInt
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 7 (3D 8node 8ip)
reshape(real([&
1, 0, 0, 0, 0, 0, 0, 0, & !
0, 1, 0, 0, 0, 0, 0, 0, & !
0, 0, 1, 0, 0, 0, 0, 0, & !
0, 0, 0, 1, 0, 0, 0, 0, & !
0, 0, 0, 0, 1, 0, 0, 0, & ! 5
0, 0, 0, 0, 0, 1, 0, 0, & !
0, 0, 0, 0, 0, 0, 1, 0, & !
0, 0, 0, 0, 0, 0, 0, 1, & !
1, 1, 0, 0, 0, 0, 0, 0, & !
0, 1, 1, 0, 0, 0, 0, 0, & ! 10
0, 0, 1, 1, 0, 0, 0, 0, & !
1, 0, 0, 1, 0, 0, 0, 0, & !
1, 0, 0, 0, 1, 0, 0, 0, & !
0, 1, 0, 0, 0, 1, 0, 0, & !
0, 0, 1, 0, 0, 0, 1, 0, & ! 15
0, 0, 0, 1, 0, 0, 0, 1, & !
0, 0, 0, 0, 1, 1, 0, 0, & !
0, 0, 0, 0, 0, 1, 1, 0, & !
0, 0, 0, 0, 0, 0, 1, 1, & !
0, 0, 0, 0, 1, 0, 0, 1, & ! 20
1, 1, 1, 1, 0, 0, 0, 0, & !
1, 1, 0, 0, 1, 1, 0, 0, & !
0, 1, 1, 0, 0, 1, 1, 0, & !
0, 0, 1, 1, 0, 0, 1, 1, & !
1, 0, 0, 1, 1, 0, 0, 1, & ! 25
0, 0, 0, 0, 1, 1, 1, 1, & !
1, 1, 1, 1, 1, 1, 1, 1 & !
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
me = me + 1_pInt
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 57 (3D 20node 8ip)
reshape(real([&
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & ! 5
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & ! 10
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, & ! 15
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, & ! 20
1, 1, 1, 1, 0, 0, 0, 0, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, & !
1, 1, 0, 0, 1, 1, 0, 0, 2, 0, 0, 0, 2, 0, 0, 0, 2, 2, 0, 0, & !
0, 1, 1, 0, 0, 1, 1, 0, 0, 2, 0, 0, 0, 2, 0, 0, 0, 2, 2, 0, & !
0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 2, 0, 0, 0, 2, 0, 0, 0, 2, 2, & !
1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 2, 0, 0, 0, 2, 2, 0, 0, 2, & ! 25
0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 2, 2, 2, 2, 0, 0, 0, 0, & !
3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4 & !
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
FE_cellnodeParentnodeWeights(1:FE_Nnodes(me),1:FE_Ncellnodes(FE_geomtype(me)),me) = & ! element 21 (3D 20node 27ip)
reshape(real([&
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & ! 5
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & ! 10
0, 1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, & ! 15
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, & !
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, & !
0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, & !
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, & !
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, & ! 20
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, & !
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, & !
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, & !
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, & !
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, & ! 25
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, & !
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, & ! 30
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, & !
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, & !
4, 1, 1, 1, 0, 0, 0, 0, 8, 2, 2, 8, 0, 0, 0, 0, 0, 0, 0, 0, & !
1, 4, 1, 1, 0, 0, 0, 0, 8, 8, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, & !
1, 1, 4, 1, 0, 0, 0, 0, 2, 8, 8, 2, 0, 0, 0, 0, 0, 0, 0, 0, & ! 35
1, 1, 1, 4, 0, 0, 0, 0, 2, 2, 8, 8, 0, 0, 0, 0, 0, 0, 0, 0, & !
4, 1, 0, 0, 1, 1, 0, 0, 8, 0, 0, 0, 2, 0, 0, 0, 8, 2, 0, 0, & !
1, 4, 0, 0, 1, 1, 0, 0, 8, 0, 0, 0, 2, 0, 0, 0, 2, 8, 0, 0, & !
0, 4, 1, 0, 0, 1, 1, 0, 0, 8, 0, 0, 0, 2, 0, 0, 0, 8, 2, 0, & !
