diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2019-12-20 21:39:19 +0530 |
|---|---|---|
| committer | Matt Arsenault <arsenm2@gmail.com> | 2019-12-21 04:55:36 -0500 |
| commit | 42a26445f9e68d55bcc4b8d1b4ced83a56d7743c (patch) | |
| tree | fe2acb04784c9e2617a687d09cdb4f119346d7ab | |
| parent | dff3f8d74240144c35e78978a73646aa34faf400 (diff) | |
| download | bcm5719-llvm-42a26445f9e68d55bcc4b8d1b4ced83a56d7743c.tar.gz bcm5719-llvm-42a26445f9e68d55bcc4b8d1b4ced83a56d7743c.zip | |
AMDGPU/GlobalISel: Fix misuse of div_scale intrinsics
Confusingly, the intrinsic operands do not match the
instruction/custom node. The order is shuffled, and the 3rd operand is
an immediate to select operands.
I'm not 100% sure I did this right, but fdiv still doesn't select end
to end and it will be easier to tell when it does. This at least
avoids an assertion in RegBankSelect and allows hitting the fallback
on selection.
| -rw-r--r-- | llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp | 10 | ||||
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-fdiv.mir | 304 |
2 files changed, 157 insertions, 157 deletions
diff --git a/llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp b/llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp index c21102fd611..055bde28cd5 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp @@ -2028,14 +2028,14 @@ bool AMDGPULegalizerInfo::legalizeFDIV32(MachineInstr &MI, auto DenominatorScaled = B.buildIntrinsic(Intrinsic::amdgcn_div_scale, {S32, S1}, false) .addUse(RHS) - .addUse(RHS) .addUse(LHS) + .addImm(1) .setMIFlags(Flags); auto NumeratorScaled = B.buildIntrinsic(Intrinsic::amdgcn_div_scale, {S32, S1}, false) .addUse(LHS) .addUse(RHS) - .addUse(LHS) + .addImm(0) .setMIFlags(Flags); auto ApproxRcp = B.buildIntrinsic(Intrinsic::amdgcn_rcp, {S32}, false) @@ -2091,9 +2091,9 @@ bool AMDGPULegalizerInfo::legalizeFDIV64(MachineInstr &MI, auto One = B.buildFConstant(S64, 1.0); auto DivScale0 = B.buildIntrinsic(Intrinsic::amdgcn_div_scale, {S64, S1}, false) - .addUse(RHS) - .addUse(RHS) .addUse(LHS) + .addUse(RHS) + .addImm(1) .setMIFlags(Flags); auto NegDivScale0 = B.buildFNeg(S64, DivScale0.getReg(0), Flags); @@ -2109,7 +2109,7 @@ bool AMDGPULegalizerInfo::legalizeFDIV64(MachineInstr &MI, auto DivScale1 = B.buildIntrinsic(Intrinsic::amdgcn_div_scale, {S64, S1}, false) .addUse(LHS) .addUse(RHS) - .addUse(LHS) + .addImm(0) .setMIFlags(Flags); auto Fma3 = B.buildFMA(S64, Fma1, Fma2, Fma1, Flags); diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-fdiv.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-fdiv.mir index 25dc9c234e7..64ee900bd19 100644 --- a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-fdiv.mir +++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-fdiv.mir @@ -24,8 +24,8 @@ body: | ; SI: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16) ; SI: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16) ; SI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -111,8 +111,8 @@ body: | ; SI: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; SI: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; SI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY1]](s32), [[COPY]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), [[COPY]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -128,8 +128,8 @@ body: | ; VI: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; VI: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; VI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY1]](s32), [[COPY]](s32) - ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), [[COPY]](s32) + ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY]](s32), 1 + ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), 0 ; VI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; VI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; VI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -145,8 +145,8 @@ body: | ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; GFX9: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY1]](s32), [[COPY]](s32) - ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), [[COPY]](s32) + ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY]](s32), 1 + ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), 0 ; GFX9: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX9: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX9: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -168,8 +168,8 @@ body: | ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; GFX10: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY1]](s32), [[COPY]](s32) - ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), [[COPY]](s32) + ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY]](s32), 1 + ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), 0 ; GFX10: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX10: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX10: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -202,8 +202,8 @@ body: | ; SI: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; SI: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; SI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY1]](s32), [[COPY]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), [[COPY]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: S_SETREG_IMM32_B32 3, 2305 @@ -221,8 +221,8 @@ body: | ; VI: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; VI: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; VI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY1]](s32), [[COPY]](s32) - ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), [[COPY]](s32) + ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY]](s32), 1 + ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), 0 ; VI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; VI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; VI: S_SETREG_IMM32_B32 3, 2305 @@ -240,8 +240,8 @@ body: | ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; GFX9: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY1]](s32), [[COPY]](s32) - ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), [[COPY]](s32) + ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY]](s32), 1 + ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), 0 ; GFX9: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX9: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX9: S_SETREG_IMM32_B32 3, 2305 @@ -265,8 +265,8 @@ body: | ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; GFX10: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY1]](s32), [[COPY]](s32) - ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), [[COPY]](s32) + ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s32), [[COPY]](s32), 1 + ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY1]](s32), 0 ; GFX10: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX10: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX10: S_DENORM_MODE 15 @@ -348,13 +348,13 @@ body: | ; SI: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 ; SI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3 ; SI: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s64), [[COPY1]](s64), [[COPY]](s64) + ; SI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), 1 ; SI: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; SI: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; SI: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; SI: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; SI: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; SI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), [[COPY]](s64) + ; SI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), 0 ; SI: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; SI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -380,13 +380,13 @@ body: | ; VI: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 ; VI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3 ; VI: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s64), [[COPY1]](s64), [[COPY]](s64) + ; VI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), 1 ; VI: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; VI: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; VI: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; VI: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; VI: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; VI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), [[COPY]](s64) + ; VI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), 0 ; VI: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; VI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -405,13 +405,13 @@ body: | ; GFX9: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 ; GFX9: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3 ; GFX9: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s64), [[COPY1]](s64), [[COPY]](s64) + ; GFX9: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), 1 ; GFX9: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; GFX9: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; GFX9: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; GFX9: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; GFX9: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; GFX9: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), [[COPY]](s64) + ; GFX9: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), 0 ; GFX9: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; GFX9: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -436,13 +436,13 @@ body: | ; GFX10: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 ; GFX10: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3 ; GFX10: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY1]](s64), [[COPY1]](s64), [[COPY]](s64) + ; GFX10: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), 1 ; GFX10: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; GFX10: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; GFX10: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; GFX10: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; GFX10: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; GFX10: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), [[COPY]](s64) + ; GFX10: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY1]](s64), 0 ; GFX10: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; GFX10: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -480,8 +480,8 @@ body: | ; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) ; SI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) ; SI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV2]](s32), [[UV]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), [[UV]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: S_SETREG_IMM32_B32 3, 2305 @@ -494,8 +494,8 @@ body: | ; SI: S_SETREG_IMM32_B32 0, 2305 ; SI: [[INT5:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; SI: [[INT6:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV2]](s32), [[UV]](s32) - ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV1]](s32) - ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), [[UV1]](s32) + ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV1]](s32), 1 + ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), 0 ; SI: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; SI: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; SI: S_SETREG_IMM32_B32 3, 2305 @@ -516,8 +516,8 @@ body: | ; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) ; VI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) ; VI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV2]](s32), [[UV]](s32) - ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), [[UV]](s32) + ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV]](s32), 1 + ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), 0 ; VI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; VI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; VI: S_SETREG_IMM32_B32 3, 2305 @@ -530,8 +530,8 @@ body: | ; VI: S_SETREG_IMM32_B32 0, 2305 ; VI: [[INT5:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; VI: [[INT6:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV2]](s32), [[UV]](s32) - ; VI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV1]](s32) - ; VI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), [[UV1]](s32) + ; VI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV1]](s32), 1 + ; VI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), 0 ; VI: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; VI: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; VI: S_SETREG_IMM32_B32 3, 2305 @@ -552,8 +552,8 @@ body: | ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) ; GFX9: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) ; GFX9: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV2]](s32), [[UV]](s32) - ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), [[UV]](s32) + ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV]](s32), 1 + ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), 0 ; GFX9: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX9: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX9: S_SETREG_IMM32_B32 3, 2305 @@ -566,8 +566,8 @@ body: | ; GFX9: S_SETREG_IMM32_B32 0, 2305 ; GFX9: [[INT5:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; GFX9: [[INT6:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV2]](s32), [[UV]](s32) - ; GFX9: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV1]](s32) - ; GFX9: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), [[UV1]](s32) + ; GFX9: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV1]](s32), 1 + ; GFX9: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), 0 ; GFX9: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; GFX9: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; GFX9: S_SETREG_IMM32_B32 3, 2305 @@ -599,8 +599,8 @@ body: | ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) ; GFX10: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) ; GFX10: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV2]](s32), [[UV]](s32) - ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), [[UV]](s32) + ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV]](s32), 1 + ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), 0 ; GFX10: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX10: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX10: S_DENORM_MODE 15 @@ -613,8 +613,8 @@ body: | ; GFX10: S_DENORM_MODE 12 ; GFX10: [[INT5:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; GFX10: [[INT6:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV2]](s32), [[UV]](s32) - ; GFX10: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV1]](s32) - ; GFX10: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), [[UV1]](s32) + ; GFX10: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV1]](s32), 1 + ; GFX10: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), 0 ; GFX10: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; GFX10: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; GFX10: S_DENORM_MODE 15 @@ -647,8 +647,8 @@ body: | ; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) ; SI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) ; SI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV2]](s32), [[UV]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), [[UV]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = nnan G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG]], [[INT4]], [[C]] @@ -659,8 +659,8 @@ body: | ; SI: [[FMA4:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG]], [[FMA3]], [[INT2]] ; SI: [[INT5:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; SI: [[INT6:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV2]](s32), [[UV]](s32) - ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV1]](s32) - ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), [[UV1]](s32) + ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV1]](s32), 1 + ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), 0 ; SI: [[INT11:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; SI: [[FNEG1:%[0-9]+]]:_(s32) = nnan G_FNEG [[INT7]] ; SI: [[FMA5:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG1]], [[INT11]], [[C]] @@ -679,8 +679,8 @@ body: | ; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) ; VI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) ; VI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV2]](s32), [[UV]](s32) - ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), [[UV]](s32) + ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV]](s32), 1 + ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), 0 ; VI: [[INT4:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; VI: [[FNEG:%[0-9]+]]:_(s32) = nnan G_FNEG [[INT]] ; VI: [[FMA:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG]], [[INT4]], [[C]] @@ -691,8 +691,8 @@ body: | ; VI: [[FMA4:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG]], [[FMA3]], [[INT2]] ; VI: [[INT5:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; VI: [[INT6:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV2]](s32), [[UV]](s32) - ; VI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV1]](s32) - ; VI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), [[UV1]](s32) + ; VI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV1]](s32), 1 + ; VI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), 0 ; VI: [[INT11:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; VI: [[FNEG1:%[0-9]+]]:_(s32) = nnan G_FNEG [[INT7]] ; VI: [[FMA5:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG1]], [[INT11]], [[C]] @@ -711,8 +711,8 @@ body: | ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) ; GFX9: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) ; GFX9: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV2]](s32), [[UV]](s32) - ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), [[UV]](s32) + ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV]](s32), 1 + ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), 0 ; GFX9: [[INT4:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX9: [[FNEG:%[0-9]+]]:_(s32) = nnan G_FNEG [[INT]] ; GFX9: [[FMA:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG]], [[INT4]], [[C]] @@ -723,8 +723,8 @@ body: | ; GFX9: [[FMA4:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG]], [[FMA3]], [[INT2]] ; GFX9: [[INT5:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; GFX9: [[INT6:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV2]](s32), [[UV]](s32) - ; GFX9: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV1]](s32) - ; GFX9: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), [[UV1]](s32) + ; GFX9: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV1]](s32), 1 + ; GFX9: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), 0 ; GFX9: [[INT11:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; GFX9: [[FNEG1:%[0-9]+]]:_(s32) = nnan G_FNEG [[INT7]] ; GFX9: [[FMA5:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG1]], [[INT11]], [[C]] @@ -754,8 +754,8 @@ body: | ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>) ; GFX10: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>) ; GFX10: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV2]](s32), [[UV]](s32) - ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), [[UV]](s32) + ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV]](s32), 1 + ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV2]](s32), 0 ; GFX10: [[INT4:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX10: [[FNEG:%[0-9]+]]:_(s32) = nnan G_FNEG [[INT]] ; GFX10: [[FMA:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG]], [[INT4]], [[C]] @@ -766,8 +766,8 @@ body: | ; GFX10: [[FMA4:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG]], [[FMA3]], [[INT2]] ; GFX10: [[INT5:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; GFX10: [[INT6:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV2]](s32), [[UV]](s32) - ; GFX10: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV1]](s32) - ; GFX10: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), [[UV1]](s32) + ; GFX10: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV1]](s32), 1 + ; GFX10: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV3]](s32), 0 ; GFX10: [[INT11:%[0-9]+]]:_(s32) = nnan G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; GFX10: [[FNEG1:%[0-9]+]]:_(s32) = nnan G_FNEG [[INT7]] ; GFX10: [[FMA5:%[0-9]+]]:_(s32) = nnan G_FMA [[FNEG1]], [[INT11]], [[C]] @@ -798,8 +798,8 @@ body: | ; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<3 x s32>) ; SI: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<3 x s32>) ; SI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV3]](s32), [[UV]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV3]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -810,8 +810,8 @@ body: | ; SI: [[FMA4:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[FMA3]], [[INT2]] ; SI: [[INT5:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; SI: [[INT6:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV3]](s32), [[UV]](s32) - ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV4]](s32), [[UV4]](s32), [[UV1]](s32) - ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV4]](s32), [[UV1]](s32) + ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV4]](s32), [[UV1]](s32), 1 + ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV4]](s32), 0 ; SI: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; SI: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; SI: [[FMA5:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[INT11]], [[C]] @@ -822,8 +822,8 @@ body: | ; SI: [[FMA9:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[FMA8]], [[INT9]] ; SI: [[INT12:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA9]](s32), [[FMA6]](s32), [[FMA8]](s32), [[INT10]](s1) ; SI: [[INT13:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT12]](s32), [[UV4]](s32), [[UV1]](s32) - ; SI: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV5]](s32), [[UV5]](s32), [[UV2]](s32) - ; SI: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV5]](s32), [[UV2]](s32) + ; SI: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV5]](s32), [[UV2]](s32), 1 + ; SI: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV5]](s32), 0 ; SI: [[INT18:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT14]](s32) ; SI: [[FNEG2:%[0-9]+]]:_(s32) = G_FNEG [[INT14]] ; SI: [[FMA10:%[0-9]+]]:_(s32) = G_FMA [[FNEG2]], [[INT18]], [[C]] @@ -842,8 +842,8 @@ body: | ; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<3 x s32>) ; VI: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<3 x s32>) ; VI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV]](s32) - ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV3]](s32), [[UV]](s32) + ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV]](s32), 1 + ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV3]](s32), 0 ; VI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; VI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; VI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -854,8 +854,8 @@ body: | ; VI: [[FMA4:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[FMA3]], [[INT2]] ; VI: [[INT5:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; VI: [[INT6:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV3]](s32), [[UV]](s32) - ; VI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV4]](s32), [[UV4]](s32), [[UV1]](s32) - ; VI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV4]](s32), [[UV1]](s32) + ; VI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV4]](s32), [[UV1]](s32), 1 + ; VI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV4]](s32), 0 ; VI: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; VI: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; VI: [[FMA5:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[INT11]], [[C]] @@ -866,8 +866,8 @@ body: | ; VI: [[FMA9:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[FMA8]], [[INT9]] ; VI: [[INT12:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA9]](s32), [[FMA6]](s32), [[FMA8]](s32), [[INT10]](s1) ; VI: [[INT13:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT12]](s32), [[UV4]](s32), [[UV1]](s32) - ; VI: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV5]](s32), [[UV5]](s32), [[UV2]](s32) - ; VI: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV5]](s32), [[UV2]](s32) + ; VI: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV5]](s32), [[UV2]](s32), 1 + ; VI: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV5]](s32), 0 ; VI: [[INT18:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT14]](s32) ; VI: [[FNEG2:%[0-9]+]]:_(s32) = G_FNEG [[INT14]] ; VI: [[FMA10:%[0-9]+]]:_(s32) = G_FMA [[FNEG2]], [[INT18]], [[C]] @@ -886,8 +886,8 @@ body: | ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<3 x s32>) ; GFX9: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<3 x s32>) ; GFX9: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV]](s32) - ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV3]](s32), [[UV]](s32) + ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV]](s32), 1 + ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV3]](s32), 0 ; GFX9: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX9: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX9: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -898,8 +898,8 @@ body: | ; GFX9: [[FMA4:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[FMA3]], [[INT2]] ; GFX9: [[INT5:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; GFX9: [[INT6:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV3]](s32), [[UV]](s32) - ; GFX9: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV4]](s32), [[UV4]](s32), [[UV1]](s32) - ; GFX9: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV4]](s32), [[UV1]](s32) + ; GFX9: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV4]](s32), [[UV1]](s32), 1 + ; GFX9: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV4]](s32), 0 ; GFX9: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; GFX9: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; GFX9: [[FMA5:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[INT11]], [[C]] @@ -910,8 +910,8 @@ body: | ; GFX9: [[FMA9:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[FMA8]], [[INT9]] ; GFX9: [[INT12:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA9]](s32), [[FMA6]](s32), [[FMA8]](s32), [[INT10]](s1) ; GFX9: [[INT13:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT12]](s32), [[UV4]](s32), [[UV1]](s32) - ; GFX9: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV5]](s32), [[UV5]](s32), [[UV2]](s32) - ; GFX9: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV5]](s32), [[UV2]](s32) + ; GFX9: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV5]](s32), [[UV2]](s32), 1 + ; GFX9: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV5]](s32), 0 ; GFX9: [[INT18:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT14]](s32) ; GFX9: [[FNEG2:%[0-9]+]]:_(s32) = G_FNEG [[INT14]] ; GFX9: [[FMA10:%[0-9]+]]:_(s32) = G_FMA [[FNEG2]], [[INT18]], [[C]] @@ -943,8 +943,8 @@ body: | ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<3 x s32>) ; GFX10: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<3 x s32>) ; GFX10: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV3]](s32), [[UV]](s32) - ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV3]](s32), [[UV]](s32) + ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s32), [[UV]](s32), 1 + ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s32), [[UV3]](s32), 0 ; GFX10: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX10: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX10: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -955,8 +955,8 @@ body: | ; GFX10: [[FMA4:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[FMA3]], [[INT2]] ; GFX10: [[INT5:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s32), [[FMA1]](s32), [[FMA3]](s32), [[INT3]](s1) ; GFX10: [[INT6:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT5]](s32), [[UV3]](s32), [[UV]](s32) - ; GFX10: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV4]](s32), [[UV4]](s32), [[UV1]](s32) - ; GFX10: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV4]](s32), [[UV1]](s32) + ; GFX10: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV4]](s32), [[UV1]](s32), 1 + ; GFX10: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s32), [[UV4]](s32), 0 ; GFX10: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; GFX10: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; GFX10: [[FMA5:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[INT11]], [[C]] @@ -967,8 +967,8 @@ body: | ; GFX10: [[FMA9:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[FMA8]], [[INT9]] ; GFX10: [[INT12:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA9]](s32), [[FMA6]](s32), [[FMA8]](s32), [[INT10]](s1) ; GFX10: [[INT13:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), [[INT12]](s32), [[UV4]](s32), [[UV1]](s32) - ; GFX10: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV5]](s32), [[UV5]](s32), [[UV2]](s32) - ; GFX10: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV5]](s32), [[UV2]](s32) + ; GFX10: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV5]](s32), [[UV2]](s32), 1 + ; GFX10: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s32), [[UV5]](s32), 0 ; GFX10: [[INT18:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT14]](s32) ; GFX10: [[FNEG2:%[0-9]+]]:_(s32) = G_FNEG [[INT14]] ; GFX10: [[FMA10:%[0-9]+]]:_(s32) = G_FMA [[FNEG2]], [[INT18]], [[C]] @@ -999,13 +999,13 @@ body: | ; SI: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>) ; SI: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>) ; SI: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s64), [[UV2]](s64), [[UV]](s64) + ; SI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), 1 ; SI: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; SI: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; SI: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; SI: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; SI: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; SI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), [[UV]](s64) + ; SI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), 0 ; SI: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; SI: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; SI: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -1026,13 +1026,13 @@ body: | ; SI: [[XOR:%[0-9]+]]:_(s1) = G_XOR [[ICMP]], [[ICMP1]] ; SI: [[INT5:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s64), [[FMA3]](s64), [[MV]](s64), [[XOR]](s1) ; SI: [[INT6:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), def %7(s64), [[INT5]](s64), [[UV2]](s64), [[UV]](s64) - ; SI: [[INT7:%[0-9]+]]:_(s64), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s64), [[UV3]](s64), [[UV1]](s64) + ; SI: [[INT7:%[0-9]+]]:_(s64), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), 1 ; SI: [[FNEG1:%[0-9]+]]:_(s64) = G_FNEG [[INT7]] ; SI: [[INT9:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s64) ; SI: [[FMA5:%[0-9]+]]:_(s64) = G_FMA [[FNEG1]], [[INT9]], [[C]] ; SI: [[FMA6:%[0-9]+]]:_(s64) = G_FMA [[INT9]], [[FMA5]], [[INT9]] ; SI: [[FMA7:%[0-9]+]]:_(s64) = G_FMA [[FNEG1]], [[FMA6]], [[C]] - ; SI: [[INT10:%[0-9]+]]:_(s64), [[INT11:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), [[UV1]](s64) + ; SI: [[INT10:%[0-9]+]]:_(s64), [[INT11:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), 0 ; SI: [[FMA8:%[0-9]+]]:_(s64) = G_FMA [[FMA6]], [[FMA7]], [[FMA6]] ; SI: [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT10]](s64) ; SI: [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA8]](s64) @@ -1061,13 +1061,13 @@ body: | ; VI: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>) ; VI: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>) ; VI: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s64), [[UV2]](s64), [[UV]](s64) + ; VI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), 1 ; VI: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; VI: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; VI: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; VI: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; VI: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; VI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), [[UV]](s64) + ; VI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), 0 ; VI: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; VI: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; VI: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -1081,13 +1081,13 @@ body: | ; VI: [[FMA4:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[MV]], [[INT3]] ; VI: [[INT5:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s64), [[FMA3]](s64), [[MV]](s64), [[INT4]](s1) ; VI: [[INT6:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), def %7(s64), [[INT5]](s64), [[UV2]](s64), [[UV]](s64) - ; VI: [[INT7:%[0-9]+]]:_(s64), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s64), [[UV3]](s64), [[UV1]](s64) + ; VI: [[INT7:%[0-9]+]]:_(s64), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), 1 ; VI: [[FNEG1:%[0-9]+]]:_(s64) = G_FNEG [[INT7]] ; VI: [[INT9:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s64) ; VI: [[FMA5:%[0-9]+]]:_(s64) = G_FMA [[FNEG1]], [[INT9]], [[C]] ; VI: [[FMA6:%[0-9]+]]:_(s64) = G_FMA [[INT9]], [[FMA5]], [[INT9]] ; VI: [[FMA7:%[0-9]+]]:_(s64) = G_FMA [[FNEG1]], [[FMA6]], [[C]] - ; VI: [[INT10:%[0-9]+]]:_(s64), [[INT11:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), [[UV1]](s64) + ; VI: [[INT10:%[0-9]+]]:_(s64), [[INT11:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), 0 ; VI: [[FMA8:%[0-9]+]]:_(s64) = G_FMA [[FMA6]], [[FMA7]], [[FMA6]] ; VI: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT10]](s64) ; VI: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA8]](s64) @@ -1109,13 +1109,13 @@ body: | ; GFX9: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>) ; GFX9: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>) ; GFX9: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s64), [[UV2]](s64), [[UV]](s64) + ; GFX9: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), 1 ; GFX9: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; GFX9: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; GFX9: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; GFX9: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; GFX9: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; GFX9: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), [[UV]](s64) + ; GFX9: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), 0 ; GFX9: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; GFX9: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; GFX9: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -1129,13 +1129,13 @@ body: | ; GFX9: [[FMA4:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[MV]], [[INT3]] ; GFX9: [[INT5:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s64), [[FMA3]](s64), [[MV]](s64), [[INT4]](s1) ; GFX9: [[INT6:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), def %7(s64), [[INT5]](s64), [[UV2]](s64), [[UV]](s64) - ; GFX9: [[INT7:%[0-9]+]]:_(s64), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s64), [[UV3]](s64), [[UV1]](s64) + ; GFX9: [[INT7:%[0-9]+]]:_(s64), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), 1 ; GFX9: [[FNEG1:%[0-9]+]]:_(s64) = G_FNEG [[INT7]] ; GFX9: [[INT9:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s64) ; GFX9: [[FMA5:%[0-9]+]]:_(s64) = G_FMA [[FNEG1]], [[INT9]], [[C]] ; GFX9: [[FMA6:%[0-9]+]]:_(s64) = G_FMA [[INT9]], [[FMA5]], [[INT9]] ; GFX9: [[FMA7:%[0-9]+]]:_(s64) = G_FMA [[FNEG1]], [[FMA6]], [[C]] - ; GFX9: [[INT10:%[0-9]+]]:_(s64), [[INT11:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), [[UV1]](s64) + ; GFX9: [[INT10:%[0-9]+]]:_(s64), [[INT11:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), 0 ; GFX9: [[FMA8:%[0-9]+]]:_(s64) = G_FMA [[FMA6]], [[FMA7]], [[FMA6]] ; GFX9: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT10]](s64) ; GFX9: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA8]](s64) @@ -1168,13 +1168,13 @@ body: | ; GFX10: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>) ; GFX10: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>) ; GFX10: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV2]](s64), [[UV2]](s64), [[UV]](s64) + ; GFX10: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), 1 ; GFX10: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; GFX10: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; GFX10: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; GFX10: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; GFX10: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; GFX10: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), [[UV]](s64) + ; GFX10: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV]](s64), [[UV2]](s64), 0 ; GFX10: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; GFX10: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; GFX10: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -1188,13 +1188,13 @@ body: | ; GFX10: [[FMA4:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[MV]], [[INT3]] ; GFX10: [[INT5:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fmas), [[FMA4]](s64), [[FMA3]](s64), [[MV]](s64), [[INT4]](s1) ; GFX10: [[INT6:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.fixup), def %7(s64), [[INT5]](s64), [[UV2]](s64), [[UV]](s64) - ; GFX10: [[INT7:%[0-9]+]]:_(s64), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV3]](s64), [[UV3]](s64), [[UV1]](s64) + ; GFX10: [[INT7:%[0-9]+]]:_(s64), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), 1 ; GFX10: [[FNEG1:%[0-9]+]]:_(s64) = G_FNEG [[INT7]] ; GFX10: [[INT9:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s64) ; GFX10: [[FMA5:%[0-9]+]]:_(s64) = G_FMA [[FNEG1]], [[INT9]], [[C]] ; GFX10: [[FMA6:%[0-9]+]]:_(s64) = G_FMA [[INT9]], [[FMA5]], [[INT9]] ; GFX10: [[FMA7:%[0-9]+]]:_(s64) = G_FMA [[FNEG1]], [[FMA6]], [[C]] - ; GFX10: [[INT10:%[0-9]+]]:_(s64), [[INT11:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), [[UV1]](s64) + ; GFX10: [[INT10:%[0-9]+]]:_(s64), [[INT11:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[UV1]](s64), [[UV3]](s64), 0 ; GFX10: [[FMA8:%[0-9]+]]:_(s64) = G_FMA [[FMA6]], [[FMA7]], [[FMA6]] ; GFX10: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT10]](s64) ; GFX10: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA8]](s64) @@ -1237,8 +1237,8 @@ body: | ; SI: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16) ; SI: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16) ; SI: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -1252,8 +1252,8 @@ body: | ; SI: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[INT6]](s32) ; SI: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16) ; SI: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16) - ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT3]](s32), [[FPEXT3]](s32), [[FPEXT2]](s32) - ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT2]](s32), [[FPEXT3]](s32), [[FPEXT2]](s32) + ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT3]](s32), [[FPEXT2]](s32), 1 + ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT2]](s32), [[FPEXT3]](s32), 0 ; SI: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; SI: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; SI: [[FMA5:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[INT11]], [[C1]] @@ -1404,8 +1404,8 @@ body: | ; SI: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16) ; SI: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16) ; SI: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -1419,8 +1419,8 @@ body: | ; SI: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[INT6]](s32) ; SI: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16) ; SI: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC4]](s16) - ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT3]](s32), [[FPEXT3]](s32), [[FPEXT2]](s32) - ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT2]](s32), [[FPEXT3]](s32), [[FPEXT2]](s32) + ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT3]](s32), [[FPEXT2]](s32), 1 + ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT2]](s32), [[FPEXT3]](s32), 0 ; SI: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; SI: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; SI: [[FMA5:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[INT11]], [[C1]] @@ -1434,8 +1434,8 @@ body: | ; SI: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[INT13]](s32) ; SI: [[FPEXT4:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16) ; SI: [[FPEXT5:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC5]](s16) - ; SI: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT5]](s32), [[FPEXT5]](s32), [[FPEXT4]](s32) - ; SI: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT4]](s32), [[FPEXT5]](s32), [[FPEXT4]](s32) + ; SI: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT5]](s32), [[FPEXT4]](s32), 1 + ; SI: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT4]](s32), [[FPEXT5]](s32), 0 ; SI: [[INT18:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT14]](s32) ; SI: [[FNEG2:%[0-9]+]]:_(s32) = G_FNEG [[INT14]] ; SI: [[FMA10:%[0-9]+]]:_(s32) = G_FMA [[FNEG2]], [[INT18]], [[C1]] @@ -1678,8 +1678,8 @@ body: | ; SI: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16) ; SI: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC4]](s16) ; SI: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -1693,8 +1693,8 @@ body: | ; SI: [[FPTRUNC:%[0-9]+]]:_(s16) = G_FPTRUNC [[INT6]](s32) ; SI: [[FPEXT2:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC1]](s16) ; SI: [[FPEXT3:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC5]](s16) - ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT3]](s32), [[FPEXT3]](s32), [[FPEXT2]](s32) - ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT2]](s32), [[FPEXT3]](s32), [[FPEXT2]](s32) + ; SI: [[INT7:%[0-9]+]]:_(s32), [[INT8:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT3]](s32), [[FPEXT2]](s32), 1 + ; SI: [[INT9:%[0-9]+]]:_(s32), [[INT10:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT2]](s32), [[FPEXT3]](s32), 0 ; SI: [[INT11:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT7]](s32) ; SI: [[FNEG1:%[0-9]+]]:_(s32) = G_FNEG [[INT7]] ; SI: [[FMA5:%[0-9]+]]:_(s32) = G_FMA [[FNEG1]], [[INT11]], [[C1]] @@ -1708,8 +1708,8 @@ body: | ; SI: [[FPTRUNC1:%[0-9]+]]:_(s16) = G_FPTRUNC [[INT13]](s32) ; SI: [[FPEXT4:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC2]](s16) ; SI: [[FPEXT5:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC6]](s16) - ; SI: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT5]](s32), [[FPEXT5]](s32), [[FPEXT4]](s32) - ; SI: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT4]](s32), [[FPEXT5]](s32), [[FPEXT4]](s32) + ; SI: [[INT14:%[0-9]+]]:_(s32), [[INT15:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT5]](s32), [[FPEXT4]](s32), 