diff options
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/sgpr-copy.ll')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/sgpr-copy.ll | 129 |
1 files changed, 52 insertions, 77 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/sgpr-copy.ll b/llvm/test/CodeGen/AMDGPU/sgpr-copy.ll index 2ad1788bb35..123b5de83f1 100644 --- a/llvm/test/CodeGen/AMDGPU/sgpr-copy.ll +++ b/llvm/test/CodeGen/AMDGPU/sgpr-copy.ll @@ -1,13 +1,10 @@ ; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck %s ; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck %s -; Function Attrs: nounwind readnone -declare <4 x float> @llvm.SI.image.sample.v2i32(<2 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #0 - ; CHECK-LABEL: {{^}}phi1: ; CHECK: s_buffer_load_dword [[DST:s[0-9]]], {{s\[[0-9]+:[0-9]+\]}}, 0x0 ; CHECK: v_mov_b32_e32 v{{[0-9]}}, [[DST]] -define amdgpu_ps void @phi1(<16 x i8> addrspace(2)* inreg %arg, <16 x i8> addrspace(2)* inreg %arg1, <8 x i32> addrspace(2)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19) #1 { +define amdgpu_ps void @phi1(<16 x i8> addrspace(2)* inreg %arg, <16 x i8> addrspace(2)* inreg %arg1, <8 x i32> addrspace(2)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19) #0 { main_body: %tmp = getelementptr <16 x i8>, <16 x i8> addrspace(2)* %arg, i32 0 %tmp20 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp, !tbaa !0 @@ -25,13 +22,13 @@ ELSE: ; preds = %main_body ENDIF: ; preds = %ELSE, %main_body %temp.0 = phi float [ %tmp26, %ELSE ], [ %tmp21, %main_body ] %tmp27 = fadd float %temp.0, %tmp23 - call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 0, float %tmp27, float %tmp27, float 0.000000e+00, float 1.000000e+00) + call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp27, float %tmp27, float 0.000000e+00, float 1.000000e+00, i1 true, i1 true) #0 ret void } ; Make sure this program doesn't crash ; CHECK-LABEL: {{^}}phi2: -define amdgpu_ps void @phi2(<16 x i8> addrspace(2)* inreg %arg, <16 x i8> addrspace(2)* inreg %arg1, <8 x i32> addrspace(2)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19) #0 { +define amdgpu_ps void @phi2(<16 x i8> addrspace(2)* inreg %arg, <16 x i8> addrspace(2)* inreg %arg1, <8 x i32> addrspace(2)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19) #1 { main_body: %tmp = getelementptr <16 x i8>, <16 x i8> addrspace(2)* %arg, i32 0 %tmp20 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp, !tbaa !0 @@ -58,32 +55,32 @@ main_body: %j.i = extractelement <2 x i32> %arg5, i32 1 %i.f.i = bitcast i32 %i.i to float %j.f.i = bitcast i32 %j.i to float - %p1.i = call float @llvm.amdgcn.interp.p1(float %i.f.i, i32 0, i32 0, i32 %arg3) #0 - %p2.i = call float @llvm.amdgcn.interp.p2(float %p1.i, float %j.f.i, i32 0, i32 0, i32 %arg3) #0 + %p1.i = call float @llvm.amdgcn.interp.p1(float %i.f.i, i32 0, i32 0, i32 %arg3) #1 + %p2.i = call float @llvm.amdgcn.interp.p2(float %p1.i, float %j.f.i, i32 0, i32 0, i32 %arg3) #1 %i.i19 = extractelement <2 x i32> %arg5, i32 0 %j.i20 = extractelement <2 x i32> %arg5, i32 1 %i.f.i21 = bitcast i32 %i.i19 to float %j.f.i22 = bitcast i32 %j.i20 to float - %p1.i23 = call float @llvm.amdgcn.interp.p1(float %i.f.i21, i32 1, i32 0, i32 %arg3) #0 - %p2.i24 = call float @llvm.amdgcn.interp.p2(float %p1.i23, float %j.f.i22, i32 1, i32 0, i32 %arg3) #0 + %p1.i23 = call float @llvm.amdgcn.interp.p1(float %i.f.i21, i32 1, i32 0, i32 %arg3) #1 + %p2.i24 = call float @llvm.amdgcn.interp.p2(float %p1.i23, float %j.f.i22, i32 1, i32 0, i32 %arg3) #1 %i.i13 = extractelement <2 x i32> %arg5, i32 