diff options
| author | Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com> | 2017-09-06 18:29:51 +0000 |
|---|---|---|
| committer | Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com> | 2017-09-06 18:29:51 +0000 |
| commit | ea134bcb133dc8147a3f44ec200348e2d699b955 (patch) | |
| tree | 9f20d01c5a5a1de13c9c90a09ed5e8416009ef88 /llvm/test | |
| parent | 0440be4a42e53e179b013f95bb805a7fe4296521 (diff) | |
| download | bcm5719-llvm-ea134bcb133dc8147a3f44ec200348e2d699b955.tar.gz bcm5719-llvm-ea134bcb133dc8147a3f44ec200348e2d699b955.zip | |
[AMDGPU] Fixed encoding of v_pk_mul_f16 in fcanonicalize
Differential Revision: https://reviews.llvm.org/D37522
llvm-svn: 312660
Diffstat (limited to 'llvm/test')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/fcanonicalize.f16.ll | 11 |
1 files changed, 5 insertions, 6 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/fcanonicalize.f16.ll b/llvm/test/CodeGen/AMDGPU/fcanonicalize.f16.ll index 52b85565176..1dfd2fe1522 100644 --- a/llvm/test/CodeGen/AMDGPU/fcanonicalize.f16.ll +++ b/llvm/test/CodeGen/AMDGPU/fcanonicalize.f16.ll @@ -211,7 +211,7 @@ define amdgpu_kernel void @test_fold_canonicalize_snan3_value_f16(half addrspace ; VI-DAG: v_max_f16_e32 [[REG1:v[0-9]+]], {{v[0-9]+}}, {{v[0-9]+}} ; VI-NOT: v_and_b32 -; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, {{v[0-9]+$}} +; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, {{v[0-9]+}} op_sel_hi:[0,1]{{$}} ; GFX9: buffer_store_dword [[REG]] define amdgpu_kernel void @v_test_canonicalize_var_v2f16(<2 x half> addrspace(1)* %out) #1 { %tid = call i32 @llvm.amdgcn.workitem.id.x() @@ -232,7 +232,7 @@ define amdgpu_kernel void @v_test_canonicalize_var_v2f16(<2 x half> addrspace(1) ; VI: v_or_b32 ; GFX9: v_and_b32_e32 [[ABS:v[0-9]+]], 0x7fff7fff, v{{[0-9]+}} -; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, [[ABS]]{{$}} +; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, [[ABS]] op_sel_hi:[0,1]{{$}} ; GCN: buffer_store_dword define amdgpu_kernel void @v_test_canonicalize_fabs_var_v2f16(<2 x half> addrspace(1)* %out) #1 { %tid = call i32 @llvm.amdgcn.workitem.id.x() @@ -251,7 +251,7 @@ define amdgpu_kernel void @v_test_canonicalize_fabs_var_v2f16(<2 x half> addrspa ; VI: v_or_b32 ; GFX9: v_and_b32_e32 [[ABS:v[0-9]+]], 0x7fff7fff, v{{[0-9]+}} -; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, [[ABS]] neg_lo:[0,1] neg_hi:[0,1]{{$}} +; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, [[ABS]] op_sel_hi:[0,1] neg_lo:[0,1] neg_hi:[0,1]{{$}} ; GCN: buffer_store_dword define amdgpu_kernel void @v_test_canonicalize_fneg_fabs_var_v2f16(<2 x half> addrspace(1)* %out) #1 { %tid = call i32 @llvm.amdgcn.workitem.id.x() @@ -264,7 +264,6 @@ define amdgpu_kernel void @v_test_canonicalize_fneg_fabs_var_v2f16(<2 x half> ad ret void } -; FIXME: Fold modifier ; GCN-LABEL: {{^}}v_test_canonicalize_fneg_var_v2f16: ; VI: v_xor_b32_e32 [[FNEG:v[0-9]+]], 0x80008000, v{{[0-9]+}} ; VI: v_lshrrev_b32_e32 [[FNEGHI:v[0-9]+]], 16, [[FNEG]] @@ -272,7 +271,7 @@ define amdgpu_kernel void @v_test_canonicalize_fneg_fabs_var_v2f16(<2 x half> ad ; VI-DAG: v_max_f16_e32 [[REG0:v[0-9]+]], [[FNEG]], [[FNEG]] ; VI-NOT: 0xffff -; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, {{v[0-9]+}} neg_lo:[0,1] neg_hi:[0,1]{{$}} +; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, {{v[0-9]+}} op_sel_hi:[0,1] neg_lo:[0,1] neg_hi:[0,1]{{$}} ; GFX9: buffer_store_dword [[REG]] define amdgpu_kernel void @v_test_canonicalize_fneg_var_v2f16(<2 x half> addrspace(1)* %out) #1 { %tid = call i32 @llvm.amdgcn.workitem.id.x() @@ -289,7 +288,7 @@ define amdgpu_kernel void @v_test_canonicalize_fneg_var_v2f16(<2 x half> addrspa ; VI: v_max_f16_e64 [[REG1:v[0-9]+]], {{s[0-9]+}}, {{s[0-9]+}} ; VI-NOT: v_and_b32 -; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, {{s[0-9]+$}} +; GFX9: v_pk_mul_f16 [[REG:v[0-9]+]], 1.0, {{s[0-9]+}} op_sel_hi:[0,1]{{$}} ; GFX9: buffer_store_dword [[REG]] define amdgpu_kernel void @s_test_canonicalize_var_v2f16(<2 x half> addrspace(1)* %out, i32 zeroext %val.arg) #1 { %val = bitcast i32 %val.arg to <2 x half> |
