added configure test to check if OpenMP is available (from petsc-dev-omp)Hg-commit: 7d3b0654647965f12166f840822aac58916263e4
#114662 update scotchHg-commit: d8b49725f826538781ae56a3cf8f5ec13fc59796
Added --with-ctetgen to turn on CTetGenHg-commit: dd04d12ab01c0f30a0f398b4635022034335a2e6
Fix operation for double on GPUHg-commit: 47e20042981ecac6ff0a9fd1e26cba3294eb93c9
backout 7d4d615e6b46 [it works for only 'hg pull -u' but not 'hg pull; hg update']Hg-commit: 11715e74c4123d17f15ce9c7f7de31255ae1b0bd
chain link buildsystem for 'hg pull'Hg-commit: 7d4d615e6b46c16b794a58a707ef043aca40fdb7
Fix for new variable coefficient FEMHg-commit: 521ac7d31c9a8e48c301686458de226499b29532
Define any extra PETSC_HAVE_* macros for CMake buildHg-commit: 68e0a24038a0845f56745b7e51a5a4542984fe5d
Add preliminary PAMI-based implementation of Iallreduce(), untestedHg-commit: 0c213f97859622665bffbe7ecf887b73517a0903
Fix small configure bugHg-commit: 2b865b17a0b6325f1837adcac667e0645293600d
Added in #defines from packages for JedHg-commit: ea90472d64db1862969b339ac100803676b7d563
Apple doesn't support __thread, use pthreads API pthread_get/setspecific() instead.Hg-commit: 758c4a646abfbe1bd113431b2081f15241d5ceaf
error out configure if using pthreadclasses and no __thread availableHg-commit: c7d6fd07395bc4fbe349637dce8ba662c7b65b73
remove duplicate definition of HAVE_PTHREADCLASSESHg-commit: c8b8601bfaeab286f9ce937c545e3650db931085
Renamed PETSC_ThreadLocal to PETSC_PTHREAD_LOCALHg-commit: 4df31eb1b3a41b153bf3bc18f26cb46c4bf6f76c
Made PetscStack to be thread local for pthreads version.Added configure test to define the thread local specifier PETSC_ThreadLocal to either __thread (Linux) or declspec(thread) (Windows)http://en
Made PetscStack to be thread local for pthreads version.Added configure test to define the thread local specifier PETSC_ThreadLocal to either __thread (Linux) or declspec(thread) (Windows)http://en.wikipedia.org/wiki/Thread-local_storageHg-commit: bf2991091f1a0ae7fada40fe84e42ea9c2be24d8
show more ...
added support for getting near null space vectors from Mat.Changed API for PCSetCoordinates to prepare for GAMG KKT solver.samll clean ups.Hg-commit: d5ce660d8f75cc2b9a446f9700efd17b7ec77a24
Fix for CUDA configureHg-commit: 814b16fb181ce5e6505aba258e9c860feeef83f0
CUDA now reports the compiler archHg-commit: 75fa2c74495232787bdd657eaca41973affba855
Added a flag to allow CUDA compiling without Thrust/Cusp- Until they are downloadable, this is necessaryHg-commit: 7ebba1f296aab4c06fa4a9ae5cf39e538e9d1be6
Made flexible scalar type for GPU assembly- ex52 tests still pass in singleHg-commit: cf64037655db7aab0ef1783e5e1afc0b186c4643
Can use xmmintrin.h when CUDA is supported, guard using __CUDACC__Hg-commit: 9f97546463cd207e5a200e458ee9b6c2b7ad3b94
removed listed run test that did not existfixed missing testexamples_X by unifying the name of X windows in the make files etc always as X not sometimes X11Hg-commit: 7f1b4c7c3dcb8b1842e10dc1e9136
removed listed run test that did not existfixed missing testexamples_X by unifying the name of X windows in the make files etc always as X not sometimes X11Hg-commit: 7f1b4c7c3dcb8b1842e10dc1e91364d5a41a2a26
merge from 32Hg-commit: 9b67cec165772d54219a7f98edc32421d226fa62
update parmetis [also metis tarball inplace] with the fix disabling building binaries.Hg-commit: a24dd87cb14e234fa57b6803cc86b30e81240e8d
1...<<31323334353637383940>>...77