diff --git a/llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp b/llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp --- a/llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp @@ -331,8 +331,9 @@ getActionDefinitionsBuilder({G_UADDO, G_USUBO, G_UADDE, G_SADDE, G_USUBE, G_SSUBE}) .legalFor({{S32, S1}, {S32, S32}}) - .clampScalar(0, S32, S32) - .scalarize(0); // TODO: Implement. + .minScalar(0, S32) + // TODO: .scalarize(0) + .lower(); getActionDefinitionsBuilder(G_BITCAST) // Don't worry about the size constraint. diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-saddo.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-saddo.mir --- a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-saddo.mir +++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-saddo.mir @@ -2,6 +2,55 @@ # RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=tahiti -O0 -run-pass=legalizer -o - %s | FileCheck %s --- +name: test_saddo_s7 +body: | + bb.0: + liveins: $vgpr0, $vgpr1 + + ; CHECK-LABEL: name: test_saddo_s7 + ; CHECK: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 + ; CHECK: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[COPY]](s32) + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[COPY1]](s32) + ; CHECK: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY2]], [[COPY3]] + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0 + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[ADD]](s32) + ; CHECK: [[SEXT_INREG:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY4]], 7 + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[COPY]](s32) + ; CHECK: [[SEXT_INREG1:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY5]], 7 + ; CHECK: [[COPY6:%[0-9]+]]:_(s32) = COPY [[SEXT_INREG]](s32) + ; CHECK: [[SEXT_INREG2:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY6]], 8 + ; CHECK: [[COPY7:%[0-9]+]]:_(s32) = COPY [[SEXT_INREG1]](s32) + ; CHECK: [[SEXT_INREG3:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY7]], 8 + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG2]](s32), [[SEXT_INREG3]] + ; CHECK: [[COPY8:%[0-9]+]]:_(s32) = COPY [[COPY1]](s32) + ; CHECK: [[SEXT_INREG4:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY8]], 7 + ; CHECK: [[COPY9:%[0-9]+]]:_(s32) = COPY [[C]](s32) + ; CHECK: [[SEXT_INREG5:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY9]], 7 + ; CHECK: [[COPY10:%[0-9]+]]:_(s32) = COPY [[SEXT_INREG4]](s32) + ; CHECK: [[SEXT_INREG6:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY10]], 8 + ; CHECK: [[COPY11:%[0-9]+]]:_(s32) = COPY [[SEXT_INREG5]](s32) + ; CHECK: [[SEXT_INREG7:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY11]], 8 + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG6]](s32), [[SEXT_INREG7]] + ; CHECK: [[XOR:%[0-9]+]]:_(s1) = G_XOR [[ICMP1]], [[ICMP]] + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 127 + ; CHECK: [[COPY12:%[0-9]+]]:_(s32) = COPY [[ADD]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY12]], [[C1]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[XOR]](s1) + ; CHECK: $vgpr0 = COPY [[AND]](s32) + ; CHECK: $vgpr1 = COPY [[ZEXT]](s32) + %0:_(s32) = COPY $vgpr0 + %1:_(s32) = COPY $vgpr1 + %2:_(s7) = G_TRUNC %0 + %3:_(s7) = G_TRUNC %1 + %4:_(s7), %5:_(s1) = G_SADDO %2, %3 + %6:_(s32) = G_ZEXT %4 + %7:_(s32) = G_ZEXT %5 + $vgpr0 = COPY %6 + $vgpr1 = COPY %7 +... + +--- name: test_saddo_s16 body: | bb.0: @@ -94,6 +143,283 @@ ... --- +name: test_saddo_v2s16 +body: | + bb.0: + liveins: $vgpr0, $vgpr1 + + ; CHECK-LABEL: name: test_saddo_v2s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0 + ; CHECK: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1 + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[BITCAST]](s32) + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[BITCAST1]](s32) + ; CHECK: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY2]], [[COPY3]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32) + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[LSHR1]](s32) + ; CHECK: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[COPY4]], [[COPY5]] + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535 + ; CHECK: [[COPY6:%[0-9]+]]:_(s32) = COPY [[ADD]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY6]], [[C1]] + ; CHECK: [[COPY7:%[0-9]+]]:_(s32) = COPY [[ADD1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY7]], [[C1]] + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]] + ; CHECK: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[C2:%[0-9]+]]:_(s16) = G_CONSTANT i16 0 + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[COPY8:%[0-9]+]]:_(s32) = COPY [[ADD]](s32) + ; CHECK: [[SEXT_INREG:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY8]], 16 + ; CHECK: [[COPY9:%[0-9]+]]:_(s32) = COPY [[BITCAST3]](s32) + ; CHECK: [[SEXT_INREG1:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY9]], 16 + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG]](s32), [[SEXT_INREG1]] + ; CHECK: [[COPY10:%[0-9]+]]:_(s32) = COPY [[ADD1]](s32) + ; CHECK: [[SEXT_INREG2:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY10]], 16 + ; CHECK: [[COPY11:%[0-9]+]]:_(s32) = COPY [[LSHR2]](s32) + ; CHECK: [[SEXT_INREG3:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY11]], 16 + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG2]](s32), [[SEXT_INREG3]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32) + ; CHECK: [[COPY12:%[0-9]+]]:_(s32) = COPY [[BITCAST4]](s32) + ; CHECK: [[SEXT_INREG4:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY12]], 16 + ; CHECK: [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP2:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG4]](s32), [[SEXT]] + ; CHECK: [[COPY13:%[0-9]+]]:_(s32) = COPY [[LSHR3]](s32) + ; CHECK: [[SEXT_INREG5:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY13]], 16 + ; CHECK: [[SEXT1:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP3:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG5]](s32), [[SEXT1]] + ; CHECK: [[XOR:%[0-9]+]]:_(s1) = G_XOR [[ICMP2]], [[ICMP]] + ; CHECK: [[XOR1:%[0-9]+]]:_(s1) = G_XOR [[ICMP3]], [[ICMP1]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR1]](s1) + ; CHECK: [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY14:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY14]], [[C3]] + ; CHECK: [[COPY15:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND3:%[0-9]+]]:_(s32) = G_AND [[COPY15]], [[C3]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[AND2]](s32), [[AND3]](s32) + ; CHECK: $vgpr0 = COPY [[BITCAST2]](<2 x s16>) + ; CHECK: $vgpr1_vgpr2 = COPY [[BUILD_VECTOR]](<2 x s32>) + %0:_(<2 x s16>) = COPY $vgpr0 + %1:_(<2 x s16>) = COPY $vgpr1 + %2:_(<2 x s16>), %3:_(<2 x s1>) = G_SADDO %0, %1 + %4:_(<2 x s32>) = G_ZEXT %3 + $vgpr0 = COPY %2 + $vgpr1_vgpr2 = COPY %4 +... + +--- +name: test_saddo_v3s16 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_saddo_v3s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr1_vgpr2 + ; CHECK: [[EXTRACT:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[COPY]](<4 x s16>), 0 + ; CHECK: [[EXTRACT1:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[COPY]](<4 x s16>), 0 + ; CHECK: [[DEF:%[0-9]+]]:_(<4 x s16>) = G_IMPLICIT_DEF + ; CHECK: [[INSERT:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT]](<3 x s16>), 0 + ; CHECK: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT]](<4 x s16>) + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[INSERT1:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT1]](<3 x s16>), 0 + ; CHECK: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT1]](<4 x