1 + ; SI: [[INT16:%[0-9]+]]:_(s32), [[INT17:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT4]](s32), [[FPEXT5]](s32), 0 ; SI: [[INT18:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT14]](s32) ; SI: [[FNEG2:%[0-9]+]]:_(s32) = G_FNEG [[INT14]] ; SI: [[FMA10:%[0-9]+]]:_(s32) = G_FMA [[FNEG2]], [[INT18]], [[C1]] @@ -1723,8 +1723,8 @@ body: | ; SI: [[FPTRUNC2:%[0-9]+]]:_(s16) = G_FPTRUNC [[INT20]](s32) ; SI: [[FPEXT6:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC3]](s16) ; SI: [[FPEXT7:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC7]](s16) - ; SI: [[INT21:%[0-9]+]]:_(s32), [[INT22:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT7]](s32), [[FPEXT7]](s32), [[FPEXT6]](s32) - ; SI: [[INT23:%[0-9]+]]:_(s32), [[INT24:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT6]](s32), [[FPEXT7]](s32), [[FPEXT6]](s32) + ; SI: [[INT21:%[0-9]+]]:_(s32), [[INT22:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT7]](s32), [[FPEXT6]](s32), 1 + ; SI: [[INT23:%[0-9]+]]:_(s32), [[INT24:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT6]](s32), [[FPEXT7]](s32), 0 ; SI: [[INT25:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT21]](s32) ; SI: [[FNEG3:%[0-9]+]]:_(s32) = G_FNEG [[INT21]] ; SI: [[FMA15:%[0-9]+]]:_(s32) = G_FMA [[FNEG3]], [[INT25]], [[C1]] @@ -1943,8 +1943,8 @@ body: | ; SI: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[C]](s16) ; SI: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16) ; SI: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -2003,8 +2003,8 @@ body: | ; SI: [[FPEXT:%[0-9]+]]:_(s32) = G_FPEXT [[C]](s16) ; SI: [[FPEXT1:%[0-9]+]]:_(s32) = G_FPEXT [[TRUNC]](s16) ; SI: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), [[FPEXT]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT1]](s32), [[FPEXT]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[FPEXT]](s32), [[FPEXT1]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -2063,8 +2063,8 @@ body: | ; SI-LABEL: name: test_fdiv_s32_constant_one_rcp ; SI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 ; SI: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY]](s32), [[C]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), [[C]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[C]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -2079,8 +2079,8 @@ body: | ; VI-LABEL: name: test_fdiv_s32_constant_one_rcp ; VI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 ; VI: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 - ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY]](s32), [[C]](s32) - ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), [[C]](s32) + ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[C]](s32), 1 + ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), 0 ; VI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; VI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; VI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -2095,8 +2095,8 @@ body: | ; GFX9-LABEL: name: test_fdiv_s32_constant_one_rcp ; GFX9: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 - ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY]](s32), [[C]](s32) - ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), [[C]](s32) + ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[C]](s32), 1 + ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), 0 ; GFX9: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX9: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX9: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -2115,8 +2115,8 @@ body: | ; GFX10-LABEL: name: test_fdiv_s32_constant_one_rcp ; GFX10: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 - ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY]](s32), [[C]](s32) - ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), [[C]](s32) + ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[C]](s32), 1 + ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), 0 ; GFX10: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX10: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX10: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C]] @@ -2144,8 +2144,8 @@ body: | ; SI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float -1.000000e+00 ; SI: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; SI: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY]](s32), [[C]](s32) - ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), [[C]](s32) + ; SI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[C]](s32), 1 + ; SI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), 0 ; SI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; SI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; SI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -2161,8 +2161,8 @@ body: | ; VI: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float -1.000000e+00 ; VI: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; VI: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY]](s32), [[C]](s32) - ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), [[C]](s32) + ; VI: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[C]](s32), 1 + ; VI: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), 0 ; VI: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; VI: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; VI: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -2178,8 +2178,8 @@ body: | ; GFX9: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float -1.000000e+00 ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; GFX9: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY]](s32), [[C]](s32) - ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), [[C]](s32) + ; GFX9: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[C]](s32), 1 + ; GFX9: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), 0 ; GFX9: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX9: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX9: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -2200,8 +2200,8 @@ body: | ; GFX10: [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float -1.000000e+00 ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; GFX10: [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[COPY]](s32), [[C]](s32) - ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), [[C]](s32) + ; GFX10: [[INT:%[0-9]+]]:_(s32), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s32), [[C]](s32), 1 + ; GFX10: [[INT2:%[0-9]+]]:_(s32), [[INT3:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s32), [[COPY]](s32), 0 ; GFX10: [[INT4:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s32) ; GFX10: [[FNEG:%[0-9]+]]:_(s32) = G_FNEG [[INT]] ; GFX10: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[FNEG]], [[INT4]], [[C1]] @@ -2233,13 +2233,13 @@ body: | ; SI-LABEL: name: test_fdiv_s64_constant_one_rcp ; SI: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 ; SI: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 - ; SI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY]](s64), [[C]](s64) + ; SI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 1 ; SI: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; SI: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; SI: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; SI: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; SI: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; SI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), [[C]](s64) + ; SI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 0 ; SI: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; SI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -2264,13 +2264,13 @@ body: | ; VI-LABEL: name: test_fdiv_s64_constant_one_rcp ; VI: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 ; VI: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 - ; VI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY]](s64), [[C]](s64) + ; VI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 1 ; VI: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; VI: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; VI: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; VI: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; VI: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; VI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), [[C]](s64) + ; VI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 0 ; VI: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; VI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -2288,13 +2288,13 @@ body: | ; GFX9-LABEL: name: test_fdiv_s64_constant_one_rcp ; GFX9: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 ; GFX9: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 - ; GFX9: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY]](s64), [[C]](s64) + ; GFX9: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 1 ; GFX9: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; GFX9: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; GFX9: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; GFX9: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; GFX9: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; GFX9: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), [[C]](s64) + ; GFX9: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 0 ; GFX9: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; GFX9: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -2316,13 +2316,13 @@ body: | ; GFX10-LABEL: name: test_fdiv_s64_constant_one_rcp ; GFX10: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 ; GFX10: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 - ; GFX10: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY]](s64), [[C]](s64) + ; GFX10: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 1 ; GFX10: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; GFX10: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; GFX10: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C]] ; GFX10: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; GFX10: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C]] - ; GFX10: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), [[C]](s64) + ; GFX10: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 0 ; GFX10: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; GFX10: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -2358,13 +2358,13 @@ body: | ; SI: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double -1.000000e+00 ; SI: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 ; SI: [[C1:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; SI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY]](s64), [[C]](s64) + ; SI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 1 ; SI: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; SI: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; SI: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C1]] ; SI: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; SI: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C1]] - ; SI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), [[C]](s64) + ; SI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 0 ; SI: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; SI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -2390,13 +2390,13 @@ body: | ; VI: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double -1.000000e+00 ; VI: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 ; VI: [[C1:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; VI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY]](s64), [[C]](s64) + ; VI: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 1 ; VI: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; VI: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; VI: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C1]] ; VI: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; VI: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C1]] - ; VI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), [[C]](s64) + ; VI: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 0 ; VI: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; VI: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -2415,13 +2415,13 @@ body: | ; GFX9: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double -1.000000e+00 ; GFX9: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 ; GFX9: [[C1:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; GFX9: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY]](s64), [[C]](s64) + ; GFX9: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 1 ; GFX9: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; GFX9: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; GFX9: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C1]] ; GFX9: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; GFX9: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C1]] - ; GFX9: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), [[C]](s64) + ; GFX9: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 0 ; GFX9: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; GFX9: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) @@ -2445,13 +2445,13 @@ body: | ; GFX10: [[C:%[0-9]+]]:_(s64) = G_FCONSTANT double -1.000000e+00 ; GFX10: [[COPY:%[0-9]+]]:_(s64) = COPY $vgpr0_vgpr1 ; GFX10: [[C1:%[0-9]+]]:_(s64) = G_FCONSTANT double 1.000000e+00 - ; GFX10: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[COPY]](s64), [[COPY]](s64), [[C]](s64) + ; GFX10: [[INT:%[0-9]+]]:_(s64), [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 1 ; GFX10: [[FNEG:%[0-9]+]]:_(s64) = G_FNEG [[INT]] ; GFX10: [[INT2:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.rcp), [[INT]](s64) ; GFX10: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[INT2]], [[C1]] ; GFX10: [[FMA1:%[0-9]+]]:_(s64) = G_FMA [[INT2]], [[FMA]], [[INT2]] ; GFX10: [[FMA2:%[0-9]+]]:_(s64) = G_FMA [[FNEG]], [[FMA1]], [[C1]] - ; GFX10: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), [[C]](s64) + ; GFX10: [[INT3:%[0-9]+]]:_(s64), [[INT4:%[0-9]+]]:_(s1) = G_INTRINSIC intrinsic(@llvm.amdgcn.div.scale), [[C]](s64), [[COPY]](s64), 0 ; GFX10: [[FMA3:%[0-9]+]]:_(s64) = G_FMA [[FMA1]], [[FMA2]], [[FMA1]] ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[INT3]](s64) ; GFX10: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA3]](s64) |
