diff options
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/llvm.fmuladd.f16.ll')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/llvm.fmuladd.f16.ll | 13 |
1 files changed, 7 insertions, 6 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.fmuladd.f16.ll b/llvm/test/CodeGen/AMDGPU/llvm.fmuladd.f16.ll index add22dbfab5..ebf8bafd3c8 100644 --- a/llvm/test/CodeGen/AMDGPU/llvm.fmuladd.f16.ll +++ b/llvm/test/CodeGen/AMDGPU/llvm.fmuladd.f16.ll @@ -95,8 +95,9 @@ define amdgpu_kernel void @fmuladd_f16_imm_b( } ; GCN-LABEL: {{^}}fmuladd_v2f16 +; VI: buffer_load_dword v[[B_V2_F16:[0-9]+]] ; GCN: buffer_load_dword v[[A_V2_F16:[0-9]+]] -; GCN: buffer_load_dword v[[B_V2_F16:[0-9]+]] +; SI: buffer_load_dword v[[B_V2_F16:[0-9]+]] ; GCN: buffer_load_dword v[[C_V2_F16:[0-9]+]] ; SI: v_cvt_f32_f16_e32 v[[A_F32_0:[0-9]+]], v[[A_V2_F16]] @@ -124,11 +125,11 @@ define amdgpu_kernel void @fmuladd_f16_imm_b( ; VI-FLUSH-NOT: v_and_b32 ; VI-FLUSH: v_or_b32_e32 v[[R_V2_F16:[0-9]+]], v[[A_V2_F16]], v[[R_F16_HI]] -; VI-DENORM: v_lshrrev_b32_e32 v[[A_F16_1:[0-9]+]], 16, v[[A_V2_F16]] -; VI-DENORM: v_lshrrev_b32_e32 v[[B_F16_1:[0-9]+]], 16, v[[B_V2_F16]] -; VI-DENORM: v_lshrrev_b32_e32 v[[C_F16_1:[0-9]+]], 16, v[[C_V2_F16]] -; VI-DENORM-DAG: v_fma_f16 v[[RES0:[0-9]+]], v[[A_V2_F16]], v[[B_V2_F16]], v[[C_V2_F16]] -; VI-DENORM-DAG: v_fma_f16 v[[RES1:[0-9]+]], v[[A_F16_1]], v[[B_F16_1]], v[[C_F16_1]] +; VI-DENORM-DAG: v_lshrrev_b32_e32 v[[A_F16_1:[0-9]+]], 16, v[[A_V2_F16]] +; VI-DENORM-DAG: v_lshrrev_b32_e32 v[[B_F16_1:[0-9]+]], 16, v[[B_V2_F16]] +; VI-DENORM-DAG: v_lshrrev_b32_e32 v[[C_F16_1:[0-9]+]], 16, v[[C_V2_F16]] +; VI-DENORM-DAG: v_fma_f16 v[[RES0:[0-9]+]], v[[B_V2_F16]], v[[A_V2_F16]], v[[C_V2_F16]] +; VI-DENORM-DAG: v_fma_f16 v[[RES1:[0-9]+]], v[[B_F16_1]], v[[A_F16_1]], v[[C_F16_1]] ; VI-DENORM-DAG: v_lshlrev_b32_e32 v[[R_F16_HI:[0-9]+]], 16, v[[RES1]] ; VI-DENORM-NOT: v_and_b32 ; VI-DENORM: v_or_b32_e32 v[[R_V2_F16:[0-9]+]], v[[RES0]], v[[R_F16_HI]] |
