Don't generate an error if the user does not free all objects while using SuperLU_DIST.Commit-type: error-checking/spend 10mReported-by: "Zhang, Hong" <hzhang@mcs.anl.gov>Reported-by: "Xiaoye S.
Don't generate an error if the user does not free all objects while using SuperLU_DIST.Commit-type: error-checking/spend 10mReported-by: "Zhang, Hong" <hzhang@mcs.anl.gov>Reported-by: "Xiaoye S. Li" <xsli@lbl.gov>Reported-by: Jin Chen <jchen@pppl.gov>
show more ...
Allow resusing SuperLU_DIST communicator and grid for multiple new SuperLU_DIST matricesNote: When possible, same size matrix, same parallel distribution, same nonzero patternone should reuse the
Allow resusing SuperLU_DIST communicator and grid for multiple new SuperLU_DIST matricesNote: When possible, same size matrix, same parallel distribution, same nonzero patternone should reuse the matrix and the KSP for new solvers, this allows reusing portions of thepreconditioner assembly process.Commit-type: optimizationFunded-by: ECP/spend 2.2hReported-by: "Xiaoye S. Li" <xsli@lbl.gov>
Single petscdir.mk
superlu_dist: update MATSOLVERSUPERLU_DIST manpage to list METIS_AT_PLUS_A and PARMETIS column permutation optionsReported-by: Yang Liu <liuyangzhuan@lbl.gov>
build: add more support for not needing to set PETSC_DIR yet build with PETSc, also support for GNU make to directly use "make test" instead of "make -f gmakefile.test"Still needs -include ../petsc
build: add more support for not needing to set PETSC_DIR yet build with PETSc, also support for GNU make to directly use "make test" instead of "make -f gmakefile.test"Still needs -include ../petscdir.mk in many of the makefilesStill needs the link for ../petscdir.mk in many directoriesGNUMakefile: call legacy makefile recursively instead of including itGNUMakefile: promote default "all" target; "libs" builds libpetsc*GNUMakefile: If gnumake < 4.2 prefer makefile wrapper interface over direct gmakefile interface to avoid the following warningsgmakefile.test:92: arch-ci-linux-cxx-cmplx-pkgs-64idx/tests/testfiles: No such file or directorygmakefile:67: arch-ci-linux-cxx-cmplx-pkgs-64idx/lib/petsc/conf/files: No such file or directoryCommit-type: testing-fix, feature, usabilityThanks-to: Jed Brown jed@jedbrown.org
Promote examples/{tests,tutorials}/ to {tests,tutorials}/This shortens paths and improves consistency between test target namesand paths to the source and output files. Most of the work wascompl
Promote examples/{tests,tutorials}/ to {tests,tutorials}/This shortens paths and improves consistency between test target namesand paths to the source and output files. Most of the work wascompleted by this script, followed by mild cleanup of nonconformingcases.for makefile in `git ls-files 'src/*makefile'`; do if rg -q 'DIRS.*\bexamples\b' $makefile; then base=$(dirname $makefile) dirs=$(cd $base/examples && ls -d tests tutorials 2>/dev/null | xargs echo) perl -pi -e "s#^(DIRS.*)\bexamples\b#\1${dirs}#" $makefile git rm $base/examples/makefile for t in $dirs; do git mv $base/examples/$t $base/ perl -pi -e "s#^(LOCDIR[[:space:]]*=).*#\1 $base/$t/#" $base/$t done fidonegit grep -l -E -z 'examples/(tutorials|tests)' | xargs -0 perl -pi -e 's#examples/(tutorials|tests)#\1#g'git checkout @ \ src/docs/website/documentation/changes/ \ src/benchmarks/results/
SuperLU_DIST: update to version 6.3.0SuperLU_DIST.py GPU code currently requires OpenMPRemove -download-superlu_dist-gpu option and use --with-cuda=1Keep backward compatibility with older super
SuperLU_DIST: update to version 6.3.0SuperLU_DIST.py GPU code currently requires OpenMPRemove -download-superlu_dist-gpu option and use --with-cuda=1Keep backward compatibility with older superlu_dist versions.