0 %j.i14 = extractelement <2 x i32> %arg5, i32 1 %i.f.i15 = bitcast i32 %i.i13 to float %j.f.i16 = bitcast i32 %j.i14 to float - %p1.i17 = call float @llvm.amdgcn.interp.p1(float %i.f.i15, i32 0, i32 1, i32 %arg3) #0 - %p2.i18 = call float @llvm.amdgcn.interp.p2(float %p1.i17, float %j.f.i16, i32 0, i32 1, i32 %arg3) #0 + %p1.i17 = call float @llvm.amdgcn.interp.p1(float %i.f.i15, i32 0, i32 1, i32 %arg3) #1 + %p2.i18 = call float @llvm.amdgcn.interp.p2(float %p1.i17, float %j.f.i16, i32 0, i32 1, i32 %arg3) #1 %i.i7 = extractelement <2 x i32> %arg5, i32 0 %j.i8 = extractelement <2 x i32> %arg5, i32 1 %i.f.i9 = bitcast i32 %i.i7 to float %j.f.i10 = bitcast i32 %j.i8 to float - %p1.i11 = call float @llvm.amdgcn.interp.p1(float %i.f.i9, i32 1, i32 1, i32 %arg3) #0 - %p2.i12 = call float @llvm.amdgcn.interp.p2(float %p1.i11, float %j.f.i10, i32 1, i32 1, i32 %arg3) #0 + %p1.i11 = call float @llvm.amdgcn.interp.p1(float %i.f.i9, i32 1, i32 1, i32 %arg3) #1 + %p2.i12 = call float @llvm.amdgcn.interp.p2(float %p1.i11, float %j.f.i10, i32 1, i32 1, i32 %arg3) #1 %i.i1 = extractelement <2 x i32> %arg5, i32 0 %j.i2 = extractelement <2 x i32> %arg5, i32 1 %i.f.i3 = bitcast i32 %i.i1 to float %j.f.i4 = bitcast i32 %j.i2 to float - %p1.i5 = call float @llvm.amdgcn.interp.p1(float %i.f.i3, i32 2, i32 1, i32 %arg3) #0 - %p2.i6 = call float @llvm.amdgcn.interp.p2(float %p1.i5, float %j.f.i4, i32 2, i32 1, i32 %arg3) #0 + %p1.i5 = call float @llvm.amdgcn.interp.p1(float %i.f.i3, i32 2, i32 1, i32 %arg3) #1 + %p2.i6 = call float @llvm.amdgcn.interp.p2(float %p1.i5, float %j.f.i4, i32 2, i32 1, i32 %arg3) #1 %tmp45 = bitcast float %p2.i to i32 %tmp46 = bitcast float %p2.i24 to i32 %tmp47 = insertelement <2 x i32> undef, i32 %tmp45, i32 0 @@ -168,16 +165,16 @@ ENDIF24: ; preds = %IF25, %ENDIF %tmp111 = fsub float -0.000000e+00, %tmp105 %tmp112 = fmul float %tmp111, %tmp106 %tmp113 = call i32 @llvm.SI.packf16(float %tmp108, float %tmp110) - %tmp114 = bitcast i32 %tmp113 to float + %tmp114 = bitcast i32 %tmp113 to <2 x half> %tmp115 = call i32 @llvm.SI.packf16(float %tmp112, float 1.000000e+00) - %tmp116 = bitcast i32 %tmp115 to float - call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 1, float %tmp114, float %tmp116, float %tmp114, float %tmp116) + %tmp116 = bitcast i32 %tmp115 to <2 x half> + call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 15, <2 x half> %tmp114, <2 x half> %tmp116, i1 true, i1 true) #0 ret void } ; We just want ot make sure the program doesn't crash ; CHECK-LABEL: {{^}}loop: -define amdgpu_ps void @loop(<16 x i8> addrspace(2)* inreg %arg, <16 x i8> addrspace(2)* inreg %arg1, <8 x i32> addrspace(2)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19) #1 { +define amdgpu_ps void @loop(<16 x i8> addrspace(2)* inreg %arg, <16 x i8> addrspace(2)* inreg %arg1, <8 x i32> addrspace(2)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19) #0 { main_body: %tmp = getelementptr <16 x i8>, <16 x i8> addrspace(2)* %arg, i32 0 %tmp20 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp, !tbaa !0 @@ -204,7 +201,7 @@ LOOP: ; preds = %ENDIF, %main_body br i1 %tmp33, label %IF, label %ENDIF IF: ; preds = %LOOP - call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 0, float %temp4.0, float %temp5.0, float %temp6.0, float 1.000000e+00) + call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %temp4.0, float %temp5.0, float %temp6.0, float 1.000000e+00, i1 true, i1 true) #0 ret void ENDIF: ; preds = %LOOP @@ -230,7 +227,7 @@ ENDIF: ; preds = %LOOP ; CHECK: image_sample