KSPCHEBYSHEV: need to call setfromoptions
SNES: replace missing direct calls to ops->converged with SNESConverged
SNESMS: fix missing norm computationsdeprecate -snes_ms_norms
Merge branch 'barry/2023-05-01/add-superlu-dist-mixed-precision' into 'main'Add support for using SuperLU_DIST in single precision from PETSc in double precisionSee merge request petsc/petsc!6402
Add support for using SuperLU_DIST in single precision from PETSc in double precision-pc_precision single (don't use an options prefix here even if the PC has such a prefix)Commit-type: featureR
Add support for using SuperLU_DIST in single precision from PETSc in double precision-pc_precision single (don't use an options prefix here even if the PC has such a prefix)Commit-type: featureReported-by: Jin Chen <jchen@pppl.gov>Thanks-to: "Xiaoye S. Li" <xsli@lbl.gov>
show more ...
SNES: move monitor after convergence testUse a developer level convergence function and fix a few corner cases
PCHYPRE: always print spgemm algorithmuse hypre for non-device buildsclean up alt files
SNES tutorials ex55: support non-kokkos runs
Merge remote-tracking branch 'origin/release'
MatFDColoringApply_AIJ: fix for device matrices
Remove unneeded output files
Fix long term bug in example where it started a SNESSolve off at a Jacobian of exactly 0Commit-type: bug
Fix bug in handling of -snes_linesearch_maxstep.It should be relative to fnorm, not an absolute valueCommit-type: bug-fix
SNESNEWTONTR: add support for NPC and sufficient decreaseadd tests
Add implementation of 4th-kind Chebyshev polynomial smoother
SNESNEWTONTR: add dogleg and Cauchy fallbacks from NTRDC
SNES ex7: Fermions on a hypercubic latticeCo-Authored-By: Joseph Pusztay <jjoep93@gmail.com>
KSPHPDDM: add support for CUDA
Register matcovert_seqaijhipsparse in HIP implementation
SNES: Updated tests
Update SNES/tutorials/ex56 with modern use of Plex, a Laplacian test in addition to the elasticity test, and some cleanup of ad hoc code.
Add yet another alt file for snes/tutorials/ex19.c test
Add HIP support to Mat class in PETScCo-authored-by: Justin Chang <jychang48@gmail.com>Fixed-by: Jacob Faibussowitsch <jacob.fai@gmail.com>Fixed-by: Junchao Zhang <jczhang@mcs.anl.gov>Fixed-by:
Add HIP support to Mat class in PETScCo-authored-by: Justin Chang <jychang48@gmail.com>Fixed-by: Jacob Faibussowitsch <jacob.fai@gmail.com>Fixed-by: Junchao Zhang <jczhang@mcs.anl.gov>Fixed-by: Satish Balay <balay@mcs.anl.gov>
Minor fixes for working with GAMG, minor optimizationsRemove the -pc_gamg_symmetrize_graph option completely. Use symmetric state of matrixCombine MatCreateGraph() and MatFilter().Minor optimiz
Minor fixes for working with GAMG, minor optimizationsRemove the -pc_gamg_symmetrize_graph option completely. Use symmetric state of matrixCombine MatCreateGraph() and MatFilter().Minor optimizations of MatCreateGraph() when numerical values are not needed skip abs(entries) and dscale (for MIS and MISK)Add PCGAMGCreateGraph() functionTODO: when using MIS and MISK only structural symmetry of matrix is needed, not actual symmetryCommit-type: feature, housekeeping/spend 2h
Fix examples (with completion)
12345678910