223031012
/
DAMASK_EICMD
mirror of https://github.com/achalhp/DAMASK_EICMD.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
4,110 Commits 5 Branches 20 Tags 292 MiB
Commit Graph

1919 Commits

Author SHA1 Message Date
Pratheek Shanthraj 1330dae432 cleaning up, removing unused variables and variable initialisation 2015-04-21 15:16:13 +00:00
Pratheek Shanthraj 0681f67570 corrected buggy previous commit 2015-04-21 14:33:38 +00:00
Pratheek Shanthraj bec17127d0 write output information only for active homogenisations. removed unused variables 2015-04-21 14:10:34 +00:00
Pratheek Shanthraj d049eadce7 write output information only for active crystallites. removed unused variables. fixed some real == 0 comparisons 2015-04-21 12:23:00 +00:00
Pratheek Shanthraj effec313ee fixed bug in cleavage system set up for ortho lattice structures 2015-04-21 12:19:41 +00:00
Martin Diehl ea1b2191f1 deleted comment, test passed probably because fftw is system-wide available at MPIE 
also, ../lib is not needed since lib is an include dir via -I
 2015-04-19 09:07:16 +00:00
Philip Eisenlohr a46b9287eb restored conditional (#ifdef Spectral) inclusion of “fftw3”. Not clear how this could pass the Marc compile test..?!? (failed in my “real-world” use) 2015-04-18 15:23:37 +00:00
Philip Eisenlohr 22f9a90bcb added informative ext_msg to error 135 2015-04-18 15:22:15 +00:00
Martin Diehl eb3173c6cc ugly hack for Abaqus in prec, it seems that it links against old (10.1) ifort where IEEE_arithmetic is not avaialbe. 
Took the chance and cleaned the env file and used #ifdef# statement in the wrappers because in fixed format Fortran some lines of code might be interpretated as comments.
This should not happen, but it happens for Abaqus exp
 2015-04-15 18:10:56 +00:00
Martin Diehl caf6611485 unused variable 2015-04-15 18:05:27 +00:00
Martin Diehl f5762209dc changed 'range' keyword to 'limit', now using proper ASCII table (with column 'intensity') for linearODF 
hybridIO_linODFsampling simplified
 2015-04-15 14:37:46 +00:00
Martin Diehl e29628b459 changed keyword for texture component/ODF simulation 
kewyword "origin" with possible values "center/voxelcenter" and "boundary/boundarycenter" is always needed
 2015-04-14 12:45:20 +00:00
Martin Diehl 7e0b79a33c also using isNaN function 2015-04-14 11:43:31 +00:00
Martin Diehl 25665b1d80 IEEE_value (like any other function from intrinsic modules) is not allowed for parameters 2015-04-14 05:21:08 +00:00
Martin Diehl 0c7e7c66f8 integer type was use for void fraction, should be real 2015-04-13 10:36:45 +00:00
Martin Diehl 5c1804e77d improved NaN checks 2015-04-13 10:02:52 +00:00
Martin Diehl 50998bd6a4 simplified makefile and remove unused variables in RGC 2015-04-12 08:11:22 +00:00
Martin Diehl 7156e3996b gfortran noticed that these functions are external 2015-04-11 20:28:07 +00:00
Martin Diehl 0ba8f27320 took some gfortran complaints serious (unused imports, implicit castings) 2015-04-11 11:47:33 +00:00
Martin Diehl 064266c0cd updated linearODF file format (including header and keywords) 2015-04-11 10:28:10 +00:00
Martin Diehl b4456e76b3 introduced two small bugs 2015-04-11 09:25:15 +00:00
Martin Diehl 470fc2dce3 comparison for zero should be abs: 
a == 0.0_pReal   becomes abs(a) <= tiny(a)
a /= 0.0_pReal   becomes abs(a) > tiny(a)
remove unused variables
 2015-04-11 08:25:23 +00:00
Martin Diehl 6170209198 use statement for marc was wrong (preprocessor) 2015-04-11 05:09:15 +00:00
Martin Diehl d7b36c2c30 gfortran complaint about implicit casting,floating point comparison, and unused imported variables. 
additionally changed use of math_invert33 to math_inv33 if det is not needed
 2015-04-10 19:09:26 +00:00
Martin Diehl 7d0a1ebea6 math_transpose33 not used 2015-04-08 18:33:08 +00:00
Martin Diehl b6481c2513 introduced simpler multiplication and forall loops 
matmul is ok for openmp, check in the web and run the state integration test.

