ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-12 09:16:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

wip fix

Browse Source
This commit is contained in:
carlushuang
2024-03-06 14:31:36 +00:00
parent f549bb5d39
commit 0e7df1999f
10 changed files with 268 additions and 68 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
include/ck_tile/ops/epilogue/default_2d_epilogue.hpp
View File

@@ -38,7 +38,11 @@ struct Default2DEpilogue
// TODO: this is ugly
if constexpr(kPadM || kPadN)
{
store_tile_raw(o_dram_window_tmp, cast_tile<ODataType>(o_acc_tile));
// o_dram_window_tmp.foo();
// ODataType{}.foo();
// o_acc_tile.foo();
auto x = cast_tile<ODataType>(o_acc_tile);
store_tile_raw(o_dram_window_tmp, x);
buffer_store_fence();
}
else

4
include/ck_tile/ops/reduce/block/block_reduce.hpp
View File

@@ -17,8 +17,8 @@ CK_TILE_DEVICE void block_tile_reduce_sync(AccDistributedTensor_& acc_tensor,
using DstrEncode = typename Dstr::DstrEncode;
using DstrEncodeDetail = typename DstrEncode::detail;
constexpr index_t NDimP = Dstr::GetNumOfDimensionP();
constexpr index_t NDimR = Dstr::GetNumOfDimensionR();
constexpr index_t NDimP = Dstr::get_num_of_dimension_p();
constexpr index_t NDimR = Dstr::get_num_of_dimension_r();
constexpr index_t idim_p_lane = NDimP - 1;