diff options
Diffstat (limited to 'lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp')
| -rw-r--r-- | lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp | 1689 |
1 files changed, 1580 insertions, 109 deletions
diff --git a/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp b/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp index 0dfbf10..cd0422d 100644 --- a/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp +++ b/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp @@ -24,19 +24,10 @@ using namespace llvm; #define DEBUG_TYPE "nvptx-isel" -unsigned FMAContractLevel = 0; - -static cl::opt<unsigned, true> -FMAContractLevelOpt("nvptx-fma-level", cl::ZeroOrMore, cl::Hidden, - cl::desc("NVPTX Specific: FMA contraction (0: don't do it" - " 1: do it 2: do it aggressively"), - cl::location(FMAContractLevel), - cl::init(2)); - static cl::opt<int> UsePrecDivF32( "nvptx-prec-divf32", cl::ZeroOrMore, cl::Hidden, cl::desc("NVPTX Specifies: 0 use div.approx, 1 use div.full, 2 use" - " IEEE Compliant F32 div.rnd if avaiable."), + " IEEE Compliant F32 div.rnd if available."), cl::init(2)); static cl::opt<bool> @@ -61,16 +52,6 @@ NVPTXDAGToDAGISel::NVPTXDAGToDAGISel(NVPTXTargetMachine &tm, CodeGenOpt::Level OptLevel) : SelectionDAGISel(tm, OptLevel), Subtarget(tm.getSubtarget<NVPTXSubtarget>()) { - - doFMAF32 = (OptLevel > 0) && Subtarget.hasFMAF32() && (FMAContractLevel >= 1); - doFMAF64 = (OptLevel > 0) && Subtarget.hasFMAF64() && (FMAContractLevel >= 1); - doFMAF32AGG = - (OptLevel > 0) && Subtarget.hasFMAF32() && (FMAContractLevel == 2); - doFMAF64AGG = - (OptLevel > 0) && Subtarget.hasFMAF64() && (FMAContractLevel == 2); - - allowFMA = (FMAContractLevel >= 1); - doMulWide = (OptLevel > 0); } @@ -116,6 +97,11 @@ bool NVPTXDAGToDAGISel::useF32FTZ() const { } } +bool NVPTXDAGToDAGISel::allowFMA() const { + const NVPTXTargetLowering *TL = Subtarget.getTargetLowering(); + return TL->allowFMA(*MF, OptLevel); +} + /// Select - Select instructions not customized! Used for /// expanded, promoted and normal instructions. SDNode *NVPTXDAGToDAGISel::Select(SDNode *N) { @@ -170,93 +156,341 @@ SDNode *NVPTXDAGToDAGISel::Select(SDNode *N) { case ISD::INTRINSIC_W_CHAIN: ResNode = SelectIntrinsicChain(N); break; - case NVPTXISD::Tex1DFloatI32: + case NVPTXISD::Tex1DFloatS32: case NVPTXISD::Tex1DFloatFloat: case NVPTXISD::Tex1DFloatFloatLevel: case NVPTXISD::Tex1DFloatFloatGrad: - case NVPTXISD::Tex1DI32I32: - case NVPTXISD::Tex1DI32Float: - case NVPTXISD::Tex1DI32FloatLevel: - case NVPTXISD::Tex1DI32FloatGrad: - case NVPTXISD::Tex1DArrayFloatI32: + case NVPTXISD::Tex1DS32S32: + case NVPTXISD::Tex1DS32Float: + case NVPTXISD::Tex1DS32FloatLevel: + case NVPTXISD::Tex1DS32FloatGrad: + case NVPTXISD::Tex1DU32S32: + case NVPTXISD::Tex1DU32Float: + case NVPTXISD::Tex1DU32FloatLevel: + case NVPTXISD::Tex1DU32FloatGrad: + case NVPTXISD::Tex1DArrayFloatS32: case NVPTXISD::Tex1DArrayFloatFloat: case NVPTXISD::Tex1DArrayFloatFloatLevel: case NVPTXISD::Tex1DArrayFloatFloatGrad: - case NVPTXISD::Tex1DArrayI32I32: - case NVPTXISD::Tex1DArrayI32Float: - case NVPTXISD::Tex1DArrayI32FloatLevel: - case NVPTXISD::Tex1DArrayI32FloatGrad: - case NVPTXISD::Tex2DFloatI32: + case NVPTXISD::Tex1DArrayS32S32: + case NVPTXISD::Tex1DArrayS32Float: + case NVPTXISD::Tex1DArrayS32FloatLevel: + case NVPTXISD::Tex1DArrayS32FloatGrad: + case NVPTXISD::Tex1DArrayU32S32: + case NVPTXISD::Tex1DArrayU32Float: + case NVPTXISD::Tex1DArrayU32FloatLevel: + case NVPTXISD::Tex1DArrayU32FloatGrad: + case NVPTXISD::Tex2DFloatS32: case NVPTXISD::Tex2DFloatFloat: case NVPTXISD::Tex2DFloatFloatLevel: case NVPTXISD::Tex2DFloatFloatGrad: - case NVPTXISD::Tex2DI32I32: - case NVPTXISD::Tex2DI32Float: - case NVPTXISD::Tex2DI32FloatLevel: - case NVPTXISD::Tex2DI32FloatGrad: - case NVPTXISD::Tex2DArrayFloatI32: + case NVPTXISD::Tex2DS32S32: + case NVPTXISD::Tex2DS32Float: + case NVPTXISD::Tex2DS32FloatLevel: + case NVPTXISD::Tex2DS32FloatGrad: + case NVPTXISD::Tex2DU32S32: + case NVPTXISD::Tex2DU32Float: + case NVPTXISD::Tex2DU32FloatLevel: + case NVPTXISD::Tex2DU32FloatGrad: + case NVPTXISD::Tex2DArrayFloatS32: case NVPTXISD::Tex2DArrayFloatFloat: case NVPTXISD::Tex2DArrayFloatFloatLevel: case NVPTXISD::Tex2DArrayFloatFloatGrad: - case NVPTXISD::Tex2DArrayI32I32: - case NVPTXISD::Tex2DArrayI32Float: - case NVPTXISD::Tex2DArrayI32FloatLevel: - case NVPTXISD::Tex2DArrayI32FloatGrad: - case NVPTXISD::Tex3DFloatI32: + case NVPTXISD::Tex2DArrayS32S32: + case NVPTXISD::Tex2DArrayS32Float: + case NVPTXISD::Tex2DArrayS32FloatLevel: + case NVPTXISD::Tex2DArrayS32FloatGrad: + case NVPTXISD::Tex2DArrayU32S32: + case NVPTXISD::Tex2DArrayU32Float: + case NVPTXISD::Tex2DArrayU32FloatLevel: + case NVPTXISD::Tex2DArrayU32FloatGrad: + case NVPTXISD::Tex3DFloatS32: case NVPTXISD::Tex3DFloatFloat: case NVPTXISD::Tex3DFloatFloatLevel: case