a3a3388855
decouple DAMASK default integer from MPI default integer
2022-01-13 10:25:45 +01:00
8223dc7fa7
polishing
...
MPI, HDF5, PETSc, and DAMASK might have different integer kinds ..
2022-01-13 08:17:31 +01:00
3fb5bd459c
pInt leftovers
2022-01-13 07:44:15 +01:00
236f0297ac
test added
2022-01-12 20:42:37 +01:00
e80d91e30a
thermal restart (WIP)
2022-01-11 16:11:18 +01:00
1e965c42b7
don't rely on ML
...
avoid dependencies to external packages as much as possible
2021-12-27 17:44:22 +01:00
1ddf1e5694
support for PETSc with 64bit integers
...
compiles, but untested
2021-12-21 23:53:46 +01:00
6ba2a08e5a
easier to read
2021-12-11 11:50:40 +01:00
f51633d43a
forall is deprecated
...
do concurrent is the successor but ifort had problems and generated
faulty code
2021-12-11 09:01:42 +01:00
7d7d0c2659
only local variable are good variables
2021-12-11 08:49:30 +01:00
2f067b544e
use variables, not descriptors
2021-12-06 07:55:13 +01:00
96e4cb591c
Merge branch 'integer-exponents' into 'development'
...
Using integer exponent
See merge request damask/DAMASK!467
2021-11-29 07:32:04 +00:00
a531b7ccae
transitioned remaining real exponents to int
2021-11-28 12:46:26 -05:00
ccd6e44b6b
Merge remote-tracking branch 'origin/development' into 134-output_none
2021-11-27 19:17:27 +01:00
bfc6b69ee2
integer exponents: faster and shorter
2021-11-25 20:52:22 +01:00
43ae4983ed
fixed misaligned/too short "writing..." statements
2021-11-15 12:58:59 -05:00
69843d0833
Merge branch 'development' into pretty-print-init
2021-11-15 12:39:14 -05:00
da9fdf53d2
consistent indentation and line-spacings in reporting
2021-11-15 12:35:44 -05:00
64d70d9be3
user option to have no output
...
f_out = none
2021-11-15 14:39:29 +01:00
b7ad5b3167
'standard' style
2021-11-10 20:53:20 +01:00
186b688b04
only look for opening part of <CellData> tag
2021-11-09 15:10:19 -05:00
256c48831e
better readable
2021-10-19 23:08:21 +02:00
29431eb8c5
same reporting as in python
2021-09-12 21:55:14 +02:00
949eea1624
bugfix: did not work with VTK 8
2021-09-01 11:07:39 +02:00
4160c4fdb4
fix for parallel HDF5
...
if filters are applied, writing from one process does not work if the
file is opened for parallel write
2021-08-15 13:26:15 +02:00
f75235f6a9
Merge branch 'more-flexible-L' into 'development'
...
more flexibility for the L in the load case
See merge request damask/DAMASK!420
2021-08-09 21:27:13 +00:00
044a048944
taking care of corner cases (e.g. restart)
...
adjusting tests to take care of new 'setup' group
2021-08-01 22:46:11 +02:00
b9d4eb23cc
only rank 0 reads file for MPI
2021-07-27 08:54:17 +02:00
812b0f07f5
read file only once (per process)
2021-07-27 08:35:52 +02:00
ddb0429a1d
store load case (full reproducibility for grid solver)
2021-07-27 07:57:04 +02:00
e01d271ee4
store geometry (for full reproducibility)
2021-07-27 07:43:35 +02:00
574cfd7034
Merge remote-tracking branch 'origin/development' into more-flexible-L
2021-07-22 00:20:33 +02:00
9d349f8a7c
symbolic notation
2021-07-21 06:50:28 +02:00
6f19113072
L, P, F, etc. are second order tensors
2021-07-21 06:19:04 +02:00
85735605f8
more flexibility for the L in the load case
...
Note that mixed boundary conditions for L introduce an ambiguity.
Consider:
L = [[1.0, x, x],
[ 0, 0, 0],
[ 0, 0, 0]]
P = [[x, 0, 0],
[x, x, x],
[x, x, x]]
What we need is F^(n+1)=F_dot^(n+1) x Delta_t, where F_dot^(n+1) is
F_dot^(n+1)_ij = L_ik F^n_kj.
So component F_11 has contributions from L_12 and L_13. We first assume
L_12=L_13=0 and then choose F^(n+1)_12 and F^(n+1)_13 to get
P_12=P_13=0. This implicitly gives a solution for L_12 and L_13, which
is however only one out of infinitely many.
