ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-26 16:04:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix macro for exp2; fix warpgemm a/b in transposedC

Browse Source
This commit is contained in:
carlushuang
2024-03-06 15:59:21 +00:00
parent 0e7df1999f
commit 7df3947819
7 changed files with 34 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
include/ck_tile/ops/gemm/warp/warp_gemm_attribute_mfma.hpp
View File

@@ -309,8 +309,8 @@ struct WarpGemmAtrributeMfmaIterateKAndTransposedCDistribution
// swap A and B, value and type
static_for<0, kKIter, 1>{}([&](auto iKIter) {
Impl{}(c_vec,
a_vec.template get_as<typename Impl::AVecType>()[iKIter],
b_vec.template get_as<typename Impl::BVecType>()[iKIter]);
b_vec.template get_as<typename Impl::AVecType>()[iKIter],
a_vec.template get_as<typename Impl::BVecType>()[iKIter]);
});
}
@@ -320,13 +320,13 @@ struct WarpGemmAtrributeMfmaIterateKAndTransposedCDistribution
constexpr auto I0 = number<0>{};
// swap A and B, value and type
auto c_vec = Impl{}(a_vec.template get_as<typename Impl::AVecType>()[I0],
b_vec.template get_as<typename Impl::BVecType>()[I0]);
auto c_vec = Impl{}(b_vec.template get_as<typename Impl::AVecType>()[I0],
a_vec.template get_as<typename Impl::BVecType>()[I0]);
static_for<1, kKIter, 1>{}([&](auto iKIter) {
Impl{}(c_vec,
a_vec.template get_as<typename Impl::AVecType>()[iKIter],
b_vec.template get_as<typename Impl::BVecType>()[iKIter]);
b_vec.template get_as<typename Impl::AVecType>()[iKIter],
a_vec.template get_as<typename Impl::BVecType>()[iKIter]);
});
return c_vec;