ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 06:49:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CK-Tile][Async gemm] add missing sync and f8 inputs test cases (#3000)

Browse Source 
* add missing sync and f8 test cases

* reformat test cases

* comment failing cases

* bump

* reintroduce compv4 shapes
This commit is contained in:
Max Podkorytov
2025-10-24 12:16:01 -07:00
committed by GitHub
parent 0584399571
commit 86d542f663
2 changed files with 55 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
include/ck_tile/ops/gemm/pipeline/gemm_pipeline_ag_bg_cr_comp_async.hpp
View File

@@ -472,6 +472,8 @@ struct GemmPipelineAgBgCrCompAsync : public BaseGemmPipelineAgBgCrCompAsync<Prob
block_gemm(c_block_tile, a_block_tile0, b_block_tile0);
}
{
// write to LDS window(0) must complete before the local prefetch
block_sync_lds_direct_load();
// read A(num_loop), B(num_loop) from LDS window(0) to pipeline registers(0)
Base::LocalPrefetch(a_block_tile0, a_lds_ld_window0, is_a_load_tr_v);
Base::LocalPrefetch(b_block_tile0, b_lds_ld_window0, is_b_load_tr_v);