Add MatDenseCUPMSetPreallocation

Merge remote-tracking branch 'origin/release'

Merge branch 'jolivet/fix-spacing' into 'release'

Docs: fix extra spacing

See merge request petsc/petsc!6443

Docs: fix extra spacing

Merge remote-tracking branch 'origin/release'

Merge branch 'jose/release/fix-man-pages' into 'release'

Fix several manual pages

See merge request petsc/petsc!6423

Fix manpage of MatSetValue()

Merge branch 'adams/mat-remove-matsetvaluesdevice' into 'main'

remove MatSetValuesDevice

See merge request petsc/petsc!6378

remove MatSetValuesDevice

Merge remote-tracking branch 'origin/release'

Merge branch 'prj--release-patch-92401' into 'release'

Update petscmat.h

See merge request petsc/petsc!6362

Update petscmat.h

Merge remote-tracking branch 'origin/release'

Merge branch 'barry/2023-03-30/fix-include-man-pages/release' into 'release'

Fix manual pages in include files

See merge request petsc/petsc!6272

Fix manual pages in include files

Merge branch 'hongzh/sell-cuda' into 'main'

SELL-based SpMV

See merge request petsc/petsc!3428

Add MatSeqSELLGetAvgSliceSize

It returns the variance of the slice sizes.

Further optimization of MatMult_SeqSELLCUDA

- Add more kernels
- Use multiple threads per row for matrices with narrow slices
- Use multiple blocks per slice for matrices with wide slices
- Add thre

Further optimization of MatMult_SeqSELLCUDA

- Add more kernels
- Use multiple threads per row for matrices with narrow slices
- Use multiple blocks per slice for matrices with wide slices
- Add three new APIs to return the irregularity ratio, the maximum slice width and the average slice width

Experiments show that column blocking gives much worse performance for wide matrices and permulation based on slice width has almost no impact on the performance.

show more ...

Initial commit for porting SELL to GPU

- Add tiled SPMV and basic SpMVfor SeqSELL
- Tested in serial
- Offloadmask is used to determine when the matrix should be copied to GPU
- Use different slice

Initial commit for porting SELL to GPU

- Add tiled SPMV and basic SpMVfor SeqSELL
- Tested in serial
- Offloadmask is used to determine when the matrix should be copied to GPU
- Use different slice height for CUDA version
- By checking the nonzerostate, PETSc can decide if the whole matrix need to be copied or just the values need to be copied
- Make the convert function public so that the very slow MatConvert_Basic can be avoided sometimes. E.g. one can use a two-step convert method: AIJ->SELL,SELL->SELLCUDA instead of the direct convert AIJ->SELLCUDA
- Make the FLOPS count for SELL same as that for AIJCUSPARSE.
- MatDisAssemble is not needed.
- Change slice height from 32 to 16 for GPU
- To overlap communication with MatMult, VecScatterBegin() should be called before MatMult() for the diagonal part.
- SLICE_HEIGHT is defined to be 32 to match the warp size of GPU. For other cases, it is still 8.

Funded-by:
Project: PETSc for GPU
Time: 42 hours
Reported-by:
Thanks-to:

show more ...

Merge branch 'jolivet/remove-sparseelemental' into 'main'

Remove MATSOLVERSPARSELEMENTAL

See merge request petsc/petsc!6296

Remove MATSOLVERSPARSELEMENTAL

-pc_factor_mat_solver_type sparseelemental causes segmentation faults, has been unsupported for many years, and there is no interest in anyone supporting or developing

Remove MATSOLVERSPARSELEMENTAL

-pc_factor_mat_solver_type sparseelemental causes segmentation faults, has been unsupported for many years, and there is no interest in anyone supporting or developing the code

show more ...

Merge remote-tracking branch 'origin/release'

Merge branch 'jolivet/fix-Wzero-as-null-pointer-constant-ci' into 'release'

Fix -Wzero-as-null-pointer-constant

See merge request petsc/petsc!6294

Fix -Wzero-as-null-pointer-constant + El warnings

Merge branch 'hsuh/mat-create-dense-from-vectype' into 'main'

Add VecCreateMatDense(): Create a matrix that matches the type of a Vec

See merge request petsc/petsc!6237