Commit Graph
318 Commits
Author SHA1 Message Date
Stefano Zampini b8281fa01a Need to restore the usedevice flag when vector has been copied 2021-03-17 16:59:16 +03:00
Tom Stitt d462dcfed6 revert changes in test_sedov, array, and vector; registering and setting VALID_DEVICE in Memory::New(size, mt) to match existing behavior is enough 2021-03-05 16:22:18 -08:00
Tom Stitt a40a90a29f a few changes to get the tests working 2021-03-05 12:52:43 -08:00
Tom Stitt 61694679ea Update Array::Copy so it Writes the dest if the source is using
the device
Use DeviceMemoryType() instead of DEVICE in a cuda routine in
vector.cpp
Clean up some stuff in mem_manager.hpp
2021-03-02 15:35:02 -08:00
Adrien M. Bernede 04fb572359 Let's forget about 2020 2021-01-26 11:55:36 -08:00
camierjs 9f7720e69f Merge master in tmopt 2020-11-18 20:03:06 -08:00
Veselin DobrevandTzanio Kolev ea01d71fb0 Update linalg/vector.cpp
Co-authored-by: Tzanio Kolev <tzanio@llnl.gov>
2020-11-16 15:09:58 -08:00
Veselin DobrevandTzanio Kolev 1b91c0c678 Update linalg/vector.cpp
Co-authored-by: Tzanio Kolev <tzanio@llnl.gov>
2020-11-16 15:09:51 -08:00
Veselin Dobrev 65f2add11d Add a deterministic way of computing the dot product with OpenMP 2020-11-15 21:18:31 -08:00
camierjs 5cbffa43c1 Merge master in tmopt 2020-11-02 10:27:05 -08:00
Veselin Dobrev 97ad695664 In HIP kernels, use MFEM_HIP_BLOCKS instead of MFEM_CUDA_BLOCKS.
There is no change in the compiled code since both macros are
defined to be 256.
2020-10-29 06:18:56 -07:00
Will Pazner 62f4c49ac6 Remove conditional compilation of resetting failbit for subnormals 2020-10-28 11:42:54 -07:00
Will Pazner d15a402464 Fix GLVis issue with subnormals Apple Clang
Zero out subnormal numbers when writing a Vector to an output stream. When
reading a Vector from an input stream with Apple Clang, ignore "out of range"
errors, which can be caused by (correctly) parsing subnormal numbers.
2020-10-27 15:05:42 -07:00
camierjs ffba287f4a Merge master in tmopt 2020-10-15 11:40:40 -07:00
Balos, Cody Joe 311eb40ac3 remove TODO comments about ToNVector function 2020-09-03 11:43:01 -07:00
camierjs fd3d8270f2 Merge master in tmopt 2020-09-01 12:39:28 -07:00
Balos, Cody Joe c7ead47f1d remove unused includes 2020-08-26 13:20:36 -07:00
Balos, Cody Joe 937db1a293 Add support for SUNDIALS GPU features
Add support for SUNDIALS CUDA features. These changes will be easily
extendable to support future GPU features in SUNDIALS, e.g. HIP
features. To keep things clean, this creates a SundialsNVector
class to interface with the SUNDIALS N_Vector. This class inherits
from the mfem::Vector class, and follows the pattern used in MFEM
interfaces to other libraries such as PETSc and hypre.
2020-08-26 13:20:36 -07:00
camierjs eaf2257263 Rename Backend::DEBUG to Backend::DEBUG_DEVICE 2020-08-21 14:34:15 -07:00
camierjs 55c3321c3b Merge branch 'master' into tmopt 2020-07-23 07:26:59 -07:00
camierjs d036fd8b1f Changelog, meld & revert MFEM_MMU_PROTECT_ERROR 2020-07-20 14:32:04 -07:00
camierjs 04eddb884c Missing PA.setup_Grad 2020-07-15 10:25:42 -07:00
camierjs 57c3853fb6 Merge fix, bump branch_max_commits to 250 and HostRead in Vector::Max 2020-07-15 09:40:35 -07:00
camierjs 08f5abc6b2 Merge branch 'master' into tmopt 2020-07-15 09:31:32 -07:00
Tzanio Kolev 01476b98cb Merge branch 'master' into CVODESSolver-dev 2020-07-14 15:26:30 -07:00
camierjs 223c47fae2 dist.Randomize(seed) 2020-06-26 12:28:09 -07:00
Ido Akkerman 0b5ee4ea04 Add missing add vector member function 2020-06-24 17:20:20 +02:00
Jonathan Wong 85349d3a95 Properly intialize quad integration vectors.