2021-07-20 07:10:28 +02:00
03b7532cc5
numpy.MaskedArray behavior
2021-07-19 23:27:10 +02:00
1c1dc9383e
symbolic names
2021-07-19 22:30:20 +02:00
f9edeb40a5
descriptive names
2021-07-17 11:50:21 +02:00
ed6b1be352
solver handles terminally ill
2021-07-16 20:43:08 +02:00
2a84aa7ae4
obvious, no need for comment
2021-07-16 20:32:21 +02:00
5f78f1753c
split up thermal
...
only for grid at the moment
2021-07-16 18:03:38 +02:00
3f0eafd640
first step towards separating of mechanics, thermal, and damage
2021-07-16 17:53:11 +02:00
2f1fa8292b
unify style to majority of occurences
2021-07-16 10:30:45 +02:00
136a4b1377
PETSc defines are rather complicated
...
now mpi_f08 can be used on newer PETSc installations if old MPI modules
are not exposed
2021-07-09 18:48:25 +02:00
637f78bd52
old name (for PETSc < 3.15)
2021-07-09 14:50:29 +02:00
8e75e87ad9
Merge branch 'MPI_F08' into polishing-for-beta
2021-07-09 11:32:32 +02:00
139f2c177a
use MPI_f08 if possible
...
most PETSc installations provide outdated MPI (f90 version)
MPI_COMM_WORLD is now of derived type (Fortran 08 style)
PETSC_COMM_WORLD is the plain integer (f90 style) alias.
Note that HDF5 is assumed to have f90 interfaces
2021-07-08 16:27:37 +02:00
58bc6e2ba6
avoid chained inclusions
2021-07-08 14:27:04 +02:00
a13e02da44
easier to follow
2021-07-06 22:33:12 +02:00
06e76c1a81
consistent default absolute tolerances
...
better use conservative values, users can easily relax if needed
2021-06-25 10:13:03 +02:00
ae9a48c124
polishing
2021-06-24 15:23:24 +02:00
1bfbd30ae2
polishing
2021-06-15 19:53:05 +02:00
218e6a79a8
VTK image data is the appropriate type, not VTK rectilinear grid
...
FFTs require constant spacing in all three directions, this is
guaranteed by the vtkImageData but not by vtkRectilinearGrid
2021-06-15 19:02:26 +02:00
3705f1d3a2
F,P, ... in loadcase are 2nd order tensors
2021-06-03 08:21:16 +02:00
5d0fc4fca3
more meaningful order
...
and intent(out) variables for read are at the front
2021-06-01 16:46:24 +02:00
d1a6607782
Merge branch 'load2Dtensor' into 'development'
...
support for 2D tensor in load case
See merge request damask/DAMASK!391
2021-05-29 13:21:47 +00:00
adb1e51e24
support for 2D tensor in load case
...
currently optional, but should become mandatory after a transition
period
2021-05-26 08:04:14 +02:00
d2855913b5
new mappings
2021-05-22 11:48:42 +02:00
c8cdd7e622
better numerical parameters for thermal solver/PETSc
...
