bugfix for single precision build: replace zeropivot=1.e-12 with 100.0*PETSC_MACHINE_EPSILONThe former hangs in src/ksp/ksp/examples/tutorials/runex32Hg-commit: 912e0bd91134ba3086e9b01a03649817fd7
bugfix for single precision build: replace zeropivot=1.e-12 with 100.0*PETSC_MACHINE_EPSILONThe former hangs in src/ksp/ksp/examples/tutorials/runex32Hg-commit: 912e0bd91134ba3086e9b01a03649817fd72bf9d
show more ...
Added flag in pc data structure for reusing the numerical factorization.Hg-commit: d07bb2ec9459eb209ab7a15d1357e143ad76cfb7
Use #include <header.h> instead of #include "header.h" when there is no intent to search the current directory for header.hHg-commit: bd7216e80cc035071a5136364ab9d04bf9f41a07
removed #define PETSCXXX_DLL from top of all source filesHg-commit: b784aa18ddcf5337d4011d34f0fa008681695ac1
more PCReset_XXX() doneKSPDestroy() now zeros the pointerHg-commit: 734a7781b5d7d5bb9df4814d2e579008497bc8d8
ISDestroy() now zeros the pointeradded a couple more PCRest_Hg-commit: 646656e619383798fae72855f3a4f260d7fdf07a
fixed .bib entries with Gail's correctionsremoved all checks of nonzero pointer before call to PetscFree and setting to zero of pointer after PetscFree, since PetscFree() does that all automatically
fixed .bib entries with Gail's correctionsremoved all checks of nonzero pointer before call to PetscFree and setting to zero of pointer after PetscFree, since PetscFree() does that all automaticallyHg-commit: 0f5bef9ea1fcd7ad8418fab6181ef990ea35c1db
MergeHg-commit: 674bc8187eb8d0423d5e5713bd4ba8a46a2b8020
Now requirepackage is only used for proper packages, functions use requiresfunction, and the rest use requiresdefineHg-commit: dd65fd3461c02753386917514e6ebe12d345a30e
added MPI to BasicVersion of streams to get some measure of parallel memory bandwidthfixed OpenMPI version of streams to actually build with MPIHg-commit: 190640f459cfcd17b88515b33cbfb760a32096c2
update MatSeqAIJFindZeroRows() to MatFindNonzeroRows()Hg-commit: fbda5d24d43b60060a19e9257c1b0ad376959dc0
changed PCFactorSetMatSolverPackage() to NOT require the Mat before being called, now it just keeps a copy of the package name and uses it lateradded PCFactorSetUpMatSolverPackage() actually calls M
changed PCFactorSetMatSolverPackage() to NOT require the Mat before being called, now it just keeps a copy of the package name and uses it lateradded PCFactorSetUpMatSolverPackage() actually calls MatGetFactor() so there now exists a factor matrix and PCFactorGetMatrix() can be called and then options set from the program.Hg-commit: 7f7727b0030448abcd98a3d596d64ed83143f965
changed Matlab spelling to MATLAB some general MATLAB cleanupHg-commit: efcb24f191dfedc98b283d9e63e7859032e8ef35
moved definition of PETSC_STDCALL to private/fortranimpl.h where it belongsremoved all DLL export/import stuff since it was never used and never will beremoved MatPermuteSparsify() since it sucked.
moved definition of PETSC_STDCALL to private/fortranimpl.h where it belongsremoved all DLL export/import stuff since it was never used and never will beremoved MatPermuteSparsify() since it sucked.Hg-commit: 72f7a5e60333f6c9780d56fcce45657940d16b13
makefile and headers do not need to be executableHg-commit: ba5f734acfa7080b87c6371551d8590561ef6909
Some extern C compliance, can't test properly because my compilers don't careHg-commit: db6fde48a972e3cd2b492e83f0e09c82d6cb4e95
PetscOptionsTruth -> PetscOptionsBool and relatedHg-commit: dfea0d194d94180043ddc1c2d9d074a0b8717c77
finished converting to use PetscTryMethod() and PetscUseMethod() whenever possible.Hg-commit: 37cf7fcea5d871042d238d3c6a46bb0914b2136d
Changed PetscTruth to PetscBool and PETSC_TRUTH to PETSC_BOOLnote this requires a new sowing versionHg-commit: 8b4c4277f05e3fb874633288193dab1105bf142a
added PCApplySymmetricRight/Left for ILUHg-commit: e6f301fecea255bcbc554e93efeebcc35920dc80
#49218 compile fixHg-commit: 89d9e2b11ecac94a17ac89ae0a211bd87752c5a1
completed PetscValidLogicallyCollectiveXXX() routines and usageHg-commit: 374cfa6c3abbd90ea5b27fc02813ca73ae37b846
started adding Logically or Neighbor-wise Collective to manual pagesHg-commit: 7a478f9b05148e2efabc037328838c5851772610
fixed bug in ilu preconditioner that created factor matrix ONLY if it already existed, missing !Hg-commit: 1daf83b6f94f37632b9f3b2adba6c2a5e3cca62d
removed unneeded and illegal fact temp variablesHg-commit: b0e3cddac8ff386121e3367e72d44fe7a0d69ec3
1...<<11121314151617