ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-13 17:55:48 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'origin/build_0705' into implicit_gemm_v4r2

Browse Source
This commit is contained in:
Chao Liu
2019-07-05 16:29:20 -05:00
parent 8133713e96 3276a5e9b9
commit 96d73c2154
17 changed files with 1008 additions and 433 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
composable_kernel/include/tensor_operation/blockwise_gemm.hpp
View File

@@ -205,10 +205,10 @@ struct BlockwiseGemmBlockABlockBThreadCTransANormalBNormalC_v2
// thread A, B for GEMM
constexpr auto a_thread_mtx =
make_ConstantMatrixDescriptor(Number<KPerThreadLoop>{}, Number<MPerThread>{});
make_ConstantMatrixDescriptor_packed(Number<KPerThreadLoop>{}, Number<MPerThread>{});
constexpr auto b_thread_mtx =
make_ConstantMatrixDescriptor(Number<KPerThreadLoop>{}, Number<NPerThread>{});
make_ConstantMatrixDescriptor_packed(Number<KPerThreadLoop>{}, Number<NPerThread>{});
// thread A-sub, B-sub for copy
constexpr auto a_thread_sub_mtx = make_ConstantMatrixDescriptor(