camierjs
30acfd7f8f
Pull new Memory Manager to get most examples working
2018-12-18 16:05:49 -08:00
Tzanio
677b6f8aad
Merge branch 'master' into okina
...
Conflicts:
fem/bilinearform.hpp
fem/pbilinearform.hpp
makefile
2018-12-18 10:07:49 -08:00
camierjs
0a58311b39
Cleanup toward getting rid of Push/Pull
2018-12-14 15:17:03 -08:00
camierjs
450905d58a
Cleanup and mm rewrite
2018-12-12 17:30:13 -08:00
camierjs
961b20e711
Laghos Sedov GPU, same results
2018-12-10 20:35:53 -08:00
camierjs
4fcf5de2c4
make check with right x_gf
2018-12-07 13:26:55 -08:00
camierjs
f9ac906687
Merge branch 'okina-occa' into okina-laghos
2018-12-06 12:42:09 -08:00
camierjs
8d7ff525ec
Laghos problems 0, 1 on CPU, GPU
2018-12-04 15:59:54 -08:00
camierjs
8fbf274863
array deletes through mm
2018-11-30 12:20:35 -08:00
camierjs
4189c54a94
Address range check for Mesh => GetNodalValues => dot product with offset
2018-11-29 18:40:38 -08:00
camierjs
431b0e6bae
Add CUDA stub to cleanup the code
2018-11-07 17:00:56 -08:00
Jakub Cerveny
00227e56ad
Merge branch 'master' into laghos-amr-dev
2018-10-26 11:18:10 +02:00
Tzanio
b17ac8f4ad
make style
2018-10-12 17:52:47 -07:00
camierjs
944ef49eeb
Melding with master
2018-10-12 15:58:30 -07:00
camierjs
620f3dab7a
Back to serial GCC
2018-10-11 13:38:10 -07:00
camierjs
840480b205
Cleanup with master
2018-10-11 10:34:24 -07:00
camierjs
ba073e9521
Add LSolve USolve kernels for master to run
2018-10-11 09:56:51 -07:00
camierjs
ec6375ea94
Merge branch 'master' into okina
2018-10-11 09:22:42 -07:00
camierjs
f9fe41f8b1
Merge branch 'master' of github.com:mfem/mfem
2018-10-11 09:21:11 -07:00
camierjs
e71d5f0013
Cleanup & Run
2018-10-11 09:19:03 -07:00
camierjs
3b615c7188
Working on mesh->SetCurvature from GPU
2018-10-10 17:44:25 -07:00
camierjs
40cacec966
ex1 with Quad UniformRefinement CPU + GPU
2018-10-10 10:59:12 -07:00
camierjs
6304c2023e
ex1 converging
2018-10-09 19:04:24 -07:00
camierjs
ccb161cbd4
PABilinearForm + PADiffusionIntegrator
2018-10-09 11:12:50 -07:00
camierjs
e79d6eadcf
Init geom without prefixed SetNodalFESpace
2018-10-08 20:07:18 -07:00
camierjs
dd2fdb529e
Diffusion, up to Vector::operator-=
2018-10-05 17:53:39 -07:00
camierjs
5a828d3207
Starting to introduce diffusion kernels, maps & geom
2018-10-04 14:04:04 -07:00
camierjs
3eaf4beb60
Cleanup and moved memcpy to mm
2018-10-03 16:39:27 -07:00
camierjs
3b0496d4a5
Ex2 OK, doing SetNodalFESpace before switching to GPU
2018-10-02 18:40:32 -07:00
camierjs
98730a3f9b
Cleanup
2018-10-01 17:35:32 -07:00
camierjs
3e2bd3b5e4
ex1 CG ok
2018-10-01 12:25:45 -07:00
camierjs
85ef967a77
WIP BilinearForm::Assemble, elmat diffs
2018-10-01 12:02:14 -07:00
camierjs
4c072af530
ex1 working, switching to GPU for PCG
2018-09-27 13:07:41 -07:00
camierjs
35dd32e5fc
Up to SparseMatrix::AddMult
2018-09-26 17:08:19 -07:00
camierjs
2f6b4c2f07
WIP, *this, When to go to GPU?
2018-09-25 19:25:16 -07:00
camierjs
fb25fb6f82
RIP to get Segmentation faults
2018-09-25 16:51:33 -07:00
camierjs
e2f7584d7b
Init CFG, DBG, MEMCPY & MEMMNG
2018-09-25 16:40:31 -07:00
Jakub Červený
253e1dc320
Added missing assert in Vector::AddElementVector (based on similar function).
2018-03-16 16:49:05 +01:00
Stowell, Mark L
8adf1bf29c
make style
2018-01-26 20:48:27 -08:00
Stowell, Mark L
bae718680d
Merge remote-tracking branch 'origin/master' into prism-dev
...
# Conflicts:
# fem/fespace.cpp
# fem/pfespace.cpp
2018-01-26 20:44:48 -08:00
Stowell, Mark L
7eaf351332
Adding debug output to an error message
2018-01-19 15:12:42 -08:00
Veselin Dobrev
42ad503d6e
Define a shortcut for std::numeric_limits<double>::infinity() called
...
mfem::infinity() and replace (almost) all usage of the std version with
the shortcut.
In SparseMatrix::MakeRef, copy the field isSorted from the original
matrix.
2018-01-02 06:36:04 -08:00
Tzanio Kolev
d1bcacf419
Merge pull request #197 from gelever/tensor_segfault
...
Segfaults from copies [tensor_segfault]
2017-12-08 15:48:24 -08:00
Veselin Dobrev
6a9cca5f44
When copying data with Vector::operator=, skip the call to memcpy if
...
the source is the same as the destination.
2017-12-01 02:51:12 -08:00
Veselin Dobrev
7f0efa4ecb
Minor tweaks.
2017-11-28 17:40:50 -08:00
Veselin Dobrev
7db2768d31
In the copy constructors of classes Vector and DenseMatrix, prohibit
...
copying when the source object is invalid, i.e. it has nonzero size
but the data pointer is NULL.
In class DenseTensor, the copy constructor always uses deep copy.
In class DenseMatrixEigensystem, in the copy constructor, initialize
'ev' to not own its data.
In classes Vector, DenseMatrix, and DenseTensor, use std::memcpy to
copy the data arrays.
2017-11-28 17:22:28 -08:00
Tzanio Kolev
89ef05810f
Merge pull request #295 from mfem/xsdk-cpissues-dev
...
Compliance with the xSDK community policies [xsdk-cpissues-dev]
2017-11-08 17:13:54 -08:00
Tzanio Kolev
60a2cd5312
Merge pull request #314 from goxberry/vector-normlp-overflow-fix-dev
...
Make Vector::Normlp less prone to overflow [vector-normlp-overflow-fix-dev]
2017-11-08 13:49:25 -08:00
Veselin Dobrev
c31bd65e6d
Restore support for SUNDIALS v2.7.0 - the used version is "autodetected"
...
in the .cpp files that use SUNDIALS.
Updated INSTALL and CHANGELOG to reflect that both v2.7.0 and v3.0.0 of
SUNDIALS are supported.
In makefile, move the addition of the SUNDIALS and PETSc examples to the
examples run by 'make test' after the inclusion of config/config.mk.
2017-11-06 16:17:25 -08:00
Veselin Dobrev
0691b61184
Fix an issue in Vector::Norml2.
2017-11-02 12:49:36 -07:00
camierjs