0, 1, 4, 0, 0, 1, 1, 0, 0, 8, 0, 0, 0, 2, 0, 0, 0, 2, 8, 0, & ! 40
0, 0, 4, 1, 0, 0, 1, 1, 0, 0, 8, 0, 0, 0, 2, 0, 0, 0, 8, 2, & !
0, 0, 1, 4, 0, 0, 1, 1, 0, 0, 8, 0, 0, 0, 2, 0, 0, 0, 2, 8, & !
1, 0, 0, 4, 1, 0, 0, 1, 0, 0, 0, 8, 0, 0, 0, 2, 2, 0, 0, 8, & !
4, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 8, 0, 0, 0, 2, 8, 0, 0, 2, & !
1, 1, 0, 0, 4, 1, 0, 0, 2, 0, 0, 0, 8, 0, 0, 0, 8, 2, 0, 0, & ! 45
1, 1, 0, 0, 1, 4, 0, 0, 2, 0, 0, 0, 8, 0, 0, 0, 2, 8, 0, 0, & !
0, 1, 1, 0, 0, 4, 1, 0, 0, 2, 0, 0, 0, 8, 0, 0, 0, 8, 2, 0, & !
0, 1, 1, 0, 0, 1, 4, 0, 0, 2, 0, 0, 0, 8, 0, 0, 0, 2, 8, 0, & !
0, 0, 1, 1, 0, 0, 4, 1, 0, 0, 2, 0, 0, 0, 8, 0, 0, 0, 8, 2, & !
0, 0, 1, 1, 0, 0, 1, 4, 0, 0, 2, 0, 0, 0, 8, 0, 0, 0, 2, 8, & ! 50
1, 0, 0, 1, 1, 0, 0, 4, 0, 0, 0, 2, 0, 0, 0, 8, 2, 0, 0, 8, & !
1, 0, 0, 1, 4, 0, 0, 1, 0, 0, 0, 2, 0, 0, 0, 8, 8, 0, 0, 2, & !
0, 0, 0, 0, 4, 1, 1, 1, 0, 0, 0, 0, 8, 2, 2, 8, 0, 0, 0, 0, & !
0, 0, 0, 0, 1, 4, 1, 1, 0, 0, 0, 0, 8, 8, 2, 2, 0, 0, 0, 0, & !
0, 0, 0, 0, 1, 1, 4, 1, 0, 0, 0, 0, 2, 8, 8, 2, 0, 0, 0, 0, & ! 55
0, 0, 0, 0, 1, 1, 1, 4, 0, 0, 0, 0, 2, 2, 8, 8, 0, 0, 0, 0, & !
24, 8, 4, 8, 8, 4, 3, 4, 32,12,12,32, 12, 4, 4,12, 32,12, 4,12, & !
8,24, 8, 4, 4, 8, 4, 3, 32,32,12,12, 12,12, 4, 4, 12,32,12, 4, & !
4, 8,24, 8, 3, 4, 8, 4, 12,32,32,12, 4,12,12, 4, 4,12,32,12, & !
8, 4, 8,24, 4, 3, 4, 8, 12,12,32,32, 4, 4,12,12, 12, 4,12,32, & ! 60
8, 4, 3, 4, 24, 8, 4, 8, 12, 4, 4,12, 32,12,12,32, 32,12, 4,12, & !
4, 8, 4, 3, 8,24, 8, 4, 12,12, 4, 4, 32,32,12,12, 12,32,12, 4, & !
3, 4, 8, 4, 4, 8,24, 8, 4,12,12, 4, 12,32,32,12, 4,12,32,12, & !
4, 3, 4, 8, 8, 4, 8,24, 4, 4,12,12, 12,12,32,32, 12, 4,12,32 & !
