Topic/migrate2v2#155
Open
bosilca wants to merge 2 commits into
Open
Conversation
Signed-off-by: George Bosilca <[email protected]>
Move GPU-enabled PTG kernels toward the common per-stream handle-key style. Add CUDA/HIP handle globals to the JDFs and initialize them in the corresponding wrappers. Add HIP bodies for the GEMM, TRSM, TRMM, POTRF, POINV, GEQRF, and GETRF-no-pivot GPU paths that currently have CUDA coverage. Modernize the GEQRF TSMQR CUDA helper to use CUBLAS v2 handles and add the matching HIPBLAS helper twin. Wire the generated HIP helper core into the cores build. For zgetrf_nopiv, add best-effort HIP support for the existing GPU trailing-update GEMM while leaving panel factorization and TRSM tasks on the CPU fallback path. Signed-off-by: George Bosilca <[email protected]>
Contributor
Author
|
At this point it is normal for this to fail to build, it needs #156 to be merged first. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR extends DPLASMA's CUDA/HIP support and moves the GPU-enabled PTG
kernels to the common handle-key style. CUDA and HIP bodies now retrieve
their BLAS handles from per-stream runtime info keys, and the wrappers
initialize those keys consistently.
The main functional additions are:
ztsmqrhelper.ztsmqrhelper.zgetrf_nopivfor the existing GPU trailing-update GEMM.GPU Coverage
zgemm_{NN,NT,TN,TT}zgemm_*_summazgemm_NN_gpuztrsm_*ztrmm_*zpotrf_{L,U}zpoinv_{L,U}zgeqrfzgetrf_nopivValidation
Tested in the debug build: