Rename petscvec.hpp -> petscveckokkos.hpp

Reducing first coarse grid to one process/device in Cuda and manual override of this and other heuristics for process reduction in GAMG coarse grids.

Merge remote-tracking branch 'origin/release' into master

Report all GPU errors as PETSC_ERR_GPU so that the test harness can retry any GPU failure

We still get occassional errors:

ot ok ksp_ksp_tutorials-ex71_bddc_cusparse # Error code: 180

That occur a

Report all GPU errors as PETSC_ERR_GPU so that the test harness can retry any GPU failure

We still get occassional errors:

ot ok ksp_ksp_tutorials-ex71_bddc_cusparse # Error code: 180

That occur after the cuSPARSE library must have been loaded since the load command already succeeded

Commit-type: error-checking, CI, enhancement
/spend 15m

show more ...

MatDuplicate_SeqAIJ_Inode: fix for null pointer

Reported by: Jose E. Roman <jroman@dsic.upv.es>

MATMPIAIJCUSPARSE: fix conversions

MatSeqAIJINode: do not change pointers for Mult,MultAdd and SOR

MatInode: set it to false when using device matrices

MATAIJ: few more fixes to KOKKOS and fixes to function composition

MPIAIJ matmults: add new nonzero location error

MatTransposeMatMultSymbolic_MPIAIJ_MPIAIJ_nonscalable: fill nonzero pattern

All the other symbolic factorizations fill the nonzero pattern. This does not

MATAIJKOKKOS: bugfixes

MatTransposeMatMult_SeqAIJ: avoid transpositions if possible

Fix building PETSC 3.14 with cuda 10.0

fix doc format

add cuda matrix for meta data method, assembly

Merge branch 'barry/2020-09-25/do-not-log-gpu-transfer-time' into 'master'

Do not log transfer time as GPU time. GPU time is used to get performance ON...

See merge request petsc/petsc!3247

Add MATAIJKOKKOS

Always check for out of range indices in MatSetValues_MPIXXAIJ()

People run much larger more interesting cases with debugging turned off so the debug version may not catch actual errors

Commit-type

Always check for out of range indices in MatSetValues_MPIXXAIJ()

People run much larger more interesting cases with debugging turned off so the debug version may not catch actual errors

Commit-type: error-checking
/spend 5m

show more ...

Do not log transfer time as GPU time. GPU time is used to get performance ON the GPU (for example flop rates)
logging transfer times distorts these numbers since they would include the time to copy d

Do not log transfer time as GPU time. GPU time is used to get performance ON the GPU (for example flop rates)
logging transfer times distorts these numbers since they would include the time to copy down the matrix.
The transfer times are already logged in the copy to and from events and the performance flop rates including the
copy times are already logged in the overall flop rate.

Commit-type: bug-fix, style-fix
/spend 5m

show more ...

fix MatGetRowMaxAbs() for MPIBAIJ

fix MatGetRowMinAbs() for AIJ matrix

fix MatGetRowMaxAbs() for AIJ matrix

fix MatGetRowMin_ for SeqAIJ and MPIAIJ

bugfix for MatGetRowMax_xxxAIJ
Reported-by: Sajid Ali <sajidsyed2021@u.northwestern.edu>