],pReal),[FE_Nnodes(me),FE_Ncellnodes(FE_geomtype(me))])
! *** FE_cellface ***
me = 0_pInt
me = me + 1_pInt
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
FE_cellface(1:FE_NcellnodesPerCellface(me),1:FE_NipNeighbors(me),me) = & ! 2D 3node, VTK_TRIANGLE (5)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
2,3, &
3,1, &
1,2 &
],pInt),[FE_NcellnodesPerCellface(me),FE_NipNeighbors(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
FE_cellface(1:FE_NcellnodesPerCellface(me),1:FE_NipNeighbors(me),me) = & ! 2D 4node, VTK_QUAD (9)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
2,3, &
4,1, &
3,4, &
1,2 &
],pInt),[FE_NcellnodesPerCellface(me),FE_NipNeighbors(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
FE_cellface(1:FE_NcellnodesPerCellface(me),1:FE_NipNeighbors(me),me) = & ! 3D 4node, VTK_TETRA (10)
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
1,3,2, &
1,2,4, &
2,3,4, &
1,4,3 &
],pInt),[FE_NcellnodesPerCellface(me),FE_NipNeighbors(me)])
switched element library to geomType based. saves to copy same geometry description for different elements that are essentially similar regarding the IP number but differ in total node count. introduced quadratic tetrahedron (Marc element 127 -- element 157 might also work, but did not perform well in fully elastic calc so far)
2012-11-16 04:15:20 +05:30
me = me + 1_pInt
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
FE_cellface(1:FE_NcellnodesPerCellface(me),1:FE_NipNeighbors(me),me) = & ! 3D 8node, VTK_HEXAHEDRON (12)
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
reshape(int([&
Major restructuring of mesh.f90 due to introduction of "cells". A cell is the volume (area in 2D elements) that surrounds an ip. It is described by up to 8 cell nodes and is used to calculate ip volume, area, and face normals. It can also be used to generate a mesh for e.g. paraview that contains ip output as cell centered data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-15 13:43:20 +05:30
2,3,7,6, &
4,1,5,8, &
3,4,8,7, &
1,2,6,5, &
5,6,7,8, &
1,4,3,2 &
],pInt),[FE_NcellnodesPerCellface(me),FE_NipNeighbors(me)])
introduced preprocessor identifiers Marc, Abaqus, and Spectral to enable conditional compilation. This allows deleted dummy functions that are used by one solver only. Mainly affected modules are IO and mesh. Most of the changes in mesh result from reordering the functions when grouping them depending on their solver. Further advantage is that FE solver do not need FFTW and kdtree2 anymore. The include files for these two libraries moved to DAMASKROO/lib now as I figured out how to use a include path in the Makefile. Put all the files I got when testing compilation with abaqus in a folder which to become the abaqus compilation test.
2012-06-15 21:40:21 +05:30
end subroutine mesh_build_FEdata
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
!> @brief writes out initial cell geometry
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
!--------------------------------------------------------------------------------------------------
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
subroutine mesh_write_cellGeom
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
use DAMASK_interface, only: &
vtk files are now correctly written to current working directory. updated Lib_VTK_IO.f90
2013-05-08 21:40:21 +05:30
getSolverJobName, &
getSolverWorkingDirectoryName
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
use Lib_VTK_IO, only: &
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
VTK_ini, &
VTK_geo, &
VTK_con, &
VTK_end
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
implicit none
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
integer(pInt), dimension(1:mesh_Ncells) :: celltype
integer(pInt), dimension(mesh_Ncells*(1_pInt+FE_maxNcellnodesPerCell)) :: cellconnection
integer(pInt):: err, g, c, e, CellID, i, j
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
cellID = 0_pInt
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
j = 0_pInt
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