s16>) + ; CHECK: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[BITCAST]](s32) + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[BITCAST2]](s32) + ; CHECK: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY2]], [[COPY3]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32) + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[LSHR2]](s32) + ; CHECK: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[COPY4]], [[COPY5]] + ; CHECK: [[COPY6:%[0-9]+]]:_(s32) = COPY [[BITCAST1]](s32) + ; CHECK: [[COPY7:%[0-9]+]]:_(s32) = COPY [[BITCAST3]](s32) + ; CHECK: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[COPY6]], [[COPY7]] + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535 + ; CHECK: [[COPY8:%[0-9]+]]:_(s32) = COPY [[ADD]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY8]], [[C1]] + ; CHECK: [[COPY9:%[0-9]+]]:_(s32) = COPY [[ADD1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY9]], [[C1]] + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[COPY10:%[0-9]+]]:_(s32) = COPY [[ADD2]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY10]], [[C1]] + ; CHECK: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 0 + ; CHECK: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[C2]], [[C]](s32) + ; CHECK: [[OR1:%[0-9]+]]:_(s32) = G_OR [[AND2]], [[SHL1]] + ; CHECK: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32) + ; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>) + ; CHECK: [[EXTRACT2:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[CONCAT_VECTORS]](<4 x s16>), 0 + ; CHECK: [[C3:%[0-9]+]]:_(s16) = G_CONSTANT i16 0 + ; CHECK: [[INSERT2:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT1]](<3 x s16>), 0 + ; CHECK: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT2]](<4 x s16>) + ; CHECK: [[BITCAST6:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>) + ; CHECK: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST6]], [[C]](s32) + ; CHECK: [[BITCAST7:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>) + ; CHECK: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST7]], [[C]](s32) + ; CHECK: [[COPY11:%[0-9]+]]:_(s32) = COPY [[BITCAST6]](s32) + ; CHECK: [[AND3:%[0-9]+]]:_(s32) = G_AND [[COPY11]], [[C1]] + ; CHECK: [[COPY12:%[0-9]+]]:_(s32) = COPY [[LSHR4]](s32) + ; CHECK: [[AND4:%[0-9]+]]:_(s32) = G_AND [[COPY12]], [[C1]] + ; CHECK: [[COPY13:%[0-9]+]]:_(s32) = COPY [[BITCAST7]](s32) + ; CHECK: [[AND5:%[0-9]+]]:_(s32) = G_AND [[COPY13]], [[C1]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[AND3]](s32), [[AND4]](s32), [[AND5]](s32) + ; CHECK: [[INSERT3:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT2]](<3 x s16>), 0 + ; CHECK: [[DEF1:%[0-9]+]]:_(<4 x s32>) = G_IMPLICIT_DEF + ; CHECK: [[INSERT4:%[0-9]+]]:_(<4 x s32>) = G_INSERT [[DEF1]], [[BUILD_VECTOR]](<3 x s32>), 0 + ; CHECK: $vgpr0_vgpr1 = COPY [[INSERT3]](<4 x s16>) + ; CHECK: $vgpr2_vgpr3_vgpr4_vgpr5 = COPY [[INSERT4]](<4 x s32>) + %0:_(<4 x s16>) = COPY $vgpr0_vgpr1 + %1:_(<4 x s16>) = COPY $vgpr1_vgpr2 + %2:_(<3 x s16>) = G_EXTRACT %0(<4 x s16>), 0 + %3:_(<3 x s16>) = G_EXTRACT %0(<4 x s16>), 0 + %4:_(<3 x s16>), %5:_(<3 x s1>) = G_SADDO %2, %3 + %6:_(<3 x s32>) = G_ZEXT %3 + %7:_(<4 x s16>) = G_IMPLICIT_DEF + %8:_(<4 x s16>) = G_INSERT %7, %4(<3 x s16>), 0 + %9:_(<4 x s32>) = G_IMPLICIT_DEF + %10:_(<4 x s32>) = G_INSERT %9, %6(<3 x s32>), 0 + $vgpr0_vgpr1 = COPY %8 + $vgpr2_vgpr3_vgpr4_vgpr5 = COPY %10 +... + +--- +name: test_saddo_v4s16 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_saddo_v4s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr1_vgpr2 + ; CHECK: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>) + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>) + ; CHECK: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[BITCAST]](s32) + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[BITCAST2]](s32) + ; CHECK: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY2]], [[COPY3]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32) + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[LSHR2]](s32) + ; CHECK: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[COPY4]], [[COPY5]] + ; CHECK: [[COPY6:%[0-9]+]]:_(s32) = COPY [[BITCAST1]](s32) + ; CHECK: [[COPY7:%[0-9]+]]:_(s32) = COPY [[BITCAST3]](s32) + ; CHECK: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[COPY6]], [[COPY7]] + ; CHECK: [[COPY8:%[0-9]+]]:_(s32) = COPY [[LSHR1]](s32) + ; CHECK: [[COPY9:%[0-9]+]]:_(s32) = COPY [[LSHR3]](s32) + ; CHECK: [[ADD3:%[0-9]+]]:_(s32) = G_ADD [[COPY8]], [[COPY9]] + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535 + ; CHECK: [[COPY10:%[0-9]+]]:_(s32) = COPY [[ADD]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY10]], [[C1]] + ; CHECK: [[COPY11:%[0-9]+]]:_(s32) = COPY [[ADD1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY11]], [[C1]] + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[COPY12:%[0-9]+]]:_(s32) = COPY [[ADD2]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY12]], [[C1]] + ; CHECK: [[COPY13:%[0-9]+]]:_(s32) = COPY [[ADD3]](s32) + ; CHECK: [[AND3:%[0-9]+]]:_(s32) = G_AND [[COPY13]], [[C1]] + ; CHECK: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[AND3]], [[C]](s32) + ; CHECK: [[OR1:%[0-9]+]]:_(s32) = G_OR [[AND2]], [[SHL1]] + ; CHECK: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32) + ; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>) + ; CHECK: [[C2:%[0-9]+]]:_(s16) = G_CONSTANT i16 0 + ; CHECK: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>) + ; CHECK: [[BITCAST6:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>) + ; CHECK: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST6]], [[C]](s32) + ; CHECK: [[BITCAST7:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>) + ; CHECK: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST7]], [[C]](s32) + ; CHECK: [[COPY14:%[0-9]+]]:_(s32) = COPY [[ADD]](s32) + ; CHECK: [[SEXT_INREG:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY14]], 16 + ; CHECK: [[COPY15:%[0-9]+]]:_(s32) = COPY [[BITCAST6]](s32) + ; CHECK: [[SEXT_INREG1:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY15]], 16 + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG]](s32), [[SEXT_INREG1]] + ; CHECK: [[COPY16:%[0-9]+]]:_(s32) = COPY [[ADD1]](s32) + ; CHECK: [[SEXT_INREG2:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY16]], 16 + ; CHECK: [[COPY17:%[0-9]+]]:_(s32) = COPY [[LSHR4]](s32) + ; CHECK: [[SEXT_INREG3:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY17]], 16 + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG2]](s32), [[SEXT_INREG3]] + ; CHECK: [[COPY18:%[0-9]+]]:_(s32) = COPY [[ADD2]](s32) + ; CHECK: [[SEXT_INREG4:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY18]], 16 + ; CHECK: [[COPY19:%[0-9]+]]:_(s32) = COPY [[BITCAST7]](s32) + ; CHECK: [[SEXT_INREG5:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY19]], 16 + ; CHECK: [[ICMP2:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG4]](s32), [[SEXT_INREG5]] + ; CHECK: [[COPY20:%[0-9]+]]:_(s32) = COPY [[ADD3]](s32) + ; CHECK: [[SEXT_INREG6:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY20]], 16 + ; CHECK: [[COPY21:%[0-9]+]]:_(s32) = COPY [[LSHR5]](s32) + ; CHECK: [[SEXT_INREG7:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY21]], 16 + ; CHECK: [[ICMP3:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG6]](s32), [[SEXT_INREG7]] + ; CHECK: [[UV6:%[0-9]+]]:_(<2 x s16>), [[UV7:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>) + ; CHECK: [[BITCAST8:%[0-9]+]]:_(s32) = G_BITCAST [[UV6]](<2 x s16>) + ; CHECK: [[LSHR6:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST8]], [[C]](s32) + ; CHECK: [[BITCAST9:%[0-9]+]]:_(s32) = G_BITCAST [[UV7]](<2 x s16>) + ; CHECK: [[LSHR7:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST9]], [[C]](s32) + ; CHECK: [[COPY22:%[0-9]+]]:_(s32) = COPY [[BITCAST8]](s32) + ; CHECK: [[SEXT_INREG8:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY22]], 16 + ; CHECK: [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP4:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG8]](s32), [[SEXT]] + ; CHECK: [[COPY23:%[0-9]+]]:_(s32) = COPY [[LSHR6]](s32) + ; CHECK: [[SEXT_INREG9:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY23]], 16 + ; CHECK: [[SEXT1:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP5:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG9]](s32), [[SEXT1]] + ; CHECK: [[COPY24:%[0-9]+]]:_(s32) = COPY [[BITCAST9]](s32) + ; CHECK: [[SEXT_INREG10:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY24]], 16 + ; CHECK: [[SEXT2:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP6:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG10]](s32), [[SEXT2]] + ; CHECK: [[COPY25:%[0-9]+]]:_(s32) = COPY [[LSHR7]](s32) + ; CHECK: [[SEXT_INREG11:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY25]], 16 + ; CHECK: [[SEXT3:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP7:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG11]](s32), [[SEXT3]] + ; CHECK: [[XOR:%[0-9]+]]:_(s1) = G_XOR [[ICMP4]], [[ICMP]] + ; CHECK: [[XOR1:%[0-9]+]]:_(s1) = G_XOR [[ICMP5]], [[ICMP1]] + ; CHECK: [[XOR2:%[0-9]+]]:_(s1) = G_XOR [[ICMP6]], [[ICMP2]] + ; CHECK: [[XOR3:%[0-9]+]]:_(s1) = G_XOR [[ICMP7]], [[ICMP3]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR1]](s1) + ; CHECK: [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR2]](s1) + ; CHECK: [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR3]](s1) + ; CHECK: [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY26:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND4:%[0-9]+]]:_(s32) = G_AND [[COPY26]], [[C3]] + ; CHECK: [[COPY27:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND5:%[0-9]+]]:_(s32) = G_AND [[COPY27]], [[C3]] + ; CHECK: [[COPY28:%[0-9]+]]:_(s32) = COPY [[ANYEXT2]](s32) + ; CHECK: [[AND6:%[0-9]+]]:_(s32) = G_AND [[COPY28]], [[C3]] + ; CHECK: [[COPY29:%[0-9]+]]:_(s32) = COPY [[ANYEXT3]](s32) + ; CHECK: [[AND7:%[0-9]+]]:_(s32) = G_AND [[COPY29]], [[C3]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[AND4]](s32), [[AND5]](s32), [[AND6]](s32), [[AND7]](s32) + ; CHECK: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>) + ; CHECK: $vgpr2_vgpr3_vgpr4_vgpr5 = COPY [[BUILD_VECTOR]](<4 x s32>) + %0:_(<4 x s16>) = COPY $vgpr0_vgpr1 + %1:_(<4 x s16>) = COPY $vgpr1_vgpr2 + %2:_(<4 x s16>), %3:_(<4 x s1>) = G_SADDO %0, %1 + %4:_(<4 x s32>) = G_ZEXT %3 + $vgpr0_vgpr1 = COPY %2 + $vgpr2_vgpr3_vgpr4_vgpr5 = COPY %4 +... + +--- name: test_saddo_v2s32 body: | bb.0: diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-ssubo.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-ssubo.mir --- a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-ssubo.mir +++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-ssubo.mir @@ -2,6 +2,55 @@ # RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=tahiti -O0 -run-pass=legalizer -o - %s | FileCheck %s --- +name: test_ssubo_s7 +body: | + bb.0: + liveins: $vgpr0, $vgpr1 + + ; CHECK-LABEL: name: test_ssubo_s7 + ; CHECK: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 + ; CHECK: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[COPY]](s32) + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[COPY1]](s32) + ; CHECK: [[SUB:%[0-9]+]]:_(s32) = G_SUB [[COPY2]], [[COPY3]] + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0 + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[SUB]](s32) + ; CHECK: [[SEXT_INREG:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY4]], 7 + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[COPY]](s32) + ; CHECK: [[SEXT_INREG1:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY5]], 7 + ; CHECK: [[COPY6:%[0-9]+]]:_(s32) = COPY [[SEXT_INREG]](s32) + ; CHECK: [[SEXT_INREG2:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY6]], 8 + ; CHECK: [[COPY7:%[0-9]+]]:_(s32) = COPY [[SEXT_INREG1]](s32) + ; CHECK: [[SEXT_INREG3:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY7]], 8 + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG2]](s32), [[SEXT_INREG3]] + ; CHECK: [[COPY8:%[0-9]+]]:_(s32) = COPY [[COPY1]](s32) + ; CHECK: [[SEXT_INREG4:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY8]], 7 + ; CHECK: [[COPY9:%[0-9]+]]:_(s32) = COPY [[C]](s32) + ; CHECK: [[SEXT_INREG5:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY9]], 7 + ; CHECK: [[COPY10:%[0-9]+]]:_(s32) = COPY [[SEXT_INREG4]](s32) + ; CHECK: [[SEXT_INREG6:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY10]], 8 + ; CHECK: [[COPY11:%[0-9]+]]:_(s32) = COPY [[SEXT_INREG5]](s32) + ; CHECK: [[SEXT_INREG7:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY11]], 8 + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(sgt), [[SEXT_INREG6]](s32), [[SEXT_INREG7]] + ; CHECK: [[XOR:%[0-9]+]]:_(s1) = G_XOR [[ICMP1]], [[ICMP]] + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 127 + ; CHECK: [[COPY12:%[0-9]+]]:_(s32) = COPY [[SUB]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY12]], [[C1]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[XOR]](s1) + ; CHECK: $vgpr0 = COPY [[AND]](s32) + ; CHECK: $vgpr1 = COPY [[ZEXT]](s32) + %0:_(s32) = COPY $vgpr0 + %1:_(s32) = COPY $vgpr1 + %2:_(s7) = G_TRUNC %0 + %3:_(s7) = G_TRUNC %1 + %4:_(s7), %5:_(s1) = G_SSUBO %2, %3 + %6:_(s32) = G_ZEXT %4 + %7:_(s32) = G_ZEXT %5 + $vgpr0 = COPY %6 + $vgpr1 = COPY %7 +... + +--- name: test_ssubo_s16 body: | bb.0: @@ -94,6 +143,283 @@ ... --- +name: test_ssubo_v2s16 +body: | + bb.0: + liveins: $vgpr0, $vgpr1 + + ; CHECK-LABEL: name: test_ssubo_v2s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0 + ; CHECK: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1 + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[BITCAST]](s32) + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[BITCAST1]](s32) + ; CHECK: [[SUB:%[0-9]+]]:_(s32) = G_SUB [[COPY2]], [[COPY3]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32) + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[LSHR1]](s32) + ; CHECK: [[SUB1:%[0-9]+]]:_(s32) = G_SUB [[COPY4]], [[COPY5]] + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535 + ; CHECK: [[COPY6:%[0-9]+]]:_(s32) = COPY [[SUB]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY6]], [[C1]] + ; CHECK: [[COPY7:%[0-9]+]]:_(s32) = COPY [[SUB1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY7]], [[C1]] + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]] + ; CHECK: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[C2:%[0-9]+]]:_(s16) = G_CONSTANT i16 0 + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[COPY8:%[0-9]+]]:_(s32) = COPY [[SUB]](s32) + ; CHECK: [[SEXT_INREG:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY8]], 16 + ; CHECK: [[COPY9:%[0-9]+]]:_(s32) = COPY [[BITCAST3]](s32) + ; CHECK: [[SEXT_INREG1:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY9]], 16 + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG]](s32), [[SEXT_INREG1]] + ; CHECK: [[COPY10:%[0-9]+]]:_(s32) = COPY [[SUB1]](s32) + ; CHECK: [[SEXT_INREG2:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY10]], 16 + ; CHECK: [[COPY11:%[0-9]+]]:_(s32) = COPY [[LSHR2]](s32) + ; CHECK: [[SEXT_INREG3:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY11]], 16 + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG2]](s32), [[SEXT_INREG3]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32) + ; CHECK: [[COPY12:%[0-9]+]]:_(s32) = COPY [[BITCAST4]](s32) + ; CHECK: [[SEXT_INREG4:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY12]], 16 + ; CHECK: [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP2:%[0-9]+]]:_(s1) = G_ICMP