works for realistic values for thermal problems
parameters copied from spectral_damage
2021-05-05 18:14:53 +02:00
887524bcc1
polishing
2021-04-11 16:09:29 +02:00
b67e837548
needs broadcast, otherwise rank>0 fail
2021-04-11 09:58:40 +02:00
547f2ffa69
cleaning
2021-04-11 09:46:11 +02:00
5f608ed572
only 2 and 3 dimension can be 1
2021-04-09 08:25:30 +02:00
1b89032086
names as in DAMASK paper
2021-04-08 23:45:26 +02:00
c4765d3742
following paper
2021-04-08 13:31:21 +02:00
d59051f576
systematic names
2021-04-07 22:42:10 +02:00
0fc7f66ef8
consistent names
2021-04-07 21:25:11 +02:00
c53927ad6f
not needed
2021-04-07 08:47:46 +02:00
1fbf14c148
encapsulation and namespace-like names
2021-04-07 07:53:24 +02:00
2e96fcf768
Merge branch 'development' into spring-cleaning
2021-03-29 07:26:17 +02:00
710f5b74b2
Merge remote-tracking branch 'origin/development' into 2d-table-yaml
2021-03-29 00:00:39 +02:00
0072ebfa64
polishing
2021-03-27 23:17:04 +01:00
7072ab0984
non-converging simulation is an error
2021-03-27 17:07:36 +01:00
6a2c107723
consistent names for debugging
2021-03-26 12:34:16 +01:00
4cc0dedf9f
Merge branch 'avoid_data_copy_restart_MPI' into spring-cleaning
2021-03-26 08:58:44 +01:00
7320120c5d
Merge branch 'development' into avoid_data_copy_restart_MPI
2021-03-26 08:58:03 +01:00
a72e2db678
Merge branch 'development' into output-rename
2021-03-25 21:03:20 +01:00
3b392281e4
consistent naming in HDF5 output
2021-03-25 19:22:59 +01:00
fc172921fb
unified citation style continued
2021-03-19 10:41:47 +01:00
07d79ef4ba
unifying style
2021-03-17 11:25:21 +01:00
4912342b1b
added missing arguments
2021-03-15 11:46:30 +01:00
a59af55f1a
read data by one process and broadcast it
2021-03-15 10:58:59 +01:00
ce91537b0f
get_asXXs --> get_as1dXX
2021-03-11 18:40:16 +01:00
adcb24d2e1
write data of average quantities non parallel
2021-03-10 16:33:02 +01:00
e4271537c5
syntax polishing; use of YAML defaults to avoid if%contains
2021-03-04 15:14:16 -05:00
2f68c43755
new style configuration for damage
2021-03-01 06:16:16 +01:00
8af0c8dbc3
using initial temperature from load case file
2021-02-28 21:51:07 +01:00
0cde43198f
modernizing:
...
- 'pass' for dummy thermal homogenization
- setting temperature in load case
2021-02-28 20:49:51 +01:00
ef543a5b49
thermal solver is defined in load case
...
not in numerics.yaml
2021-02-28 20:09:08 +01:00
b2fea6b149
solver not specific to load case number
2021-02-28 19:24:44 +01:00
4dd99d4c39
solver is selected in load case, not numerics.yaml
2021-02-28 19:13:20 +01:00
9679c5403e
continue with cell based map (solver + homog)
2021-02-23 15:44:58 +01:00
5348305ac8
restart key part of loadstep and not time discretization
2021-02-22 14:42:54 +01:00
4d5e5cfb70
Merge branch 'fix-IntelMPI' into 'development'
...
Bugfix for access of unallocated variable
See merge request damask/DAMASK!342
2021-02-22 12:40:37 +00:00
d54e49e3bc
restore functionality to write non-parallel
...
not needed at the moment, but in general useful. If PETSc = parallel
should always hold, we can simplify much more
2021-02-22 13:37:21 +01:00
6c7201610e
Bugfix for access of unallocated variable
...
IntelMPI seems to access sendbuf for root!=0 in MPI_Scatterv
2021-02-17 17:56:39 +01:00
4eb2a981ca
keeping variables local
2021-02-12 15:31:43 +01:00
e855083964
systematic names
2021-02-11 14:19:04 +01:00
8048d69a97
[skip ci] logic inverted
2021-02-05 16:59:28 +01:00
ba2e8b3c63
variable name follows label
2021-02-05 11:57:18 +01:00
a1a7a339d4
specify solver,initial and boundary conditions in load file
2021-02-05 11:20:28 +01:00
2e173b4b7b
'estimate_rate' better describes what we want to do than 'drop_guessing'
...
true by default
2021-02-04 19:17:32 +01:00
b58465415b
store damage parameter like temperature
2021-01-24 18:20:47 +01:00
26c7969837
not needed anymore
2021-01-24 17:12:34 +01:00
c2ae2c919b
use new structure
2021-01-24 15:19:57 +01:00
e22d76be9e
using new structure for thermal
2021-01-24 13:26:01 +01:00
4f059910ab
fix broken statistics reporting in case of multi-physics
2021-01-24 10:39:32 +01:00
6fe8077b81
Merge remote-tracking branch 'origin/separate-damage' into development
2021-01-22 14:25:32 +01:00
da93031cbc
Merge branch 'signal_input_handling_1' into 'development'
...