NVPTXISD::Tex3DFloatFloatGrad: - case NVPTXISD::Tex3DI32I32: - case NVPTXISD::Tex3DI32Float: - case NVPTXISD::Tex3DI32FloatLevel: - case NVPTXISD::Tex3DI32FloatGrad: + case NVPTXISD::Tex3DS32S32: + case NVPTXISD::Tex3DS32Float: + case NVPTXISD::Tex3DS32FloatLevel: + case NVPTXISD::Tex3DS32FloatGrad: + case NVPTXISD::Tex3DU32S32: + case NVPTXISD::Tex3DU32Float: + case NVPTXISD::Tex3DU32FloatLevel: + case NVPTXISD::Tex3DU32FloatGrad: + case NVPTXISD::TexCubeFloatFloat: + case NVPTXISD::TexCubeFloatFloatLevel: + case NVPTXISD::TexCubeS32Float: + case NVPTXISD::TexCubeS32FloatLevel: + case NVPTXISD::TexCubeU32Float: + case NVPTXISD::TexCubeU32FloatLevel: + case NVPTXISD::TexCubeArrayFloatFloat: + case NVPTXISD::TexCubeArrayFloatFloatLevel: + case NVPTXISD::TexCubeArrayS32Float: + case NVPTXISD::TexCubeArrayS32FloatLevel: + case NVPTXISD::TexCubeArrayU32Float: + case NVPTXISD::TexCubeArrayU32FloatLevel: + case NVPTXISD::Tld4R2DFloatFloat: + case NVPTXISD::Tld4G2DFloatFloat: + case NVPTXISD::Tld4B2DFloatFloat: + case NVPTXISD::Tld4A2DFloatFloat: + case NVPTXISD::Tld4R2DS64Float: + case NVPTXISD::Tld4G2DS64Float: + case NVPTXISD::Tld4B2DS64Float: + case NVPTXISD::Tld4A2DS64Float: + case NVPTXISD::Tld4R2DU64Float: + case NVPTXISD::Tld4G2DU64Float: + case NVPTXISD::Tld4B2DU64Float: + case NVPTXISD::Tld4A2DU64Float: + case NVPTXISD::TexUnified1DFloatS32: + case NVPTXISD::TexUnified1DFloatFloat: + case NVPTXISD::TexUnified1DFloatFloatLevel: + case NVPTXISD::TexUnified1DFloatFloatGrad: + case NVPTXISD::TexUnified1DS32S32: + case NVPTXISD::TexUnified1DS32Float: + case NVPTXISD::TexUnified1DS32FloatLevel: + case NVPTXISD::TexUnified1DS32FloatGrad: + case NVPTXISD::TexUnified1DU32S32: + case NVPTXISD::TexUnified1DU32Float: + case NVPTXISD::TexUnified1DU32FloatLevel: + case NVPTXISD::TexUnified1DU32FloatGrad: + case NVPTXISD::TexUnified1DArrayFloatS32: + case NVPTXISD::TexUnified1DArrayFloatFloat: + case NVPTXISD::TexUnified1DArrayFloatFloatLevel: + case NVPTXISD::TexUnified1DArrayFloatFloatGrad: + case NVPTXISD::TexUnified1DArrayS32S32: + case NVPTXISD::TexUnified1DArrayS32Float: + case NVPTXISD::TexUnified1DArrayS32FloatLevel: + case NVPTXISD::TexUnified1DArrayS32FloatGrad: + case NVPTXISD::TexUnified1DArrayU32S32: + case NVPTXISD::TexUnified1DArrayU32Float: + case NVPTXISD::TexUnified1DArrayU32FloatLevel: + case NVPTXISD::TexUnified1DArrayU32FloatGrad: + case NVPTXISD::TexUnified2DFloatS32: + case NVPTXISD::TexUnified2DFloatFloat: + case NVPTXISD::TexUnified2DFloatFloatLevel: + case NVPTXISD::TexUnified2DFloatFloatGrad: + case NVPTXISD::TexUnified2DS32S32: + case NVPTXISD::TexUnified2DS32Float: + case NVPTXISD::TexUnified2DS32FloatLevel: + case NVPTXISD::TexUnified2DS32FloatGrad: + case NVPTXISD::TexUnified2DU32S32: + case NVPTXISD::TexUnified2DU32Float: + case NVPTXISD::TexUnified2DU32FloatLevel: + case NVPTXISD::TexUnified2DU32FloatGrad: + case NVPTXISD::TexUnified2DArrayFloatS32: + case NVPTXISD::TexUnified2DArrayFloatFloat: + case NVPTXISD::TexUnified2DArrayFloatFloatLevel: + case NVPTXISD::TexUnified2DArrayFloatFloatGrad: + case NVPTXISD::TexUnified2DArrayS32S32: + case NVPTXISD::TexUnified2DArrayS32Float: + case NVPTXISD::TexUnified2DArrayS32FloatLevel: + case NVPTXISD::TexUnified2DArrayS32FloatGrad: + case NVPTXISD::TexUnified2DArrayU32S32: + case NVPTXISD::TexUnified2DArrayU32Float: + case NVPTXISD::TexUnified2DArrayU32FloatLevel: + case NVPTXISD::TexUnified2DArrayU32FloatGrad: + case NVPTXISD::TexUnified3DFloatS32: + case NVPTXISD::TexUnified3DFloatFloat: + case NVPTXISD::TexUnified3DFloatFloatLevel: + case NVPTXISD::TexUnified3DFloatFloatGrad: + case NVPTXISD::TexUnified3DS32S32: + case NVPTXISD::TexUnified3DS32Float: + case NVPTXISD::TexUnified3DS32FloatLevel: + case NVPTXISD::TexUnified3DS32FloatGrad: + case NVPTXISD::TexUnified3DU32S32: + case NVPTXISD::TexUnified3DU32Float: + case NVPTXISD::TexUnified3DU32FloatLevel: + case NVPTXISD::TexUnified3DU32FloatGrad: + case NVPTXISD::TexUnifiedCubeFloatFloat: + case NVPTXISD::TexUnifiedCubeFloatFloatLevel: + case NVPTXISD::TexUnifiedCubeS32Float: + case NVPTXISD::TexUnifiedCubeS32FloatLevel: + case NVPTXISD::TexUnifiedCubeU32Float: + case NVPTXISD::TexUnifiedCubeU32FloatLevel: + case NVPTXISD::TexUnifiedCubeArrayFloatFloat: + case NVPTXISD::TexUnifiedCubeArrayFloatFloatLevel: + case NVPTXISD::TexUnifiedCubeArrayS32Float: + case NVPTXISD::TexUnifiedCubeArrayS32FloatLevel: + case NVPTXISD::TexUnifiedCubeArrayU32Float: + case NVPTXISD::TexUnifiedCubeArrayU32FloatLevel: + case NVPTXISD::Tld4UnifiedR2DFloatFloat: + case NVPTXISD::Tld4UnifiedG2DFloatFloat: + case NVPTXISD::Tld4UnifiedB2DFloatFloat: + case NVPTXISD::Tld4UnifiedA2DFloatFloat: + case NVPTXISD::Tld4UnifiedR2DS64Float: + case NVPTXISD::Tld4UnifiedG2DS64Float: + case NVPTXISD::Tld4UnifiedB2DS64Float: + case NVPTXISD::Tld4UnifiedA2DS64Float: + case NVPTXISD::Tld4UnifiedR2DU64Float: + case NVPTXISD::Tld4UnifiedG2DU64Float: + case NVPTXISD::Tld4UnifiedB2DU64Float: + case NVPTXISD::Tld4UnifiedA2DU64Float: ResNode = SelectTextureIntrinsic(N); break; + case NVPTXISD::Suld1DI8Clamp: + case NVPTXISD::Suld1DI16Clamp: + case NVPTXISD::Suld1DI32Clamp: + case NVPTXISD::Suld1DI64Clamp: + case NVPTXISD::Suld1DV2I8Clamp: + case NVPTXISD::Suld1DV2I16Clamp: + case NVPTXISD::Suld1DV2I32Clamp: + case NVPTXISD::Suld1DV2I64Clamp: + case NVPTXISD::Suld1DV4I8Clamp: + case NVPTXISD::Suld1DV4I16Clamp: + case NVPTXISD::Suld1DV4I32Clamp: + case NVPTXISD::Suld1DArrayI8Clamp: + case NVPTXISD::Suld1DArrayI16Clamp: + case NVPTXISD::Suld1DArrayI32Clamp: + case NVPTXISD::Suld1DArrayI64Clamp: + case NVPTXISD::Suld1DArrayV2I8Clamp: + case NVPTXISD::Suld1DArrayV2I16Clamp: + case NVPTXISD::Suld1DArrayV2I32Clamp: + case NVPTXISD::Suld1DArrayV2I64Clamp: + case NVPTXISD::Suld1DArrayV4I8Clamp: + case NVPTXISD::Suld1DArrayV4I16Clamp: + case NVPTXISD::Suld1DArrayV4I32Clamp: + case NVPTXISD::Suld2DI8Clamp: + case NVPTXISD::Suld2DI16Clamp: + case NVPTXISD::Suld2DI32Clamp: + case NVPTXISD::Suld2DI64Clamp: + case NVPTXISD::Suld2DV2I8Clamp: + case NVPTXISD::Suld2DV2I16Clamp: + case NVPTXISD::Suld2DV2I32Clamp: + case NVPTXISD::Suld2DV2I64Clamp: + case NVPTXISD::Suld2DV4I8Clamp: + case NVPTXISD::Suld2DV4I16Clamp: + case NVPTXISD::Suld2DV4I32Clamp: + case NVPTXISD::Suld2DArrayI8Clamp: + case NVPTXISD::Suld2DArrayI16Clamp: + case NVPTXISD::Suld2DArrayI32Clamp: + case NVPTXISD::Suld2DArrayI64Clamp: + case NVPTXISD::Suld2DArrayV2I8Clamp: + case NVPTXISD::Suld2DArrayV2I16Clamp: + case NVPTXISD::Suld2DArrayV2I32Clamp: + case NVPTXISD::Suld2DArrayV2I64Clamp: + case NVPTXISD::Suld2DArrayV4I8Clamp: + case NVPTXISD::Suld2DArrayV4I16Clamp: + case NVPTXISD::Suld2DArrayV4I32Clamp: + case NVPTXISD::Suld3DI8Clamp: + case NVPTXISD::Suld3DI16Clamp: + case NVPTXISD::Suld3DI32Clamp: + case NVPTXISD::Suld3DI64Clamp: + case NVPTXISD::Suld3DV2I8Clamp: + case NVPTXISD::Suld3DV2I16Clamp: + case NVPTXISD::Suld3DV2I32Clamp: + case NVPTXISD::Suld3DV2I64Clamp: + case NVPTXISD::Suld3DV4I8Clamp: + case NVPTXISD::Suld3DV4I16Clamp: + case NVPTXISD::Suld3DV4I32Clamp: case NVPTXISD::Suld1DI8Trap: case NVPTXISD::Suld1DI16Trap: case NVPTXISD::Suld1DI32Trap: + case NVPTXISD::Suld1DI64Trap: case NVPTXISD::Suld1DV2I8Trap: case NVPTXISD::Suld1DV2I16Trap: case NVPTXISD::Suld1DV2I32Trap: + case NVPTXISD::Suld1DV2I64Trap: case NVPTXISD::Suld1DV4I8Trap: case NVPTXISD::Suld1DV4I16Trap: case NVPTXISD::Suld1DV4I32Trap: case NVPTXISD::Suld1DArrayI8Trap: case NVPTXISD::Suld1DArrayI16Trap: case NVPTXISD::Suld1DArrayI32Trap: + case NVPTXISD::Suld1DArrayI64Trap: case NVPTXISD::Suld1DArrayV2I8Trap: case NVPTXISD::Suld1DArrayV2I16Trap: case NVPTXISD::Suld1DArrayV2I32Trap: + case NVPTXISD::Suld1DArrayV2I64Trap: case NVPTXISD::Suld1DArrayV4I8Trap: case NVPTXISD::Suld1DArrayV4I16Trap: case NVPTXISD::Suld1DArrayV4I32Trap: case NVPTXISD::Suld2DI8Trap: case NVPTXISD::Suld2DI16Trap: case NVPTXISD::Suld2DI32Trap: + case NVPTXISD::Suld2DI64Trap: case NVPTXISD::Suld2DV2I8Trap: case NVPTXISD::Suld2DV2I16Trap: case NVPTXISD::Suld2DV2I32Trap: + case NVPTXISD::Suld2DV2I64Trap: case NVPTXISD::Suld2DV4I8Trap: case NVPTXISD::Suld2DV4I16Trap: case NVPTXISD::Suld2DV4I32Trap: case NVPTXISD::Suld2DArrayI8Trap: case NVPTXISD::Suld2DArrayI16Trap: case NVPTXISD::Suld2DArrayI32Trap: + case NVPTXISD::Suld2DArrayI64Trap: case NVPTXISD::Suld2DArrayV2I8Trap: case NVPTXISD::Suld2DArrayV2I16Trap: case NVPTXISD::Suld2DArrayV2I32Trap: + case NVPTXISD::Suld2DArrayV2I64Trap: case NVPTXISD::Suld2DArrayV4I8Trap: case NVPTXISD::Suld2DArrayV4I16Trap: case NVPTXISD::Suld2DArrayV4I32Trap: case NVPTXISD::Suld3DI8Trap: case NVPTXISD::Suld3DI16Trap: case NVPTXISD::Suld3DI32Trap: + case NVPTXISD::Suld3DI64Trap: case NVPTXISD::Suld3DV2I8Trap: case NVPTXISD::Suld3DV2I16Trap: case NVPTXISD::Suld3DV2I32Trap: + case NVPTXISD::Suld3DV2I64Trap: case NVPTXISD::Suld3DV4I8Trap: case NVPTXISD::Suld3DV4I16Trap: case NVPTXISD::Suld3DV4I32Trap: + case NVPTXISD::Suld1DI8Zero: + case NVPTXISD::Suld1DI16Zero: + case NVPTXISD::Suld1DI32Zero: + case NVPTXISD::Suld1DI64Zero: + case NVPTXISD::Suld1DV2I8Zero: + case NVPTXISD::Suld1DV2I16Zero: + case NVPTXISD::Suld1DV2I32Zero: + case NVPTXISD::Suld1DV2I64Zero: + case NVPTXISD::Suld1DV4I8Zero: + case NVPTXISD::Suld1DV4I16Zero: + case NVPTXISD::Suld1DV4I32Zero: + case NVPTXISD::Suld1DArrayI8Zero: + case NVPTXISD::Suld1DArrayI16Zero: + case NVPTXISD::Suld1DArrayI32Zero: + case NVPTXISD::Suld1DArrayI64Zero: + case NVPTXISD::Suld1DArrayV2I8Zero: + case NVPTXISD::Suld1DArrayV2I16Zero: + case NVPTXISD::Suld1DArrayV2I32Zero: + case NVPTXISD::Suld1DArrayV2I64Zero: + case NVPTXISD::Suld1DArrayV4I8Zero: + case NVPTXISD::Suld1DArrayV4I16Zero: + case NVPTXISD::Suld1DArrayV4I32Zero: + case NVPTXISD::Suld2DI8Zero: + case NVPTXISD::Suld2DI16Zero: + case NVPTXISD::Suld2DI32Zero: + case