intpred(sgt), [[SEXT_INREG4]](s32), [[SEXT]] + ; CHECK: [[COPY13:%[0-9]+]]:_(s32) = COPY [[LSHR3]](s32) + ; CHECK: [[SEXT_INREG5:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY13]], 16 + ; CHECK: [[SEXT1:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP3:%[0-9]+]]:_(s1) = G_ICMP intpred(sgt), [[SEXT_INREG5]](s32), [[SEXT1]] + ; CHECK: [[XOR:%[0-9]+]]:_(s1) = G_XOR [[ICMP2]], [[ICMP]] + ; CHECK: [[XOR1:%[0-9]+]]:_(s1) = G_XOR [[ICMP3]], [[ICMP1]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR1]](s1) + ; CHECK: [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY14:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY14]], [[C3]] + ; CHECK: [[COPY15:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND3:%[0-9]+]]:_(s32) = G_AND [[COPY15]], [[C3]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[AND2]](s32), [[AND3]](s32) + ; CHECK: $vgpr0 = COPY [[BITCAST2]](<2 x s16>) + ; CHECK: $vgpr1_vgpr2 = COPY [[BUILD_VECTOR]](<2 x s32>) + %0:_(<2 x s16>) = COPY $vgpr0 + %1:_(<2 x s16>) = COPY $vgpr1 + %2:_(<2 x s16>), %3:_(<2 x s1>) = G_SSUBO %0, %1 + %4:_(<2 x s32>) = G_ZEXT %3 + $vgpr0 = COPY %2 + $vgpr1_vgpr2 = COPY %4 +... + +--- +name: test_ssubo_v3s16 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_ssubo_v3s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr1_vgpr2 + ; CHECK: [[EXTRACT:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[COPY]](<4 x s16>), 0 + ; CHECK: [[EXTRACT1:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[COPY]](<4 x s16>), 0 + ; CHECK: [[DEF:%[0-9]+]]:_(<4 x s16>) = G_IMPLICIT_DEF + ; CHECK: [[INSERT:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT]](<3 x s16>), 0 + ; CHECK: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT]](<4 x s16>) + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[INSERT1:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT1]](<3 x s16>), 0 + ; CHECK: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT1]](<4 x s16>) + ; CHECK: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[BITCAST]](s32) + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[BITCAST2]](s32) + ; CHECK: [[SUB:%[0-9]+]]:_(s32) = G_SUB [[COPY2]], [[COPY3]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32) + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[LSHR2]](s32) + ; CHECK: [[SUB1:%[0-9]+]]:_(s32) = G_SUB [[COPY4]], [[COPY5]] + ; CHECK: [[COPY6:%[0-9]+]]:_(s32) = COPY [[BITCAST1]](s32) + ; CHECK: [[COPY7:%[0-9]+]]:_(s32) = COPY [[BITCAST3]](s32) + ; CHECK: [[SUB2:%[0-9]+]]:_(s32) = G_SUB [[COPY6]], [[COPY7]] + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535 + ; CHECK: [[COPY8:%[0-9]+]]:_(s32) = COPY [[SUB]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY8]], [[C1]] + ; CHECK: [[COPY9:%[0-9]+]]:_(s32) = COPY [[SUB1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY9]], [[C1]] + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[COPY10:%[0-9]+]]:_(s32) = COPY [[SUB2]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY10]], [[C1]] + ; CHECK: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 0 + ; CHECK: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[C2]], [[C]](s32) + ; CHECK: [[OR1:%[0-9]+]]:_(s32) = G_OR [[AND2]], [[SHL1]] + ; CHECK: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32) + ; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>) + ; CHECK: [[EXTRACT2:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[CONCAT_VECTORS]](<4 x s16>), 0 + ; CHECK: [[C3:%[0-9]+]]:_(s16) = G_CONSTANT i16 0 + ; CHECK: [[INSERT2:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT1]](<3 x s16>), 0 + ; CHECK: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT2]](<4 x s16>) + ; CHECK: [[BITCAST6:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>) + ; CHECK: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST6]], [[C]](s32) + ; CHECK: [[BITCAST7:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>) + ; CHECK: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST7]], [[C]](s32) + ; CHECK: [[COPY11:%[0-9]+]]:_(s32) = COPY [[BITCAST6]](s32) + ; CHECK: [[AND3:%[0-9]+]]:_(s32) = G_AND [[COPY11]], [[C1]] + ; CHECK: [[COPY12:%[0-9]+]]:_(s32) = COPY [[LSHR4]](s32) + ; CHECK: [[AND4:%[0-9]+]]:_(s32) = G_AND [[COPY12]], [[C1]] + ; CHECK: [[COPY13:%[0-9]+]]:_(s32) = COPY [[BITCAST7]](s32) + ; CHECK: [[AND5:%[0-9]+]]:_(s32) = G_AND [[COPY13]], [[C1]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[AND3]](s32), [[AND4]](s32), [[AND5]](s32) + ; CHECK: [[INSERT3:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT2]](<3 x s16>), 0 + ; CHECK: [[DEF1:%[0-9]+]]:_(<4 x s32>) = G_IMPLICIT_DEF + ; CHECK: [[INSERT4:%[0-9]+]]:_(<4 x s32>) = G_INSERT [[DEF1]], [[BUILD_VECTOR]](<3 x s32>), 0 + ; CHECK: $vgpr0_vgpr1 = COPY [[INSERT3]](<4 x s16>) + ; CHECK: $vgpr2_vgpr3_vgpr4_vgpr5 = COPY [[INSERT4]](<4 x s32>) + %0:_(<4 x s16>) = COPY $vgpr0_vgpr1 + %1:_(<4 x s16>) = COPY $vgpr1_vgpr2 + %2:_(<3 x s16>) = G_EXTRACT %0(<4 x s16>), 0 + %3:_(<3 x s16>) = G_EXTRACT %0(<4 x s16>), 0 + %4:_(<3 x s16>), %5:_(<3 x s1>) = G_SSUBO %2, %3 + %6:_(<3 x s32>) = G_ZEXT %3 + %7:_(<4 x s16>) = G_IMPLICIT_DEF + %8:_(<4 x s16>) = G_INSERT %7, %4(<3 x s16>), 0 + %9:_(<4 x s32>) = G_IMPLICIT_DEF + %10:_(<4 x s32>) = G_INSERT %9, %6(<3 x s32>), 0 + $vgpr0_vgpr1 = COPY %8 + $vgpr2_vgpr3_vgpr4_vgpr5 = COPY %10 +... + +--- +name: test_ssubo_v4s16 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_ssubo_v4s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr1_vgpr2 + ; CHECK: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>) + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>) + ; CHECK: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[BITCAST]](s32) + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[BITCAST2]](s32) + ; CHECK: [[SUB:%[0-9]+]]:_(s32) = G_SUB [[COPY2]], [[COPY3]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32) + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[LSHR2]](s32) + ; CHECK: [[SUB1:%[0-9]+]]:_(s32) = G_SUB [[COPY4]], [[COPY5]] + ; CHECK: [[COPY6:%[0-9]+]]:_(s32) = COPY [[BITCAST1]](s32) + ; CHECK: [[COPY7:%[0-9]+]]:_(s32) = COPY [[BITCAST3]](s32) + ; CHECK: [[SUB2:%[0-9]+]]:_(s32) = G_SUB [[COPY6]], [[COPY7]] + ; CHECK: [[COPY8:%[0-9]+]]:_(s32) = COPY [[LSHR1]](s32) + ; CHECK: [[COPY9:%[0-9]+]]:_(s32) = COPY [[LSHR3]](s32) + ; CHECK: [[SUB3:%[0-9]+]]:_(s32) = G_SUB [[COPY8]], [[COPY9]] + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535 + ; CHECK: [[COPY10:%[0-9]+]]:_(s32) = COPY [[SUB]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY10]], [[C1]] + ; CHECK: [[COPY11:%[0-9]+]]:_(s32) = COPY [[SUB1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY11]], [[C1]] + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[COPY12:%[0-9]+]]:_(s32) = COPY [[SUB2]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY12]], [[C1]] + ; CHECK: [[COPY13:%[0-9]+]]:_(s32) = COPY [[SUB3]](s32) + ; CHECK: [[AND3:%[0-9]+]]:_(s32) = G_AND [[COPY13]], [[C1]] + ; CHECK: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[AND3]], [[C]](s32) + ; CHECK: [[OR1:%[0-9]+]]:_(s32) = G_OR [[AND2]], [[SHL1]] + ; CHECK: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32) + ; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>) + ; CHECK: [[C2:%[0-9]+]]:_(s16) = G_CONSTANT i16 0 + ; CHECK: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>) + ; CHECK: [[BITCAST6:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>) + ; CHECK: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST6]], [[C]](s32) + ; CHECK: [[BITCAST7:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>) + ; CHECK: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST7]], [[C]](s32) + ; CHECK: [[COPY14:%[0-9]+]]:_(s32) = COPY [[SUB]](s32) + ; CHECK: [[SEXT_INREG:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY14]], 16 + ; CHECK: [[COPY15:%[0-9]+]]:_(s32) = COPY [[BITCAST6]](s32) + ; CHECK: [[SEXT_INREG1:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY15]], 16 + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG]](s32), [[SEXT_INREG1]] + ; CHECK: [[COPY16:%[0-9]+]]:_(s32) = COPY [[SUB1]](s32) + ; CHECK: [[SEXT_INREG2:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY16]], 16 + ; CHECK: [[COPY17:%[0-9]+]]:_(s32) = COPY [[LSHR4]](s32) + ; CHECK: [[SEXT_INREG3:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY17]], 16 + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG2]](s32), [[SEXT_INREG3]] + ; CHECK: [[COPY18:%[0-9]+]]:_(s32) = COPY [[SUB2]](s32) + ; CHECK: [[SEXT_INREG4:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY18]], 16 + ; CHECK: [[COPY19:%[0-9]+]]:_(s32) = COPY [[BITCAST7]](s32) + ; CHECK: [[SEXT_INREG5:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY19]], 16 + ; CHECK: [[ICMP2:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG4]](s32), [[SEXT_INREG5]] + ; CHECK: [[COPY20:%[0-9]+]]:_(s32) = COPY [[SUB3]](s32) + ; CHECK: [[SEXT_INREG6:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY20]], 16 + ; CHECK: [[COPY21:%[0-9]+]]:_(s32) = COPY [[LSHR5]](s32) + ; CHECK: [[SEXT_INREG7:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY21]], 16 + ; CHECK: [[ICMP3:%[0-9]+]]:_(s1) = G_ICMP intpred(slt), [[SEXT_INREG6]](s32), [[SEXT_INREG7]] + ; CHECK: [[UV6:%[0-9]+]]:_(<2 x s16>), [[UV7:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>) + ; CHECK: [[BITCAST8:%[0-9]+]]:_(s32) = G_BITCAST [[UV6]](<2 x s16>) + ; CHECK: [[LSHR6:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST8]], [[C]](s32) + ; CHECK: [[BITCAST9:%[0-9]+]]:_(s32) = G_BITCAST [[UV7]](<2 x s16>) + ; CHECK: [[LSHR7:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST9]], [[C]](s32) + ; CHECK: [[COPY22:%[0-9]+]]:_(s32) = COPY [[BITCAST8]](s32) + ; CHECK: [[SEXT_INREG8:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY22]], 16 + ; CHECK: [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP4:%[0-9]+]]:_(s1) = G_ICMP intpred(sgt), [[SEXT_INREG8]](s32), [[SEXT]] + ; CHECK: [[COPY23:%[0-9]+]]:_(s32) = COPY [[LSHR6]](s32) + ; CHECK: [[SEXT_INREG9:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY23]], 16 + ; CHECK: [[SEXT1:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP5:%[0-9]+]]:_(s1) = G_ICMP intpred(sgt), [[SEXT_INREG9]](s32), [[SEXT1]] + ; CHECK: [[COPY24:%[0-9]+]]:_(s32) = COPY [[BITCAST9]](s32) + ; CHECK: [[SEXT_INREG10:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY24]], 16 + ; CHECK: [[SEXT2:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP6:%[0-9]+]]:_(s1) = G_ICMP intpred(sgt), [[SEXT_INREG10]](s32), [[SEXT2]] + ; CHECK: [[COPY25:%[0-9]+]]:_(s32) = COPY [[LSHR7]](s32) + ; CHECK: [[SEXT_INREG11:%[0-9]+]]:_(s32) = G_SEXT_INREG [[COPY25]], 16 + ; CHECK: [[SEXT3:%[0-9]+]]:_(s32) = G_SEXT [[C2]](s16) + ; CHECK: [[ICMP7:%[0-9]+]]:_(s1) = G_ICMP intpred(sgt), [[SEXT_INREG11]](s32), [[SEXT3]] + ; CHECK: [[XOR:%[0-9]+]]:_(s1) = G_XOR [[ICMP4]], [[ICMP]] + ; CHECK: [[XOR1:%[0-9]+]]:_(s1) = G_XOR [[ICMP5]], [[ICMP1]] + ; CHECK: [[XOR2:%[0-9]+]]:_(s1) = G_XOR [[ICMP6]], [[ICMP2]] + ; CHECK: [[XOR3:%[0-9]+]]:_(s1) = G_XOR [[ICMP7]], [[ICMP3]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR1]](s1) + ; CHECK: [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR2]](s1) + ; CHECK: [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[XOR3]](s1) + ; CHECK: [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY26:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND4:%[0-9]+]]:_(s32) = G_AND [[COPY26]], [[C3]] + ; CHECK: [[COPY27:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND5:%[0-9]+]]:_(s32) = G_AND [[COPY27]], [[C3]] + ; CHECK: [[COPY28:%[0-9]+]]:_(s32) = COPY [[ANYEXT2]](s32) + ; CHECK: [[AND6:%[0-9]+]]:_(s32) = G_AND [[COPY28]], [[C3]] + ; CHECK: [[COPY29:%[0-9]+]]:_(s32) = COPY [[ANYEXT3]](s32) + ; CHECK: [[AND7:%[0-9]+]]:_(s32) = G_AND [[COPY29]], [[C3]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[AND4]](s32), [[AND5]](s32), [[AND6]](s32), [[AND7]](s32) + ; CHECK: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>) + ; CHECK: $vgpr2_vgpr3_vgpr4_vgpr5 = COPY [[BUILD_VECTOR]](<4 x s32>) + %0:_(<4 x s16>) = COPY $vgpr0_vgpr1 + %1:_(<4 x s16>) = COPY $vgpr1_vgpr2 + %2:_(<4 x s16>), %3:_(<4 x s1>) = G_SSUBO %0, %1 + %4:_(<4 x s32>) = G_ZEXT %3 + $vgpr0_vgpr1 = COPY %2 + $vgpr2_vgpr3_vgpr4_vgpr5 = COPY %4 +... + +--- name: test_ssubo_v2s32 body: | bb.0: diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-uaddo.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-uaddo.mir --- a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-uaddo.mir +++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-uaddo.mir @@ -89,3 +89,272 @@ $vgpr0 = COPY %6 $vgpr1 = COPY %7 ... + +--- +name: test_uaddo_s64 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_uaddo_s64 + ; CHECK: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3 + ; CHECK: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64) + ; CHECK: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64) + ; CHECK: [[UADDO:%[0-9]+]]:_(s32), [[UADDO1:%[0-9]+]]:_(s1) = G_UADDO [[UV]], [[UV2]] + ; CHECK: [[UADDE:%[0-9]+]]:_(s32), [[UADDE1:%[0-9]+]]:_(s1) = G_UADDE [[UV1]], [[UV3]], [[UADDO1]] + ; CHECK: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO]](s32), [[UADDE]](s32) + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[MV]](s64), [[COPY1]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ICMP]](s1) + ; CHECK: $vgpr0_vgpr1 = COPY [[MV]](s64) + ; CHECK: $vgpr2 = COPY [[ZEXT]](s32) + %0:_(s64) = COPY $vgpr0_vgpr1 + %1:_(s64) = COPY $vgpr2_vgpr3 + %2:_(s64), %3:_(s1) = G_UADDO %0, %1 + %4:_(s32) = G_ZEXT %3 + $vgpr0_vgpr1 = COPY %2 + $vgpr2 = COPY %4 +... + +--- +name: test_uaddo_v2s16 +body: | + bb.0: + liveins: $vgpr0, $vgpr1 + + ; CHECK-LABEL: name: test_uaddo_v2s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0 + ; CHECK: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1 + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>) + ; CHECK: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>) + ; CHECK: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32) + ; CHECK: [[ADD:%[0-9]+]]:_(s16) = G_ADD [[TRUNC]], [[TRUNC2]] + ; CHECK: [[ADD1:%[0-9]+]]:_(s16) = G_ADD [[TRUNC1]], [[TRUNC3]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ADD]](s16) + ; CHECK: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[ADD1]](s16) + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]] + ; CHECK: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>) + ; CHECK: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32) + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[ADD]](s16), [[TRUNC4]] + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[ADD1]](s16), [[TRUNC5]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP1]](s1) + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C1]] + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C1]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[AND]](s32), [[AND1]](s32) + ; CHECK: $vgpr0 = COPY [[BITCAST2]](<2 x s16>) + ; CHECK: $vgpr1_vgpr2 = COPY [[BUILD_VECTOR]](<2 x s32>) + %0:_(<2 x s16>) = COPY $vgpr0 + %1:_(<2 x s16>) = COPY $vgpr1 + %2:_(<2 x s16>), %3:_(<2 x s1>) = G_UADDO %0, %1 + %4:_(<2 x s32>) = G_ZEXT %3 + $vgpr0 = COPY %2 + $vgpr1_vgpr2 = COPY %4 +... + +--- +name: test_uaddo_v3s16 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_uaddo_v3s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr1_vgpr2 + ; CHECK: [[EXTRACT:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[COPY]](<4 x s16>), 0 + ; CHECK: [[EXTRACT1:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[COPY]](<4 x s16>), 0 + ; CHECK: [[DEF:%[0-9]+]]:_(<4 x s16>) = G_IMPLICIT_DEF + ; CHECK: [[INSERT:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT]](<3 x s16>), 0 + ; CHECK: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT]](<4 x s16>) + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>) + ; CHECK: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>) + ; CHECK: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[INSERT1:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT1]](<3 x s16>), 0 + ; CHECK: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT1]](<4 x s16>) + ; CHECK: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>) + ; CHECK: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32) + ; CHECK: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>) + ; CHECK: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[ADD:%[0-9]+]]:_(s16) = G_ADD [[TRUNC]], [[TRUNC3]] + ; CHECK: [[ADD1:%[0-9]+]]:_(s16) = G_ADD [[TRUNC1]], [[TRUNC4]] + ; CHECK: [[ADD2:%[0-9]+]]:_(s16) = G_ADD [[TRUNC2]], [[TRUNC5]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ADD]](s16) + ; CHECK: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[ADD1]](s16) + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[ADD2]](s16) + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 0 + ; CHECK: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[C1]], [[C]](s32) + ; CHECK: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]] + ; CHECK: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32) + ; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>) + ; CHECK: [[EXTRACT2:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[CONCAT_VECTORS]](<4 x s16>), 0 + ; CHECK: [[INSERT2:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT1]](<3 x s16>), 0 + ; CHECK: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT2]](<4 x s16>) + ; CHECK: [[BITCAST6:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>) + ; CHECK: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST6]], [[C]](s32) + ; CHECK: [[BITCAST7:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>) + ; CHECK: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST7]], [[C]](s32) + ; CHECK: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[BITCAST6]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C2]] + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[LSHR4]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C2]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[BITCAST7]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY4]], [[C2]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[AND]](s32), [[AND1]](s32), [[AND2]](s32) + ; CHECK: [[INSERT3:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT2]](<3 x s16>), 0 + ; CHECK: [[DEF1:%[0-9]+]]:_(<4 x s32>) = G_IMPLICIT_DEF + ; CHECK: [[INSERT4:%[0-9]+]]:_(<4 x s32>) = G_INSERT [[DEF1]], [[BUILD_VECTOR]](<3 x s32>), 0 + ; CHECK: $vgpr0_vgpr1 = COPY [[INSERT3]](<4 x s16>) + ; CHECK: $vgpr2_vgpr3_vgpr4_vgpr5 = COPY [[INSERT4]](<4 x s32>) + %0:_(<4 x s16>) = COPY $vgpr0_vgpr1 + %1:_(<4 x s16>) = COPY $vgpr1_vgpr2 + %2:_(<3 x s16>) = G_EXTRACT %0(<4 x s16>), 0 + %3:_(<3 x s16>) = G_EXTRACT %0(<4 x s16>), 0 + %4:_(<3 x s16>), %5:_(<3 x s1>) = G_UADDO %2, %3 + %6:_(<3 x s32>) = G_ZEXT %3 + %7:_(<4 x s16>) = G_IMPLICIT_DEF + %8:_(<4 x s16>) = G_INSERT %7, %4(<3 x s16>), 0 + %9:_(<4 x s32>) = G_IMPLICIT_DEF + %10:_(<4 x s32>) = G_INSERT %9, %6(<3 x s32>), 0 + $vgpr0_vgpr1 = COPY %8 + $vgpr2_vgpr3_vgpr4_vgpr5 = COPY %10 +... + +--- +name: test_uaddo_v4s16 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_uaddo_v4s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr1_vgpr2 + ; CHECK: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>) + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>) + ; CHECK: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>) + ; CHECK: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32) + ; CHECK: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>) + ; CHECK: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>) + ; CHECK: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32) + ; CHECK: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>) + ; CHECK: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32) + ; CHECK: [[ADD:%[0-9]+]]:_(s16) = G_ADD [[TRUNC]], [[TRUNC4]] + ; CHECK: [[ADD1:%[0-9]+]]:_(s16) = G_ADD [[TRUNC1]], [[TRUNC5]] + ; CHECK: [[ADD2:%[0-9]+]]:_(s16) = G_ADD [[TRUNC2]], [[TRUNC6]] + ; CHECK: [[ADD3:%[0-9]+]]:_(s16) = G_ADD [[TRUNC3]], [[TRUNC7]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ADD]](s16) + ; CHECK: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[ADD1]](s16) + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[ADD2]](s16) + ; CHECK: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[ADD3]](s16) + ; CHECK: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32) + ; CHECK: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]] + ; CHECK: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32) + ; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>) + ; CHECK: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>) + ; CHECK: [[BITCAST6:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>) + ; CHECK: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST6]](s32) + ; CHECK: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST6]], [[C]](s32) + ; CHECK: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32) + ; CHECK: [[BITCAST7:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>) + ; CHECK: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST7]](s32) + ; CHECK: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST7]], [[C]](s32) + ; CHECK: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32) + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[ADD]](s16), [[TRUNC8]] + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[ADD1]](s16), [[TRUNC9]] + ; CHECK: [[ICMP2:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[ADD2]](s16), [[TRUNC10]] + ; CHECK: [[ICMP3:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[ADD3]](s16), [[TRUNC11]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP1]](s1) + ; CHECK: [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP2]](s1) + ; CHECK: [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP3]](s1) + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C1]] + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C1]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[ANYEXT2]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY4]], [[C1]] + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[ANYEXT3]](s32) + ; CHECK: [[AND3:%[0-9]+]]:_(s32) = G_AND [[COPY5]], [[C1]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[AND]](s32), [[AND1]](s32), [[AND2]](s32), [[AND3]](s32) + ; CHECK: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>) + ; CHECK: $vgpr2_vgpr3_vgpr4_vgpr5 = COPY [[BUILD_VECTOR]](<4 x s32>) + %0:_(<4 x s16>) = COPY $vgpr0_vgpr1 + %1:_(<4 x s16>) = COPY $vgpr1_vgpr2 + %2:_(<4 x s16>), %3:_(<4 x s1>) = G_UADDO %0, %1 + %4:_(<4 x s32>) = G_ZEXT %3 + $vgpr0_vgpr1 = COPY %2 + $vgpr2_vgpr3_vgpr4_vgpr5 = COPY %4 +... + +--- +name: test_uaddo_v2s32 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_uaddo_v2s32 + ; CHECK: [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3 + ; CHECK: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) + ; CHECK: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) + ; CHECK: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[UV]], [[UV2]] + ; CHECK: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[UV1]], [[UV3]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[ADD]](s32), [[ADD1]](s32) + ; CHECK: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[ADD]](s32), [[UV4]] + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[ADD1]](s32), [[UV5]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP1]](s1) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C]] + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C]] + ; CHECK: [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[AND]](s32), [[AND1]](s32) + ; CHECK: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>) + ; CHECK: $vgpr2_vgpr3 = COPY [[BUILD_VECTOR1]](<2 x s32>) + %0:_(<2 x s32>) = COPY $vgpr0_vgpr1 + %1:_(<2 x s32>) = COPY $vgpr2_vgpr3 + %2:_(<2 x s32>), %3:_(<2 x s1>) = G_UADDO %0, %1 + %4:_(<2 x s32>) = G_ZEXT %3 + $vgpr0_vgpr1 = COPY %2 + $vgpr2_vgpr3 = COPY %4 +... diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-usubo.