v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[SAMPLE_LO]]:[[SAMPLE_HI]]{{\]}} ; CHECK: exp ; CHECK: s_endpgm -define amdgpu_ps void @sample_v3([17 x <16 x i8>] addrspace(2)* byval %arg, [32 x <16 x i8>] addrspace(2)* byval %arg1, [16 x <8 x i32>] addrspace(2)* byval %arg2, float inreg %arg3, i32 inreg %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <2 x i32> %arg7, <3 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, <2 x i32> %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, float %arg20) #1 { +define amdgpu_ps void @sample_v3([17 x <16 x i8>] addrspace(2)* byval %arg, [32 x <16 x i8>] addrspace(2)* byval %arg1, [16 x <8 x i32>] addrspace(2)* byval %arg2, float inreg %arg3, i32 inreg %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <2 x i32> %arg7, <3 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, <2 x i32> %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, float %arg20) #0 { entry: %tmp = getelementptr [17 x <16 x i8>], [17 x <16 x i8>] addrspace(2)* %arg, i64 0, i32 0 %tmp21 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp, !tbaa !0 @@ -261,7 +258,7 @@ endif: ; preds = %else, %if %val.0 = phi float [ %val.if.0, %if ], [ %val.else.0, %else ] %val.1 = phi float [ %val.if.1, %if ], [ %val.else.1, %else ] %val.2 = phi float [ %val.if.2, %if ], [ %val.else.2, %else ] - call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 1, float %val.0, float %val.1, float %val.2, float 0.000000e+00) + call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %val.0, float %val.1, float %val.2, float 0.000000e+00, i1 true, i1 true) #0 ret void } @@ -294,7 +291,7 @@ endif: ; preds = %if1, %if0, %entry ; This test is just checking that we don't crash / assertion fail. ; CHECK-LABEL: {{^}}copy2: ; CHECK: s_endpgm -define amdgpu_ps void @copy2([17 x <16 x i8>] addrspace(2)* byval %arg, [32 x <16 x i8>] addrspace(2)* byval %arg1, [16 x <8 x i32>] addrspace(2)* byval %arg2, float inreg %arg3, i32 inreg %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <2 x i32> %arg7, <3 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, <2 x i32> %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, float %arg20) #1 { +define amdgpu_ps void @copy2([17 x <16 x i8>] addrspace(2)* byval %arg, [32 x <16 x i8>] addrspace(2)* byval %arg1, [16 x <8 x i32>] addrspace(2)* byval %arg2, float inreg %arg3, i32 inreg %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <2 x i32> %arg7, <3 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, <2 x i32> %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, float %arg20) #0 { entry: br label %LOOP68 @@ -308,7 +305,7 @@ LOOP68: ; preds = %ENDIF69, %entry IF70: ; preds = %LOOP68 %q = icmp ne i32 %l, 13 %temp.8 = select i1 %q, float 1.000000e+00, float 0.000000e+00 - call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 0, float %temp.8, float 0.000000e+00, float 0.000000e+00, float 1.000000e+00) + call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %temp.8, float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, i1 true, i1 true) #0 ret void ENDIF69: ; preds = %LOOP68 @@ -330,7 +327,7 @@ ENDIF69: ; preds = %LOOP68 ; [[END]]: ; CHECK: image_sample v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+}}:[[ADD]]{{\]}} ; CHECK: s_endpgm -define amdgpu_ps void @sample_rsrc([6 x <16 x i8>] addrspace(2)* byval %arg, [17 x <16 x i8>] addrspace(2)* byval %arg1, [16 x <4 x i32>] addrspace(2)* byval %arg2, [32 x <8 x i32>] addrspace(2)* byval %arg3, float inreg %arg4, i32 inreg %arg5, <2 x i32> %arg6, <2 x i32> %arg7, <2 