Example program testing for new state update for rkck dot state:

program test

real, dimension(6,10) :: dotState=reshape(&
                      [1,1,1,1,1,1,1,1,1,1,&
                       2,2,2,2,2,2,2,2,2,2,&
                       3,3,3,3,3,3,3,3,3,3,&
                       4,4,4,4,4,4,4,4,4,4,&
                       5,5,5,5,5,5,5,5,5,5,&
                       6,6,6,6,6,6,6,6,6,6],[6,10])

real, dimension(10) :: residuum
real, dimension(6) :: B=2.5

integer :: i

residuum = B(1)*dotState(1,:)+& 
           B(2)*dotState(2,:)+& 
           B(3)*dotState(3,:)+& 
           B(4)*dotState(4,:)+& 
           B(5)*dotState(5,:)+& 
           B(6)*dotState(6,:)
do i =1,10
  print*,residuum(i)
enddo

residuum =  matmul(transpose(dotState),B)
do i =1,10
  print*,residuum(i)
enddo

end program test
 2015-04-01 16:45:53 +00:00
Martin Diehl 693efcaa58 comment was confused by search and replace 2015-04-01 16:04:33 +00:00
Martin Diehl 63491bf268 for some reasons, ifort was complaining about mixed integer types. 
also deleted #ifdef PETSC since we do not support the solver without PETSC
 2015-03-30 20:49:17 +00:00
Pratheek Shanthraj c7418db9bd FEM : hierarchical ordering of h5 output data and more meaningful visualisation for multiple grains/crystallites/phases 2015-03-30 09:45:10 +00:00
Martin Diehl e0f1132a17 gfortran complaints: equal comparison of reals and external (MPI) functions 2015-03-29 12:54:13 +00:00
Martin Diehl 9e824d029e renamed to new names: spectralsolver and spectralfilter 2015-03-28 21:14:11 +00:00
Martin Diehl 1b2e4c0239 removed old solver 2015-03-28 17:41:40 +00:00
Martin Diehl 8a30021202 cleaned up core modul 2015-03-27 14:10:18 +00:00
Pratheek Shanthraj c8554a5641 corrected alignment of restart output byte offset. restart test working now 2015-03-27 07:17:24 +00:00
Pratheek Shanthraj 1ee81e74ea more patches to get the tests running again: 
- fixed increment counting in postResults to current output scheme

- corrected node coordinates calculation

- some work on restart (restart test still not running)
 2015-03-26 21:19:28 +00:00
Pratheek Shanthraj 86283359f1 PetscFinalize should be the last call before exiting 2015-03-26 13:43:18 +00:00
Pratheek Shanthraj a9cf8e8051 some patches to compile damask core module without fftw-mpi 2015-03-26 13:41:19 +00:00
Pratheek Shanthraj 2417877bdb dropped non petsc basic spectral solver 2015-03-25 16:08:41 +00:00
Pratheek Shanthraj d44fce4a76 Spectral solver now fully parallel (parallel IO, domain decomposition, FFTs and restart). Working but not extensively tested so please report bugs to me 2015-03-25 16:06:19 +00:00
Pratheek Shanthraj 37a7364a3e multi physics output now working for all solvers 2015-03-25 16:02:30 +00:00
Pratheek Shanthraj 2d4e7fb8e7 calculate dFi/dS (9x9 matrix inverse) for the analytic jacobian only when needed. should improve performance 2015-03-20 07:41:11 +00:00
Pratheek Shanthraj 2e44a846af need to keep track of temperature again since we can no longer get this from F_T 2015-03-18 18:03:18 +00:00
Pratheek Shanthraj c5a8543f17 updated petsc 2015-03-18 17:18:43 +00:00
Martin Diehl 4a0c1c1717 comments 2015-03-15 15:30:14 +00:00
Martin Diehl 12cb5e56fe fixed interpretation 'spectralOut' header. Seem to have been buggy all the time, but manifested itself only after keyword change. Also added ':' after keywords to be more robust 2015-03-15 15:29:38 +00:00
Martin Diehl ac3d82493e changed output keywords form resolution to grid and from dimension to size 2015-03-13 14:45:33 +00:00
Martin Diehl e4d70e971d cp is commercially pure ;) 2015-03-13 13:49:09 +00:00
Martin Diehl d8debc3396 updated reference geom files (used old names), updated numerics.config to reflect use of analytic jaco 2015-03-13 06:47:26 +00:00
Martin Diehl aa125bac5b aliases for grid and size not supported any more 2015-03-12 22:30:24 +00:00
Martin Diehl 00cba25a44 improved update of ip coordinates for spectral solver. 
do not need to create a new fftw plan all the time, using data already defined for the convolution
 2015-03-12 22:28:33 +00:00