do e = 1_pInt, mesh_NcpElems ! loop over cpElems
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
g = FE_geomtype(mesh_element(2_pInt,e)) ! get geometry type
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
c = FE_celltype(g) ! get cell type
do i = 1_pInt,FE_Nips(g) ! loop over ips=cells in this element
cellID = cellID + 1_pInt
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
celltype(cellID) = MESH_VTKCELLTYPE(c)
cellconnection(j+1_pInt:j+FE_NcellnodesPerCell(c)+1_pInt) &
= [FE_NcellnodesPerCell(c),mesh_cell(1:FE_NcellnodesPerCell(c),i,e)-1_pInt] ! number of cellnodes per cell & list of global cellnode IDs belnging to this cell (cellnode counting starts at 0)
j = j + FE_NcellnodesPerCell(c) + 1_pInt
enddo
enddo
err = VTK_ini(output_format = 'ASCII', &
vtk files are now correctly written to current working directory. updated Lib_VTK_IO.f90
2013-05-08 21:40:21 +05:30
title=trim(getSolverJobName())//' cell mesh', &
filename = trim(getSolverWorkingDirectoryName())//trim(getSolverJobName())//'_ipbased.vtk', &
mesh_topology = 'UNSTRUCTURED_GRID')
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
err = VTK_geo(NN = mesh_Ncellnodes, &
some lines in crystallite.f90 exceeded limit of 132 characters fixed bug in Lib_VTK_IO.f90 (informed author as well) added zip file containing source of Lib_VTK_IO.f90
2013-05-08 23:18:00 +05:30
X = mesh_cellnode(1,1:mesh_Ncellnodes), &
Y = mesh_cellnode(2,1:mesh_Ncellnodes), &
Z = mesh_cellnode(3,1:mesh_Ncellnodes))
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
err = VTK_con(NC = mesh_Ncells, &
vtk files are now correctly written to current working directory. updated Lib_VTK_IO.f90
2013-05-08 21:40:21 +05:30
connect = cellconnection(1:j), &
cell_type = celltype)
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
err = VTK_end()
end subroutine mesh_write_cellGeom
!--------------------------------------------------------------------------------------------------
!> @brief writes out initial element geometry
!--------------------------------------------------------------------------------------------------
subroutine mesh_write_elemGeom
use DAMASK_interface, only: &
vtk files are now correctly written to current working directory. updated Lib_VTK_IO.f90
2013-05-08 21:40:21 +05:30
getSolverJobName, &
getSolverWorkingDirectoryName
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
use Lib_VTK_IO, only: &
VTK_ini, &
VTK_geo, &
VTK_con, &
VTK_end
implicit none
integer(pInt), dimension(1:mesh_NcpElems) :: elemtype
integer(pInt), dimension(mesh_NcpElems*(1_pInt+FE_maxNnodes)) :: elementconnection
integer(pInt):: err, e, t, n, i
i = 0_pInt
do e = 1_pInt, mesh_NcpElems ! loop over cpElems
t = mesh_element(2,e) ! get element type
elemtype(e) = MESH_VTKELEMTYPE(t)
elementconnection(i+1_pInt) = FE_Nnodes(t) ! number of nodes per element
do n = 1_pInt,FE_Nnodes(t)
elementconnection(i+1_pInt+n) = mesh_element(4_pInt+n,e) - 1_pInt ! global node ID of node that belongs to this element (node counting starts at 0)
enddo
i = i + 1_pInt + FE_Nnodes(t)
enddo
some lines in crystallite.f90 exceeded limit of 132 characters fixed bug in Lib_VTK_IO.f90 (informed author as well) added zip file containing source of Lib_VTK_IO.f90
2013-05-08 23:18:00 +05:30
err =VTK_ini(output_format = 'ASCII', &
title=trim(getSolverJobName())//' element mesh', &
filename = trim(getSolverWorkingDirectoryName())//trim(getSolverJobName())//'_nodebased.vtk', &
mesh_topology = 'UNSTRUCTURED_GRID')
err =VTK_geo(NN = mesh_Nnodes, &
X = mesh_node0(1,1:mesh_Nnodes), &
Y = mesh_node0(2,1:mesh_Nnodes), &
Z = mesh_node0(3,1:mesh_Nnodes))
err =VTK_con(NC = mesh_Nelems, &
connect = elementconnection(1:i), &
cell_type = elemtype)
err =VTK_end()
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
end subroutine mesh_write_elemGeom
!--------------------------------------------------------------------------------------------------
!> @brief writes description file for mesh
!--------------------------------------------------------------------------------------------------
subroutine mesh_write_meshfile
use IO, only: &