MATSUPERLUDIST: fix MPIAIJ with commsize 1 case
Enable MatMatSolve(F,RHS,RHS) for some packagesRequested-by: Jaysaval, Piyoosh <Piyoosh.Jaysaval@pnnl.gov>
Fix spelling mistake.Reported-by: Matt Knepley <knepley@gmail.com>
-package-prefix-hash=directory will prefix the external packages based on a hash of the config directoryand configure argumentsBased on a suggestion by Satish BalayAll PETSc builds go as usual f
-package-prefix-hash=directory will prefix the external packages based on a hash of the config directoryand configure argumentsBased on a suggestion by Satish BalayAll PETSc builds go as usual for testing into PETSC_DIR/PETSC_ARCH.Running with a different branch with the same config will generate the same hash and hencereuse the previous external package builds.Any change in the hash results in a COMPLETELY new external packages prefixIf one assumes that most merge requests do not involve changes to config this will result in a dramaticsavings in configure time plus no longer require someone to manually manage the pre-building of externallibraries and keeping them up to date. The same directory can be used for tests for maint and master sincemaint will generate different hashes and thus have its own copies of the external libraries.Also supports --arch-hash that sets PETSC_ARCH based on the same config hash.Commit-type: testing-fix, featureThanks-to: Satish Balay <balay@mcs.anl.gov>
Remove testing and inserting into petscconf.h items that are not actually used by PETSc1) PETSC_HAVE_LIBXXXX - which was rarely used be careful with the package libpng and libjpeg since they hav
Remove testing and inserting into petscconf.h items that are not actually used by PETSc1) PETSC_HAVE_LIBXXXX - which was rarely used be careful with the package libpng and libjpeg since they have lib in the name of the package2) various system include files that are never used or always exist: for example stdlib.h3) various system functions that are never used or always exist4) fixes for requires for MUMPS and SuperLU_DIST when dependent packages are installed or not installed (unrelated to the rest of this pull request)5) packages that always exist such as PETSC_HAVE_BLASLAPACK, or are not used by PETSc such as PETSC_HAVE_NETCFD6) remove a couple of uses of HAVE_LIB* in the code that were not needed by adjusting the configure code slightly7) remove all the #if guards for each entry in petscconf.h since petscconf.h already has a guard and the values are never defined else where the extra guards just make the file clutteredFor a build with about 10 external packages this reduced the size of petscconf.h from 1236 lines to 828/4 around 220 entries.Commit-type: style-fix, cleanupReported-by: Jed Brown <jed@jedbrown.org
Add PetscArraycmp(), PetscArraycpy(), PetscArrayzero(), PetscArraymove(), update code to use new formsReplace a couple multi-mallocs with a single mallocReplace backwards loops in MatSetValues_*()
Add PetscArraycmp(), PetscArraycpy(), PetscArrayzero(), PetscArraymove(), update code to use new formsReplace a couple multi-mallocs with a single mallocReplace backwards loops in MatSetValues_*() with PetscArraymove()Commit-type: style-fix, feature
initialized pointers in MatLUFactorNumeric_SuperLU_DIST()
bug fix for uninitialized 'rstart'
rm SuperLU_MatInputMode
rm all codes of MatInputMode == GLOBAL
deprecate MatInputMode == GLOBAL in superlu_dist interface as requested by S. Xiaoye Li
Changed VecScatterCreate() to take a communicator and return scatter as other first class PETSc objects do.Changed old VecScatterCreate() to VecScatterCreateWithData()Completed basic refactorizati
Changed VecScatterCreate() to take a communicator and return scatter as other first class PETSc objects do.Changed old VecScatterCreate() to VecScatterCreateWithData()Completed basic refactorization of VecScatter code, ready for new implementations to be added.(There is still some code duplication between the VecScatter MPI1 and MPI3 code but it can be fixed at anytime without requiring changes to the VecScatter API.Part 5 of a refactorization of VecScatter to be treated as a first class PETSc objectCommit-type: style-fix
superlu_dist: update to version v5.4.0$ git diff v5.3.0..v5.4.0 SRC/superlu_enum_consts.hdiff --git a/SRC/superlu_enum_consts.h b/SRC/superlu_enum_consts.hindex 628eecf..515643a 100644--- a/SRC
superlu_dist: update to version v5.4.0$ git diff v5.3.0..v5.4.0 SRC/superlu_enum_consts.hdiff --git a/SRC/superlu_enum_consts.h b/SRC/superlu_enum_consts.hindex 628eecf..515643a 100644--- a/SRC/superlu_enum_consts.h+++ b/SRC/superlu_enum_consts.h@@ -26,7 +26,7 @@ at the top-level directory. ***********************************************************************/ typedef enum {NO, YES} yes_no_t; typedef enum {DOFACT, SamePattern, SamePattern_SameRowPerm, FACTORED} fact_t;-typedef enum {NOROWPERM, LargeDiag, MY_PERMR} rowperm_t;+typedef enum {NOROWPERM, LargeDiag_MC64, LargeDiag_AWPM, MY_PERMR} rowperm_t; typedef enum {NATURAL, MMD_ATA, MMD_AT_PLUS_A, COLAMD, METIS_AT_PLUS_A, PARMETIS, ZOLTAN, MY_PERMC} colperm_t; typedef enum {NOTRANS, TRANS, CONJ} trans_t;
add MatGetInertia_SuperLU_DIST for complex Hermitian matrix
Fix MatGetInertia_SuperLU_DIST()Reported-by: Jose E. Roman <jroman@dsic.upv.es>
initialize pointers to 'NULL'
add Cholesky factorization when A->symmetric=true
add MatGetInertia_SuperLU_DIST()Requested-by: Xiaoye S. Li <xsli@lbl.gov>
12345678910>>...13