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-usubo.mir --- a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-usubo.mir +++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-usubo.mir @@ -89,3 +89,286 @@ $vgpr0 = COPY %6 $vgpr1 = COPY %7 ... + +--- +name: test_usubo_s64 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_usubo_s64 + ; CHECK: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3 + ; CHECK: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](s64) + ; CHECK: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64) + ; CHECK: [[USUBO:%[0-9]+]]:_(s32), [[USUBO1:%[0-9]+]]:_(s1) = G_USUBO [[UV]], [[UV2]] + ; CHECK: [[USUBE:%[0-9]+]]:_(s32), [[USUBE1:%[0-9]+]]:_(s1) = G_USUBE [[UV1]], [[UV3]], [[USUBO1]] + ; CHECK: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[USUBO]](s32), [[USUBE]](s32) + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[COPY]](s64), [[COPY1]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ICMP]](s1) + ; CHECK: $vgpr0_vgpr1 = COPY [[MV]](s64) + ; CHECK: $vgpr2 = COPY [[ZEXT]](s32) + %0:_(s64) = COPY $vgpr0_vgpr1 + %1:_(s64) = COPY $vgpr2_vgpr3 + %2:_(s64), %3:_(s1) = G_USUBO %0, %1 + %4:_(s32) = G_ZEXT %3 + $vgpr0_vgpr1 = COPY %2 + $vgpr2 = COPY %4 +... + +--- +name: test_usubo_v2s16 +body: | + bb.0: + liveins: $vgpr0, $vgpr1 + + ; CHECK-LABEL: name: test_usubo_v2s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0 + ; CHECK: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1 + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>) + ; CHECK: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>) + ; CHECK: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32) + ; CHECK: [[SUB:%[0-9]+]]:_(s16) = G_SUB [[TRUNC]], [[TRUNC2]] + ; CHECK: [[SUB1:%[0-9]+]]:_(s16) = G_SUB [[TRUNC1]], [[TRUNC3]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[SUB]](s16) + ; CHECK: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[SUB1]](s16) + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]] + ; CHECK: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>) + ; CHECK: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32) + ; CHECK: [[BITCAST4:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>) + ; CHECK: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST4]](s32) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST4]], [[C]](s32) + ; CHECK: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32) + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[TRUNC4]](s16), [[TRUNC6]] + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[TRUNC5]](s16), [[TRUNC7]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP1]](s1) + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C1]] + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C1]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[AND]](s32), [[AND1]](s32) + ; CHECK: $vgpr0 = COPY [[BITCAST2]](<2 x s16>) + ; CHECK: $vgpr1_vgpr2 = COPY [[BUILD_VECTOR]](<2 x s32>) + %0:_(<2 x s16>) = COPY $vgpr0 + %1:_(<2 x s16>) = COPY $vgpr1 + %2:_(<2 x s16>), %3:_(<2 x s1>) = G_USUBO %0, %1 + %4:_(<2 x s32>) = G_ZEXT %3 + $vgpr0 = COPY %2 + $vgpr1_vgpr2 = COPY %4 +... + +--- +name: test_usubo_v3s16 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_usubo_v3s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr1_vgpr2 + ; CHECK: [[EXTRACT:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[COPY]](<4 x s16>), 0 + ; CHECK: [[EXTRACT1:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[COPY]](<4 x s16>), 0 + ; CHECK: [[DEF:%[0-9]+]]:_(<4 x s16>) = G_IMPLICIT_DEF + ; CHECK: [[INSERT:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT]](<3 x s16>), 0 + ; CHECK: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT]](<4 x s16>) + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>) + ; CHECK: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>) + ; CHECK: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[INSERT1:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT1]](<3 x s16>), 0 + ; CHECK: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT1]](<4 x s16>) + ; CHECK: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>) + ; CHECK: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32) + ; CHECK: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>) + ; CHECK: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[SUB:%[0-9]+]]:_(s16) = G_SUB [[TRUNC]], [[TRUNC3]] + ; CHECK: [[SUB1:%[0-9]+]]:_(s16) = G_SUB [[TRUNC1]], [[TRUNC4]] + ; CHECK: [[SUB2:%[0-9]+]]:_(s16) = G_SUB [[TRUNC2]], [[TRUNC5]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[SUB]](s16) + ; CHECK: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[SUB1]](s16) + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[SUB2]](s16) + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 0 + ; CHECK: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[C1]], [[C]](s32) + ; CHECK: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]] + ; CHECK: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32) + ; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>) + ; CHECK: [[EXTRACT2:%[0-9]+]]:_(<3 x s16>) = G_EXTRACT [[CONCAT_VECTORS]](<4 x s16>), 0 + ; CHECK: [[INSERT2:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT1]](<3 x s16>), 0 + ; CHECK: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[INSERT2]](<4 x s16>) + ; CHECK: [[BITCAST6:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>) + ; CHECK: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST6]], [[C]](s32) + ; CHECK: [[BITCAST7:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>) + ; CHECK: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST7]], [[C]](s32) + ; CHECK: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[BITCAST6]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C2]] + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[LSHR4]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C2]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[BITCAST7]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY4]], [[C2]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[AND]](s32), [[AND1]](s32), [[AND2]](s32) + ; CHECK: [[INSERT3:%[0-9]+]]:_(<4 x s16>) = G_INSERT [[DEF]], [[EXTRACT2]](<3 x s16>), 0 + ; CHECK: [[DEF1:%[0-9]+]]:_(<4 x s32>) = G_IMPLICIT_DEF + ; CHECK: [[INSERT4:%[0-9]+]]:_(<4 x s32>) = G_INSERT [[DEF1]], [[BUILD_VECTOR]](<3 x s32>), 0 + ; CHECK: $vgpr0_vgpr1 = COPY [[INSERT3]](<4 x s16>) + ; CHECK: $vgpr2_vgpr3_vgpr4_vgpr5 = COPY [[INSERT4]](<4 x s32>) + %0:_(<4 x s16>) = COPY $vgpr0_vgpr1 + %1:_(<4 x s16>) = COPY $vgpr1_vgpr2 + %2:_(<3 x s16>) = G_EXTRACT %0(<4 x s16>), 0 + %3:_(<3 x s16>) = G_EXTRACT %0(<4 x s16>), 0 + %4:_(<3 x s16>), %5:_(<3 x s1>) = G_USUBO %2, %3 + %6:_(<3 