x i32> %arg8, <3 x i32> %arg9, <2 x i32> %arg10, <2 x i32> %arg11, <2 x i32> %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, i32 %arg19, float %arg20, float %arg21) #1 { +define amdgpu_ps void @sample_rsrc([6 x <16 x i8>] addrspace(2)* byval %arg, [17 x <16 x i8>] addrspace(2)* byval %arg1, [16 x <4 x i32>] addrspace(2)* byval %arg2, [32 x <8 x i32>] addrspace(2)* byval %arg3, float inreg %arg4, i32 inreg %arg5, <2 x i32> %arg6, <2 x i32> %arg7, <2 x i32> %arg8, <3 x i32> %arg9, <2 x i32> %arg10, <2 x i32> %arg11, <2 x i32> %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, i32 %arg19, float %arg20, float %arg21) #0 { bb: %tmp = getelementptr [17 x <16 x i8>], [17 x <16 x i8>] addrspace(2)* %arg1, i32 0, i32 0 %tmp22 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp, !tbaa !3 @@ -343,14 +340,14 @@ bb: %j.i = extractelement <2 x i32> %arg7, i32 1 %i.f.i = bitcast i32 %i.i to float %j.f.i = bitcast i32 %j.i to float - %p1.i = call float @llvm.amdgcn.interp.p1(float %i.f.i, i32 0, i32 0, i32 %arg5) #1 - %p2.i = call float @llvm.amdgcn.interp.p2(float %p1.i, float %j.f.i, i32 0, i32 0, i32 %arg5) #1 + %p1.i = call float @llvm.amdgcn.interp.p1(float %i.f.i, i32 0, i32 0, i32 %arg5) #0 + %p2.i = call float @llvm.amdgcn.interp.p2(float %p1.i, float %j.f.i, i32 0, i32 0, i32 %arg5) #0 %i.i1 = extractelement <2 x i32> %arg7, i32 0 %j.i2 = extractelement <2 x i32> %arg7, i32 1 %i.f.i3 = bitcast i32 %i.i1 to float %j.f.i4 = bitcast i32 %j.i2 to float - %p1.i5 = call float @llvm.amdgcn.interp.p1(float %i.f.i3, i32 1, i32 0, i32 %arg5) #1 - %p2.i6 = call float @llvm.amdgcn.interp.p2(float %p1.i5, float %j.f.i4, i32 1, i32 0, i32 %arg5) #1 + %p1.i5 = call float @llvm.amdgcn.interp.p1(float %i.f.i3, i32 1, i32 0, i32 %arg5) #0 + %p2.i6 = call float @llvm.amdgcn.interp.p2(float %p1.i5, float %j.f.i4, i32 1, i32 0, i32 %arg5) #0 %tmp31 = bitcast float %tmp23 to i32 %tmp36 = icmp ne i32 %tmp31, 0 br i1 %tmp36, label %bb38, label %bb80 @@ -377,80 +374,58 @@ bb80: ; preds = %bb bb71: ; preds = %bb80, %bb38 %tmp72 = phi <4 x float> [ %tmp58, %bb38 ], [ %tmp87, %bb80 ] %tmp88 = extractelement <4 x float> %tmp72, i32 0 - call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 1, float %tmp88, float %tmp88, float %tmp88, float %tmp88) + call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp88, float %tmp88, float %tmp88, float %tmp88, i1 true, i1 true) #0 ret void } ; Check the the resource descriptor is stored in an sgpr. ; CHECK-LABEL: {{^}}mimg_srsrc_sgpr: ; CHECK: image_sample v{{[0-9]+}}, v[{{[0-9]+:[0-9]+}}], s[{{[0-9]+:[0-9]+}}], s[{{[0-9]+:[0-9]+}}] dmask:0x1 -define amdgpu_ps void @mimg_srsrc_sgpr([34 x <8 x i32>] addrspace(2)* byval %arg) #1 { +define amdgpu_ps void @mimg_srsrc_sgpr([34 x <8 x i32>] addrspace(2)* byval %arg) #0 { bb: - %tid = call i32 @llvm.amdgcn.mbcnt.lo(i32 -1, i32 0) #1 + %tid = call i32 @llvm.amdgcn.mbcnt.lo(i32 -1, i32 0) #0 %tmp7 = getelementptr [34 x <8 x i32>], [34 x <8 x i32>] addrspace(2)* %arg, i32 0, i32 %tid %tmp8 = load <8 x i32>, <8 x i32> addrspace(2)* %tmp7, align 32, !tbaa !0 %tmp9 = call <4 x float> @llvm.SI.image.sample.v2i32(<2 x i32> <i32 1061158912, i32 1048576000>, <8 x i32> %tmp8, <4 x i32> undef, i32 15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0) %tmp10 = extractelement <4 x float> %tmp9, i32 0 %tmp12 = call i32 @llvm.SI.packf16(float undef, float %tmp10) - %tmp13 = bitcast i32 %tmp12 to float - call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 1, float %tmp13, float undef, float undef, float undef) + %tmp13 = bitcast i32 %tmp12 to <2 x half> + call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 15, <2 x half> %tmp13, <2 x half> undef, i1 true, i1 true) #0 ret void } ; Check the the sampler is stored in an sgpr. ; CHECK-LABEL: {{^}}mimg_ssamp_sgpr: ; CHECK: image_sample v{{[0-9]+}}, v[{{[0-9]+:[0-9]+}}], s[{{[0-9]+:[0-9]+}}], s[{{[0-9]+:[0-9]+}}] dmask:0x1 -define amdgpu_ps void @mimg_ssamp_sgpr([17 x <4 x i32>] addrspace(2)* byval %arg) #1 { +define amdgpu_ps void @mimg_ssamp_sgpr([17 x <4 x i32>] addrspace(2)* byval %arg) #0 { bb: - %tid = call i32 @llvm.amdgcn.mbcnt.lo(i32 -1, i32 0) #1 + %tid = call i32 @llvm.amdgcn.mbcnt.lo(i32 -1, i32 0) #0 %tmp7 = getelementptr [17 x <4 x i32>], [17 x <4 x i32>] addrspace(2)* %arg, i32 0, i32 %tid %tmp8 = load <4 x i32>, <4 x i32> addrspace(2)* %tmp7, align 16, !tbaa !0 %tmp9 = call <4 x float> @llvm.SI.image.sample.v2i32(<2 x i32> <i32 1061158912, i32 1048576000>, <8 x i32> undef, <4 x i32> %tmp8, i32 15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0) %tmp10 = extractelement <4 x float> %tmp9, i32 0 %tmp12 = call i32 @llvm.SI.packf16(float %tmp10, float undef) - %tmp13 = bitcast i32 %tmp12 to float - call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 1, float %tmp13, float undef, float undef, float undef) + %tmp13 = bitcast i32 %tmp12 to <2 x half> + call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 15, <2 x half> %tmp13, <2 x half> undef, i1 true, i1 true) #0 ret void } -; Function Attrs: nounwind readnone -declare float @llvm.SI.load.const(<16 x i8>, i32) #0 - -; Function Attrs: nounwind readnone -declare float @llvm.fabs.f32(float) #0 - -declare void @llvm.SI.export(i32, i32, i32, i32, i32, float, float, float, float) - -; Function Attrs: nounwind readnone -declare <4 x float> @llvm.SI.sample.v2i32(<2 x i32>, <8 x i32>, <16 x i8>, i32) #0 - -; Function Attrs: nounwind readnone -declare float @llvm.amdgcn.rsq.f32(float) #0 - -; Function Attrs: nounwind readnone -declare float @llvm.exp2.f32(float) #0 - -; Function Attrs: nounwind readnone -declare float @llvm.pow.f32(float, float) #0 - -; Function Attrs: nounwind readnone -declare i32 @llvm.SI.packf16(float, float) #0 - -; Function Attrs: nounwind readnone -declare i32 @llvm.amdgcn.mbcnt.lo(i32, i32) #0 - -; Function Attrs: nounwind readnone -declare float @llvm.amdgcn.interp.p1(float, i32, i32, i32) #0 - -; Function Attrs: nounwind readnone -declare float @llvm.amdgcn.interp.p2(float, float, i32, i32, i32) #0 - -; Function Attrs: nounwind readnone -declare float @llvm.amdgcn.interp.mov(i32, i32, i32, i32) #0 - -attributes #0 = { nounwind readnone } -attributes #1 = { nounwind } -attributes #2 = { nounwind readonly } +declare float @llvm.fabs.f32(float) #1 +declare float @llvm.amdgcn.rsq.f32(float) #1 +declare float @llvm.exp2.f32(float) #1 +declare float @llvm.pow.f32(float, float) #1 +declare i32 @llvm.amdgcn.mbcnt.lo(i32, i32) #1 +declare float @llvm.amdgcn.interp.p1(float, i32, i32, i32) #1 +declare float @llvm.amdgcn.interp.p2(float, float, i32, i32, i32) #1 +declare void @llvm.amdgcn.exp.f32(i32, i32, float, float, float, float, i1, i1) #0 +declare void @llvm.amdgcn.exp.compr.v2f16(i32, i32, <2 x half>, <2 x half>, i1, i1) #0 + +declare i32 @llvm.SI.packf16(float, float) #1 +declare <4 x float> @llvm.SI.image.sample.v2i32(<2 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1 +declare float @llvm.SI.load.const(<16 x i8>, i32) #1 + +attributes #0 = { nounwind } +attributes #1 = { nounwind readnone } !0 = !{!1, !1, i64 0, i32 1} !1 = !{!"const", !2} |