Fixed memory errors in ToNVector and initialize in CVODESSolver::RHSB
2020-06-22 23:34:43 -07:00
camierjs 134436f064 Vector::operator* of size 0 fix 2020-06-09 17:54:48 -07:00
Jonathan Wong 023e63dfb1 Reran astyle with astyle 2.05.1 2020-03-30 12:51:23 -07:00
Jonathan Wong e63ac279c1 Merge branch 'master' of github.com:mfem/mfem into CVODESSolver-dev
Fixed merge conflict in linalg/operator.hpp.
Cleaned up linalg/sundials.cpp a little bit.
Reverted defaults.mk change.
2020-03-30 12:49:40 -07:00
Jonathan Wong 8b7f15cdca revert astyle changes 2020-03-30 12:45:43 -07:00
Jonathan Wong 3f4a6ce0d2 ran astyle 2020-03-25 12:58:46 -07:00
Jonathan Wong 4ee2b18d97 Changed from SetN_Vector to .ToNVector. Fixed implementation as suggested by gardner
Added an extra argument to Vector::ToNVector() to better handle parallel vectors
2020-03-25 12:58:33 -07:00
William F Godoy 9c379891bf Resolving conflicts with latest master 2020-03-18 09:41:46 -04:00
Tzanio f1e04e7282 Merge branch 'master' into mfem-4.1-dev
Conflicts:
	fem/quadinterpolator.cpp
	fem/restriction.cpp
2020-03-10 16:56:01 -07:00
Tzanio 05125e19d5 Removed a space 2020-03-09 15:36:41 -07:00
camierjs a3bb229bc9 Templated code for kernels 2020-03-05 18:46:07 -08:00
camierjs 28f3373ed8 Switch blas to kernels 2020-03-05 16:52:49 -08:00
camierjs a79cadc840 Merge branch 'master' into laghos 2020-03-05 16:37:55 -08:00
Tzanio 2e37ba7d97 Merge branch 'master' into yohann/dg
Conflicts:
x	fem/bilinearform_ext.cpp
2020-03-04 20:42:09 -08:00
Tzanio 58872c9e7b Merge branch 'master' into okina-mmu
Conflicts:
	fem/datacollection.cpp
	tests/unit/CMakeLists.txt
2020-03-04 15:19:33 -08:00
Tzanio ca55ef0217 The release number for mfem-4.1 is LLNL-CODE-806117 2020-03-02 17:18:46 -08:00
camierjs 602dce5818 CUDA => DEVICE in hipVectorMin and hipVectorDot 2020-02-25 15:54:59 -08:00
Tzanio 2caa222edf A number of changes to switch the code from LGPL-2.1 to BSD-3:
- Updated LICENSE, README, CHANGELOG and CONTRIBUTING.md
- Removed COPYRIGHT, added NOTICE
- Updated the headers in all source files, makefiles, etc.
2020-02-23 11:16:46 -08:00
camierjs 8f30f76e85 Move DEVICE macros to cuda header
Use stream sync
2020-02-20 15:08:49 -08:00
Yohann e2858823e1 Merge branch 'master' into yohann/dg 2020-01-13 15:16:15 -08:00
camierjs 371d7c411b Now use blas:: code instead of densemat & cleanup 2020-01-10 12:12:53 -08:00
Yohann Dudouit a563c5be9e Fix a bug in optimized version of Convection integrator. 2020-01-08 13:41:18 -08:00
camierjs 6a61e78bf9 Merge branch 'master' into okina-mmu 2020-01-06 11:01:36 -08:00