IO_write_jobFile
implicit none
integer(pInt), parameter :: fileUnit = 223_pInt
integer(pInt) :: e,i,t,g,c,n
call IO_write_jobFile(fileUnit,'mesh')
corrected format strings in mesh_read_meshfile and mesh_write_meshfile
2013-05-07 18:57:51 +05:30
write(fileUnit,'(A16,E10.3)') 'unitlength', mesh_unitlength
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
write(fileUnit,'(A16,I10)') 'maxNcellnodes', mesh_maxNcellnodes
write(fileUnit,'(A16,I10)') 'maxNips', mesh_maxNips
write(fileUnit,'(A16,I10)') 'maxNnodes', mesh_maxNnodes
write(fileUnit,'(A16,I10)') 'Nnodes', mesh_Nnodes
write(fileUnit,'(A16,I10)') 'NcpElems', mesh_NcpElems
do e = 1_pInt,mesh_NcpElems
t = mesh_element(2,e)
write(fileUnit,'(20(I10))') mesh_element(1_pInt:4_pInt+FE_Nnodes(t),e)
enddo
write(fileUnit,'(A16,I10)') 'Ncellnodes', mesh_Ncellnodes
do n = 1_pInt,mesh_Ncellnodes
write(fileUnit,'(2(I10))') mesh_cellnodeParent(1:2,n)
enddo
write(fileUnit,'(A16,I10)') 'Ncells', mesh_Ncells
do e = 1_pInt,mesh_NcpElems
t = mesh_element(2,e)
g = FE_geomtype(t)
c = FE_celltype(g)
do i = 1_pInt,FE_Nips(g)
write(fileUnit,'(8(I10))') &
mesh_cell(1_pInt:FE_NcellnodesPerCell(c),i,e)
enddo
enddo
close(fileUnit)
end subroutine mesh_write_meshfile
!--------------------------------------------------------------------------------------------------
!> @brief reads mesh description file
!--------------------------------------------------------------------------------------------------
integer function mesh_read_meshfile(filepath)
implicit none
character(len=*), intent(in) :: filepath
integer(pInt), parameter :: fileUnit = 223_pInt
integer(pInt) :: e,i,t,g,n
open(fileUnit,status='old',err=100,iostat=mesh_read_meshfile,action='read',file=filepath)
corrected format strings in mesh_read_meshfile and mesh_write_meshfile
2013-05-07 18:57:51 +05:30
read(fileUnit,'(TR16,E10.3)',err=100,iostat=mesh_read_meshfile) mesh_unitlength
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
read(fileUnit,'(TR16,I10)',err=100,iostat=mesh_read_meshfile) mesh_maxNcellnodes
read(fileUnit,'(TR16,I10)',err=100,iostat=mesh_read_meshfile) mesh_maxNips
read(fileUnit,'(TR16,I10)',err=100,iostat=mesh_read_meshfile) mesh_maxNnodes
read(fileUnit,'(TR16,I10)',err=100,iostat=mesh_read_meshfile) mesh_Nnodes
read(fileUnit,'(TR16,I10)',err=100,iostat=mesh_read_meshfile) mesh_NcpElems
if (.not. allocated(mesh_element)) allocate(mesh_element(4_pInt+mesh_maxNnodes,mesh_NcpElems))
mesh_element = 0_pInt
do e = 1_pInt,mesh_NcpElems
read(fileUnit,'(20(I10))',err=100,iostat=mesh_read_meshfile) &
mesh_element(:,e)
enddo
read(fileUnit,'(TR16,I10)',err=100,iostat=mesh_read_meshfile) mesh_Ncellnodes
if (.not. allocated(mesh_cellnodeParent)) allocate(mesh_cellnodeParent(2_pInt,mesh_Ncellnodes))
do n = 1_pInt,mesh_Ncellnodes
read(fileUnit,'(2(I10))',err=100,iostat=mesh_read_meshfile) mesh_cellnodeParent(1:2,n)
enddo
read(fileUnit,'(TR16,I10)',err=100,iostat=mesh_read_meshfile) mesh_Ncells
if (.not. allocated(mesh_cell)) allocate(mesh_cell(FE_maxNcellnodesPerCell,mesh_maxNips,mesh_NcpElems))
do e = 1_pInt,mesh_NcpElems
t = mesh_element(2,e)
g = FE_geomtype(t)
do i = 1_pInt,FE_Nips(g)
read(fileUnit,'(8(I10))',err=100,iostat=mesh_read_meshfile) mesh_cell(:,i,e)
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
enddo
enddo
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
close(fileUnit)
mesh_read_meshfile = 0 ! successfully read data
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
100 continue
end function mesh_read_meshfile
!--------------------------------------------------------------------------------------------------
!> @brief initializes mesh data for use in post processing
!--------------------------------------------------------------------------------------------------
integer function mesh_init_postprocessing(filepath)
implicit none
character(len=*), intent(in) :: filepath
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
call mesh_build_FEdata
mesh_init_postprocessing = mesh_read_meshfile(filepath)
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