Enabling signal input handling in DAMASK
See merge request damask/DAMASK!325
2021-01-22 14:07:09 +01:00
32bb0d8c6e
new variables for damage
2021-01-20 21:37:43 +01:00
a2b5178b45
polishing
2021-01-17 21:56:19 +01:00
8af53472ee
Merge remote-tracking branch 'origin/development' into signal_input_handling_1
2021-01-17 16:52:48 +01:00
d2ae859b1c
use new structure
2021-01-17 15:26:01 +01:00
6b452a73bb
not needed
2021-01-17 12:45:28 +01:00
9b698e78c1
standard name
2021-01-17 12:35:47 +01:00
b5bfb1dba9
tangent is not included anymore
2021-01-11 16:34:43 +01:00
350466dd5f
not needed
2021-01-08 09:01:38 +01:00
e8ea815d92
simplified
2020-12-27 11:48:02 +01:00
0a3d43e0fc
only one process gets the signal: use OR, not AND
2020-12-17 15:17:20 +01:00
d83e8dac59
need to share variable before if
2020-12-17 13:20:18 +01:00
da1b006315
MPI communication for SIGUSRXXX
2020-12-17 12:47:33 +01:00
3884549e19
store field variables as 1D array
...
first step of simplifying layout:
1) Solver translates from ip,el tuple (FEM) or cells(1),cells(2),cells(3) triple to list.
2) DAMASK iterates over all points
3) homogenization knows mapping (point,constituent) -> (instance,member)
2020-12-16 21:56:02 +01:00
f8756ad95a
simplifying
...
no extral handling for homogeneous temperature (the memory that was
saved was consumed by the extra mapping)
2020-12-15 19:05:01 +01:00
afbcb82502
Merge remote-tracking branch 'origin/development' into misc-improvements
2020-12-15 10:32:59 +01:00
2dd520b4a2
P_aim should be independent from P_av
...
P_av is not defined after restart or cutback. Restart with change of
load case is probably still an issue
2020-12-15 08:06:50 +01:00
8fbadef524
print instead of write
2020-12-14 22:37:14 +01:00
768d139768
reset the signal terms to false after task is done
2020-12-14 09:34:35 +01:00
40bd5db361
signals enabling writing results,restart and termination
2020-12-14 09:17:15 +01:00
a881e6e57a
Merge remote-tracking branch 'origin/development' into misc-improvements
2020-12-10 23:41:59 +01:00
97ee7e6ee5
not needed anymore
2020-12-07 17:26:50 +01:00
78a246b44a
avoid constant reallocation, it is slow for large vtr files
2020-12-05 12:46:48 +01:00
42eb802126
not needed anymore
2020-12-05 11:29:23 +01:00
ac0a20696c
rename: grid -> cells
2020-12-03 22:16:34 +01:00
80c8c9e51e
fixing use for new Intel Fortran
...
visibility of use associated entities can now be limited
2020-11-28 07:23:55 +01:00
5bed0787ed
replace log scaling with geometric regression scaling
...
more flexibility in time discretization via parameter 'r'.
Also, changed variable names to reflect names in loadcase.yaml
2020-11-25 18:56:03 +01:00
2157bf50cd
mech --> mechanics
2020-11-17 22:21:35 +01:00
523a0979eb
Merge branch 'mpi-single-process-read' into 'development'
...
MPI single process read of grid
See merge request damask/DAMASK!277
2020-11-17 17:00:09 +01:00
c16ad845bc
Merge branch 'modernize-testing' into 'development'
...
Modernize testing
See merge request damask/DAMASK!280
2020-11-16 20:10:38 +01:00
cf3a28952a
fixed tests
2020-11-16 16:38:17 +01:00
0bd8a410ad
check all errors, avoid intermediate wait
2020-11-16 09:22:55 +01:00
e419eea507
read vtr file only by process 1
2020-11-12 14:40:25 +01:00
bb471cb923
it's a process, not a thread
2020-11-12 07:51:07 +01:00
cd7ada0da9
use Fortran internals for error handling
2020-11-11 20:09:14 +01:00
3dd5eaf1c1
clean exit with stack trace
2020-11-11 14:06:21 +01:00
5531f892c3
z = 0 for invalid number of processes
...
e.g. using 24 processes for a 128x128x128 grid would need 21 processes
with z=6 and 1 process with z=2
2020-11-11 12:47:13 +01:00
429c226008
IO is not inherited when using ifort 19
...
this is intented, public modules should only bring their own functions
and variables into global scope
2020-11-11 12:19:39 +01:00
b46a25a7d2
C capitalization
2020-11-11 11:47:23 +01:00
9119254210
materialpoint => homogenization
...
especially as prefix for global data (clear name spaces)
2020-10-24 17:26:42 +02:00
a1ab526456
avoid global variables
2020-10-24 16:01:56 +02:00
1c13737935
better readable
2020-10-21 21:12:55 +02:00
Martin Diehl