x s32>) = G_ZEXT %3 + %7:_(<4 x s16>) = G_IMPLICIT_DEF + %8:_(<4 x s16>) = G_INSERT %7, %4(<3 x s16>), 0 + %9:_(<4 x s32>) = G_IMPLICIT_DEF + %10:_(<4 x s32>) = G_INSERT %9, %6(<3 x s32>), 0 + $vgpr0_vgpr1 = COPY %8 + $vgpr2_vgpr3_vgpr4_vgpr5 = COPY %10 +... + +--- +name: test_usubo_v4s16 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_usubo_v4s16 + ; CHECK: [[COPY:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<4 x s16>) = COPY $vgpr1_vgpr2 + ; CHECK: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>) + ; CHECK: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[UV]](<2 x s16>) + ; CHECK: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16 + ; CHECK: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32) + ; CHECK: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32) + ; CHECK: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[UV1]](<2 x s16>) + ; CHECK: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32) + ; CHECK: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32) + ; CHECK: [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32) + ; CHECK: [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>) + ; CHECK: [[BITCAST2:%[0-9]+]]:_(s32) = G_BITCAST [[UV2]](<2 x s16>) + ; CHECK: [[TRUNC4:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST2]](s32) + ; CHECK: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST2]], [[C]](s32) + ; CHECK: [[TRUNC5:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR2]](s32) + ; CHECK: [[BITCAST3:%[0-9]+]]:_(s32) = G_BITCAST [[UV3]](<2 x s16>) + ; CHECK: [[TRUNC6:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST3]](s32) + ; CHECK: [[LSHR3:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST3]], [[C]](s32) + ; CHECK: [[TRUNC7:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR3]](s32) + ; CHECK: [[SUB:%[0-9]+]]:_(s16) = G_SUB [[TRUNC]], [[TRUNC4]] + ; CHECK: [[SUB1:%[0-9]+]]:_(s16) = G_SUB [[TRUNC1]], [[TRUNC5]] + ; CHECK: [[SUB2:%[0-9]+]]:_(s16) = G_SUB [[TRUNC2]], [[TRUNC6]] + ; CHECK: [[SUB3:%[0-9]+]]:_(s16) = G_SUB [[TRUNC3]], [[TRUNC7]] + ; CHECK: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[SUB]](s16) + ; CHECK: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[SUB1]](s16) + ; CHECK: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32) + ; CHECK: [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]] + ; CHECK: [[BITCAST4:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32) + ; CHECK: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[SUB2]](s16) + ; CHECK: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[SUB3]](s16) + ; CHECK: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[ZEXT3]], [[C]](s32) + ; CHECK: [[OR1:%[0-9]+]]:_(s32) = G_OR [[ZEXT2]], [[SHL1]] + ; CHECK: [[BITCAST5:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR1]](s32) + ; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[BITCAST4]](<2 x s16>), [[BITCAST5]](<2 x s16>) + ; CHECK: [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY]](<4 x s16>) + ; CHECK: [[BITCAST6:%[0-9]+]]:_(s32) = G_BITCAST [[UV4]](<2 x s16>) + ; CHECK: [[TRUNC8:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST6]](s32) + ; CHECK: [[LSHR4:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST6]], [[C]](s32) + ; CHECK: [[TRUNC9:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR4]](s32) + ; CHECK: [[BITCAST7:%[0-9]+]]:_(s32) = G_BITCAST [[UV5]](<2 x s16>) + ; CHECK: [[TRUNC10:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST7]](s32) + ; CHECK: [[LSHR5:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST7]], [[C]](s32) + ; CHECK: [[TRUNC11:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR5]](s32) + ; CHECK: [[UV6:%[0-9]+]]:_(<2 x s16>), [[UV7:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[COPY1]](<4 x s16>) + ; CHECK: [[BITCAST8:%[0-9]+]]:_(s32) = G_BITCAST [[UV6]](<2 x s16>) + ; CHECK: [[TRUNC12:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST8]](s32) + ; CHECK: [[LSHR6:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST8]], [[C]](s32) + ; CHECK: [[TRUNC13:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR6]](s32) + ; CHECK: [[BITCAST9:%[0-9]+]]:_(s32) = G_BITCAST [[UV7]](<2 x s16>) + ; CHECK: [[TRUNC14:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST9]](s32) + ; CHECK: [[LSHR7:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST9]], [[C]](s32) + ; CHECK: [[TRUNC15:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR7]](s32) + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[TRUNC8]](s16), [[TRUNC12]] + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[TRUNC9]](s16), [[TRUNC13]] + ; CHECK: [[ICMP2:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[TRUNC10]](s16), [[TRUNC14]] + ; CHECK: [[ICMP3:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[TRUNC11]](s16), [[TRUNC15]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP1]](s1) + ; CHECK: [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP2]](s1) + ; CHECK: [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP3]](s1) + ; CHECK: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C1]] + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C1]] + ; CHECK: [[COPY4:%[0-9]+]]:_(s32) = COPY [[ANYEXT2]](s32) + ; CHECK: [[AND2:%[0-9]+]]:_(s32) = G_AND [[COPY4]], [[C1]] + ; CHECK: [[COPY5:%[0-9]+]]:_(s32) = COPY [[ANYEXT3]](s32) + ; CHECK: [[AND3:%[0-9]+]]:_(s32) = G_AND [[COPY5]], [[C1]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[AND]](s32), [[AND1]](s32), [[AND2]](s32), [[AND3]](s32) + ; CHECK: $vgpr0_vgpr1 = COPY [[CONCAT_VECTORS]](<4 x s16>) + ; CHECK: $vgpr2_vgpr3_vgpr4_vgpr5 = COPY [[BUILD_VECTOR]](<4 x s32>) + %0:_(<4 x s16>) = COPY $vgpr0_vgpr1 + %1:_(<4 x s16>) = COPY $vgpr1_vgpr2 + %2:_(<4 x s16>), %3:_(<4 x s1>) = G_USUBO %0, %1 + %4:_(<4 x s32>) = G_ZEXT %3 + $vgpr0_vgpr1 = COPY %2 + $vgpr2_vgpr3_vgpr4_vgpr5 = COPY %4 +... + +--- +name: test_usubo_v2s32 +body: | + bb.0: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3 + + ; CHECK-LABEL: name: test_usubo_v2s32 + ; CHECK: [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr0_vgpr1 + ; CHECK: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3 + ; CHECK: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) + ; CHECK: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) + ; CHECK: [[SUB:%[0-9]+]]:_(s32) = G_SUB [[UV]], [[UV2]] + ; CHECK: [[SUB1:%[0-9]+]]:_(s32) = G_SUB [[UV1]], [[UV3]] + ; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[SUB]](s32), [[SUB1]](s32) + ; CHECK: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) + ; CHECK: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) + ; CHECK: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[UV4]](s32), [[UV6]] + ; CHECK: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(ult), [[UV5]](s32), [[UV7]] + ; CHECK: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP]](s1) + ; CHECK: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ICMP1]](s1) + ; CHECK: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 1 + ; CHECK: [[COPY2:%[0-9]+]]:_(s32) = COPY [[ANYEXT]](s32) + ; CHECK: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C]] + ; CHECK: [[COPY3:%[0-9]+]]:_(s32) = COPY [[ANYEXT1]](s32) + ; CHECK: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C]] + ; CHECK: [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[AND]](s32), [[AND1]](s32) + ; CHECK: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>) + ; CHECK: $vgpr2_vgpr3 = COPY [[BUILD_VECTOR1]](<2 x s32>) + %0:_(<2 x s32>) = COPY $vgpr0_vgpr1 + %1:_(<2 x s32>) = COPY $vgpr2_vgpr3 + %2:_(<2 x s32>), %3:_(<2 x s1>) = G_USUBO %0, %1 + %4:_(<2 x s32>) = G_ZEXT %3 + $vgpr0_vgpr1 = COPY %2 + $vgpr2_vgpr3 = COPY %4 +...