ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-24 23:05:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

added (1x4)x(2x4) threadwise gemm

Browse Source
This commit is contained in:
Chao Liu
2019-07-30 18:20:55 -05:00
parent c5e5a9307b
commit 08cbac98cc
4 changed files with 166 additions and 232 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
composable_kernel/include/utility/amd_inline_asm.hpp
View File

@@ -105,6 +105,15 @@ __device__ void outerProduct1x4(const float& a,
outerProduct1x4(&a, reinterpret_cast<const float*>(&b), reinterpret_cast<float*>(&c));
}
__device__ void outerProduct2x4(const vector_type<float, 2>::MemoryType& a,
const vector_type<float, 4>::MemoryType& b,
vector_type<float, 4>::MemoryType& c0,
vector_type<float, 4>::MemoryType& c1)
{
outerProduct1x4(a.x, b, c0);
outerProduct1x4(a.y, b, c1);
}
__device__ void outerProduct4x4(const vector_type<float, 4>::MemoryType& a,
const vector_type<float, 4>::MemoryType& b,
vector_type<float, 4>::MemoryType& c0,