diff --git a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_bwd_weight.hpp b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_bwd_weight.hpp index 2b66898b1e..2da92466b5 100644 --- a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_bwd_weight.hpp +++ b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_bwd_weight.hpp @@ -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, ck::bhalf_t, FloatAB>; #else using FloatABAdjusted = FloatAB; diff --git a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_v2r3.hpp b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_v2r3.hpp index 02b0081349..51c578385f 100644 --- a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_v2r3.hpp +++ b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_v2r3.hpp @@ -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, ck::bhalf_t, FloatAB>; #else using FloatABAdjusted = FloatAB;