diff options
| author | Navdeep Kumar <navdeep.navdeep37@gmail.com> | 2021-05-21 21:19:37 +0530 |
|---|---|---|
| committer | Uday Bondhugula <uday@polymagelabs.com> | 2021-05-21 21:20:33 +0530 |
| commit | eaaf7a6a09da905cc314201f93e2be11773726a0 (patch) | |
| tree | 8d48473d522a928269613a75fcfa52e05dcb3cc2 /llvm/lib/CodeGen/MachineDebugify.cpp | |
| parent | c2d44bd2309c1e232d900fd6979aba320c913357 (diff) | |
| download | llvm-eaaf7a6a09da905cc314201f93e2be11773726a0.zip llvm-eaaf7a6a09da905cc314201f93e2be11773726a0.tar.gz llvm-eaaf7a6a09da905cc314201f93e2be11773726a0.tar.bz2 | |
[MLIR][GPU][NVVM] Add conversion of warp synchronous matrix-multiply accumulate GPU ops
Add conversion of warp synchronous matrix-multiply
accumulate GPU ops
Add conversion of warp synchronous matrix-multiply accumulate GPU ops to
NVVM ops. The following conversions are added :-
1.) subgroup_mma_load_matrix -> wmma.m16n16k16.load.[a,b,c]..row.stride
2.) subgroup_mma_store_matrix -> wmma.m16n16k16.store.d.[f16,f32].row.stride
3.) subgroup_mma_compute -> wmma.m16n16k16.mma.row.row.[f16,f32].[f16,f32]
Reviewed By: bondhugula, ftynse
Differential Revision: https://reviews.llvm.org/D95331
Diffstat (limited to 'llvm/lib/CodeGen/MachineDebugify.cpp')
0 files changed, 0 insertions, 0 deletions