Slightly improved the handling of PCFailedReason on multiple rankspc->failedreasonrank now contains the failed reason from the current rankpc->failedreason continues the unified common value colle
Slightly improved the handling of PCFailedReason on multiple rankspc->failedreasonrank now contains the failed reason from the current rankpc->failedreason continues the unified common value collected in KSPCheckNorm/DotThis parallel code can check the common value and be sure to do the same action on each processAlso printed message now provide the accurate value instead of the value on rank 0Changed KSPReasonView() and KSPReasonViewFromOptions() to KSPCOnvergeReasonView() and KSPConvergedReasonViewFromOptions()Added the PetscViewerFormat argument to the public KSPReasonViewFromOptions() and removed the private functionRemoved some off the PCFailedReaon processing code from KSPSolve_PREONLY since buggily assumed the same reason all ranksThe error message on KSP failure now also prints the PCFailedReason if appropriateNeeded to add a new alt fileReported-by: Yu, Xiangyu <Xiangyu.Yu@bp.com>
show more ...
SNES ex62: Completely reworked
Add Kokkos test example, testing infrastructure and to CIMove updates to gmake system for double suffix from adams/feature-dmplex-snes-landau-kokkos-barryCommit-type: featureFunded-by: ECP/spen
Add Kokkos test example, testing infrastructure and to CIMove updates to gmake system for double suffix from adams/feature-dmplex-snes-landau-kokkos-barryCommit-type: featureFunded-by: ECP/spend 15h30mThanks-to: Junchao Zhang <junchao.zhang@gmail.com>
Factored matrices should not view information about MatSetValues() mallocsCommit-type: style-fix, feature/spend 40m
Add multiple MPI rank AIJMKL test case to SNES tutorial ex56.
Add seqaijmkl case to SNES ex56 to test PtAP for symmetric matrices.
Most external factorization packages do not use the ordering provided by PETSc, therefor only provide it when needed.Adds MatOrderingType external to indicate not to generate an ordering and use wh
Most external factorization packages do not use the ordering provided by PETSc, therefor only provide it when needed.Adds MatOrderingType external to indicate not to generate an ordering and use what the package needsNow -pc_view will not print the wrong PETSc ordering when the ordering is done externallySaves some memory and compute time.Commit-type: optimizationTime: 1.2 hoursReported-by: Junchao Zhang <junchao.zhang@gmail.com>Thanks-to: Stefano Zampini <stefano.zampini@gmail.com>
SNES ex56: add viennacl and cuda tests
make cg default for cheby with MAT_SPD
Merge branch 'jczhang/update-ex19-output' into 'master'Update snes/tutorials/ex19_fieldsplit_5.outSee merge request petsc/petsc!2832
SNES tutorials ex35: update output
cleaning up ex56, eigest matsetblocksize
Update ex19_fieldsplit_5.out since it is the output
Update to hypre 2.19.0 and to MUMPS 5.3.1
Update HPDDM outputs
Merge branch 'stefanozampini/feature-add-hpackages-rebased' into 'master'Fixes to MatProduct + many bug fixes + new hierarchical matrix supportSee merge request petsc/petsc!2800
SNES ex23: Test of a split domain
SNES ex12: update HPDDM output
Merge branch 'psanan/asm-view-fix' into 'master'PC: ASM, Block Jacobi: when viewing, note local solver on rank 0 is shownSee merge request petsc/petsc!2563
ci: update alt files for osx- src/snes/tutorials/output/ex12_44.out likely incorrectly updated in c4762a1b19cd2af06abeed90e8f9d34fb975dd94- src/ts/tutorials/output/ex12_2_alt.out added for one of
ci: update alt files for osx- src/snes/tutorials/output/ex12_44.out likely incorrectly updated in c4762a1b19cd2af06abeed90e8f9d34fb975dd94- src/ts/tutorials/output/ex12_2_alt.out added for one of the 3 OSX boxes (jpro) - producing slightly different numbers
PC: ASM, Block Jacobi: when viewing, note local solver on rank 0 is shownThe KSPView() call shows the (sub-)KSP on rank 0, which should have the sametype, PC, parameters, etc., but will not be the
PC: ASM, Block Jacobi: when viewing, note local solver on rank 0 is shownThe KSPView() call shows the (sub-)KSP on rank 0, which should have the sametype, PC, parameters, etc., but will not be the identical solver,as the size of the local subdomains and the entries of the localsubmatrices will be different. For instance, the output only showsthe local Mat for rank 0, which might confuse a user into thinking thatthe subdomains are all of the same size when they are not. Similarly,a PC like LU would show different factorization details on differentranks, even though the solver settings are the same.Make the viewer output more precise by noting explicitly that isdisplaying information for the rank 0 subsolver.Update test output with the new string.
DM+Plex: Example Fixes for cell type rewrite
SNES: changing misleading example to testReported-by: Xiaodong Liu<xliu29@ncsu.edu>Thanks-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/
12345678910