calculation of cellnode positions now available in python via damask.core first call damask.core.mesh.mesh_init_postprocessing(meshfilename) to initialize all necessary mesh variables then damask.core.mesh.mesh_build_cellnodes(nodes) calculates the cellnode positions for a given list of node positions the meshfile that is needed for the init is created automatically by mesh_init in DAMASK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! For marc simulations, run ./code/setup/setup_code.sh !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2013-04-22 00:18:59 +05:30
end function mesh_init_postprocessing
introduced output of initial geometry (vtk file) to mesh
2013-04-18 22:10:49 +05:30
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
!--------------------------------------------------------------------------------------------------
added two new scripts: "marc_extractData" is a simple variant of the "postResult" script that extracts all data from a macro *.t16 file (quadratic elements not yet supported) and writes it into an asciitable "addDataToGeometry" searches for corresponding *.txt and *.vtk file in a given directory and adds the data from the *.txt file as SCALARS to the *.vtk file first script needed an additional function in mesh that returns the corresponding node of a given ip in a specific element
2013-04-29 02:26:17 +05:30
!> @brief returns global variable mesh_Ncellnodes
added python script that produces vtk files with the (node based and ip based) deformed mesh from marc output file
2013-04-24 00:00:56 +05:30
!--------------------------------------------------------------------------------------------------
integer(pInt) function mesh_get_Ncellnodes()
implicit none
mesh_get_Ncellnodes = mesh_Ncellnodes
end function mesh_get_Ncellnodes
"unitlength" parameter, which determines the physical size of the mesh, now available as a global mesh variable "mesh_unitlength" and written to the .mesh file during init. Hence, it is available to the marc_deformedGeometry.py script via "mesh_init_postprocessing()" and "mesh_get_unitlength()", so no need for setting the scaling of the displacement vectors explicitly through an option; now displacements and nodal positions are always consistent
2013-05-07 18:36:29 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief returns global variable mesh_unitlength
!--------------------------------------------------------------------------------------------------
real(pReal) function mesh_get_unitlength()
implicit none
mesh_get_unitlength = mesh_unitlength
end function mesh_get_unitlength
added two new scripts: "marc_extractData" is a simple variant of the "postResult" script that extracts all data from a macro *.t16 file (quadratic elements not yet supported) and writes it into an asciitable "addDataToGeometry" searches for corresponding *.txt and *.vtk file in a given directory and adds the data from the *.txt file as SCALARS to the *.vtk file first script needed an additional function in mesh that returns the corresponding node of a given ip in a specific element
2013-04-29 02:26:17 +05:30
!--------------------------------------------------------------------------------------------------
!> @brief returns node that is located at an ip
!> @details return zero if requested ip does not exist or not available (more ips than nodes)
!--------------------------------------------------------------------------------------------------
integer(pInt) function mesh_get_nodeAtIP(elemtypeFE,ip)
implicit none
character(len=*), intent(in) :: elemtypeFE
integer(pInt), intent(in) :: ip
integer(pInt) :: elemtype
integer(pInt) :: geomtype
mesh_get_nodeAtIP = 0_pInt
elemtype = FE_mapElemtype(elemtypeFE)
geomtype = FE_geomtype(elemtype)
if (FE_Nips(geomtype) >= ip .and. FE_Nips(geomtype) <= FE_Nnodes(elemtype)) &
mesh_get_nodeAtIP = FE_nodesAtIP(1,ip,geomtype)
end function mesh_get_nodeAtIP
fixed reading/writing of integer arrays with function for real arrays
2012-08-16 20:25:23 +05:30
end module mesh
integrated subroutine for regridding into mesh.f90 and made it available for python
2012-04-10 20:45:46 +05:30