ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-03 21:21:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add more macros to turn on/off denorm fix (#678)

Browse Source 
Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com>
This commit is contained in:
Rostyslav Geyyer
2023-04-15 21:56:07 -05:00
committed by GitHub
parent e85178b4ca
commit 03eaee6ae6
2 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_bwd_weight.hpp
View File

@@ -265,7 +265,7 @@ struct GridwiseGemm_bk0mk1_bk0nk1_mn_xdlops_bwd_weight
// we convert fp16->fp32->bf16 and execute bf16 mfma instruction
// when mfma if fixed, remove this section and update
// FloatABAdjusted -> FloatAB throughout this file
#if defined(__gfx90a__)
#if CK_WORKAROUND_DENORM_FIX && defined(__gfx90a__)
using FloatABAdjusted = conditional_t<is_same_v<FloatAB, ck::half_t>, ck::bhalf_t, FloatAB>;
#else
using FloatABAdjusted = FloatAB;

2
include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_v2r3.hpp
View File

@@ -135,7 +135,7 @@ struct GridwiseGemm_k0mk1_k0nk1_mn_xdlops_v2r3
// we convert fp16->fp32->bf16 and execute bf16 mfma instruction
// when mfma if fixed, remove this section and update
// FloatABAdjusted -> FloatAB throughout this file
#if defined(__gfx90a__)
#if CK_WORKAROUND_DENORM_FIX && defined(__gfx90a__)
using FloatABAdjusted = conditional_t<is_same_v<FloatAB, ck::half_t>, ck::bhalf_t, FloatAB>;
#else
using FloatABAdjusted = FloatAB;