ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-02 04:31:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

use inline asm for 4x4 int8 transposition (#187)

Browse Source
This commit is contained in:
Anthony Chang
2022-04-23 04:47:31 +08:00
committed by GitHub
parent 1a0cd5d160
commit 08a979f188
2 changed files with 88 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
include/ck/tensor_operation/gpu/thread/threadwise_tensor_slice_transfer_v3r1.hpp
View File

@@ -277,9 +277,12 @@ struct ThreadwiseTensorSliceTransfer_v3r1
// sub-dword transpose between src_thread_scratch_ and dst_thread_scratch_
// TODO make this logic more generic for more sub-dword datatype
if constexpr(SrcVectorDim != DstVectorDim &&
is_same<half_t, remove_cvref_t<SrcData>>::value &&
is_same<half_t, remove_cvref_t<DstData>>::value &&
SrcScalarPerVector % 2 == 0 && DstScalarPerVector % 2 == 0)
((is_same<half_t, remove_cvref_t<SrcData>>::value &&
is_same<half_t, remove_cvref_t<DstData>>::value &&
SrcScalarPerVector % 2 == 0 && DstScalarPerVector % 2 == 0) ||
(is_same<int8_t, remove_cvref_t<SrcData>>::value &&
is_same<int8_t, remove_cvref_t<DstData>>::value &&
SrcScalarPerVector % 4 == 0 && DstScalarPerVector % 4 == 0)))
{
// each transpose does
// DstScalarPerVector # of src vectors in src_thread_scratch_