NVPTXISD::Suld2DI64Zero: + case NVPTXISD::Suld2DV2I8Zero: + case NVPTXISD::Suld2DV2I16Zero: + case NVPTXISD::Suld2DV2I32Zero: + case NVPTXISD::Suld2DV2I64Zero: + case NVPTXISD::Suld2DV4I8Zero: + case NVPTXISD::Suld2DV4I16Zero: + case NVPTXISD::Suld2DV4I32Zero: + case NVPTXISD::Suld2DArrayI8Zero: + case NVPTXISD::Suld2DArrayI16Zero: + case NVPTXISD::Suld2DArrayI32Zero: + case NVPTXISD::Suld2DArrayI64Zero: + case NVPTXISD::Suld2DArrayV2I8Zero: + case NVPTXISD::Suld2DArrayV2I16Zero: + case NVPTXISD::Suld2DArrayV2I32Zero: + case NVPTXISD::Suld2DArrayV2I64Zero: + case NVPTXISD::Suld2DArrayV4I8Zero: + case NVPTXISD::Suld2DArrayV4I16Zero: + case NVPTXISD::Suld2DArrayV4I32Zero: + case NVPTXISD::Suld3DI8Zero: + case NVPTXISD::Suld3DI16Zero: + case NVPTXISD::Suld3DI32Zero: + case NVPTXISD::Suld3DI64Zero: + case NVPTXISD::Suld3DV2I8Zero: + case NVPTXISD::Suld3DV2I16Zero: + case NVPTXISD::Suld3DV2I32Zero: + case NVPTXISD::Suld3DV2I64Zero: + case NVPTXISD::Suld3DV4I8Zero: + case NVPTXISD::Suld3DV4I16Zero: + case NVPTXISD::Suld3DV4I32Zero: ResNode = SelectSurfaceIntrinsic(N); break; case ISD::AND: @@ -2781,16 +3015,14 @@ SDNode *NVPTXDAGToDAGISel::SelectStoreParam(SDNode *N) { SDNode *NVPTXDAGToDAGISel::SelectTextureIntrinsic(SDNode *N) { SDValue Chain = N->getOperand(0); - SDValue TexRef = N->getOperand(1); - SDValue SampRef = N->getOperand(2); SDNode *Ret = nullptr; unsigned Opc = 0; SmallVector<SDValue, 8> Ops; switch (N->getOpcode()) { default: return nullptr; - case NVPTXISD::Tex1DFloatI32: - Opc = NVPTX::TEX_1D_F32_I32; + case NVPTXISD::Tex1DFloatS32: + Opc = NVPTX::TEX_1D_F32_S32; break; case NVPTXISD::Tex1DFloatFloat: Opc = NVPTX::TEX_1D_F32_F32; @@ -2801,20 +3033,32 @@ SDNode *NVPTXDAGToDAGISel::SelectTextureIntrinsic(SDNode *N) { case NVPTXISD::Tex1DFloatFloatGrad: Opc = NVPTX::TEX_1D_F32_F32_GRAD; break; - case NVPTXISD::Tex1DI32I32: - Opc = NVPTX::TEX_1D_I32_I32; + case NVPTXISD::Tex1DS32S32: + Opc = NVPTX::TEX_1D_S32_S32; + break; + case NVPTXISD::Tex1DS32Float: + Opc = NVPTX::TEX_1D_S32_F32; + break; + case NVPTXISD::Tex1DS32FloatLevel: + Opc = NVPTX::TEX_1D_S32_F32_LEVEL; break; - case NVPTXISD::Tex1DI32Float: - Opc = NVPTX::TEX_1D_I32_F32; + case NVPTXISD::Tex1DS32FloatGrad: + Opc = NVPTX::TEX_1D_S32_F32_GRAD; break; - case NVPTXISD::Tex1DI32FloatLevel: - Opc = NVPTX::TEX_1D_I32_F32_LEVEL; + case NVPTXISD::Tex1DU32S32: + Opc = NVPTX::TEX_1D_U32_S32; break; - case NVPTXISD::Tex1DI32FloatGrad: - Opc = NVPTX::TEX_1D_I32_F32_GRAD; + case NVPTXISD::Tex1DU32Float: + Opc = NVPTX::TEX_1D_U32_F32; break; - case NVPTXISD::Tex1DArrayFloatI32: - Opc = NVPTX::TEX_1D_ARRAY_F32_I32; + case NVPTXISD::Tex1DU32FloatLevel: + Opc = NVPTX::TEX_1D_U32_F32_LEVEL; + break; + case NVPTXISD::Tex1DU32FloatGrad: + Opc = NVPTX::TEX_1D_U32_F32_GRAD; + break; + case NVPTXISD::Tex1DArrayFloatS32: + Opc = NVPTX::TEX_1D_ARRAY_F32_S32; break; case NVPTXISD::Tex1DArrayFloatFloat: Opc = NVPTX::TEX_1D_ARRAY_F32_F32; @@ -2825,20 +3069,32 @@ SDNode *NVPTXDAGToDAGISel::SelectTextureIntrinsic(SDNode *N) { case NVPTXISD::Tex1DArrayFloatFloatGrad: Opc = NVPTX::TEX_1D_ARRAY_F32_F32_GRAD; break; - case NVPTXISD::Tex1DArrayI32I32: - Opc = NVPTX::TEX_1D_ARRAY_I32_I32; + case NVPTXISD::Tex1DArrayS32S32: + Opc = NVPTX::TEX_1D_ARRAY_S32_S32; + break; + case NVPTXISD::Tex1DArrayS32Float: + Opc = NVPTX::TEX_1D_ARRAY_S32_F32; break; - case NVPTXISD::Tex1DArrayI32Float: - Opc = NVPTX::TEX_1D_ARRAY_I32_F32; + case NVPTXISD::Tex1DArrayS32FloatLevel: + Opc = NVPTX::TEX_1D_ARRAY_S32_F32_LEVEL; break; - case NVPTXISD::Tex1DArrayI32FloatLevel: - Opc = NVPTX::TEX_1D_ARRAY_I32_F32_LEVEL; + case NVPTXISD::Tex1DArrayS32FloatGrad: + Opc = NVPTX::TEX_1D_ARRAY_S32_F32_GRAD; break; - case NVPTXISD::Tex1DArrayI32FloatGrad: - Opc = NVPTX::TEX_1D_ARRAY_I32_F32_GRAD; + case NVPTXISD::Tex1DArrayU32S32: + Opc = NVPTX::TEX_1D_ARRAY_U32_S32; break; - case NVPTXISD::Tex2DFloatI32: - Opc = NVPTX::TEX_2D_F32_I32; + case NVPTXISD::Tex1DArrayU32Float: + Opc = NVPTX::TEX_1D_ARRAY_U32_F32; + break; + case NVPTXISD::Tex1DArrayU32FloatLevel: + Opc = NVPTX::TEX_1D_ARRAY_U32_F32_LEVEL; + break; + case NVPTXISD::Tex1DArrayU32FloatGrad: + Opc = NVPTX::TEX_1D_ARRAY_U32_F32_GRAD; + break; + case NVPTXISD::Tex2DFloatS32: + Opc = NVPTX::TEX_2D_F32_S32; break; case NVPTXISD::Tex2DFloatFloat: Opc = NVPTX::TEX_2D_F32_F32; @@ -2849,20 +3105,32 @@ SDNode *NVPTXDAGToDAGISel::SelectTextureIntrinsic(SDNode *N) { case NVPTXISD::Tex2DFloatFloatGrad: Opc = NVPTX::TEX_2D_F32_F32_GRAD; break; - case NVPTXISD::Tex2DI32I32: - Opc = NVPTX::TEX_2D_I32_I32; + case NVPTXISD::Tex2DS32S32: + Opc = NVPTX::TEX_2D_S32_S32; + break; + case NVPTXISD::Tex2DS32Float: + Opc = NVPTX::TEX_2D_S32_F32; + break; + case NVPTXISD::Tex2DS32FloatLevel: + Opc = NVPTX::TEX_2D_S32_F32_LEVEL; + break; + case NVPTXISD::Tex2DS32FloatGrad: + Opc = NVPTX::TEX_2D_S32_F32_GRAD; break; - case NVPTXISD::Tex2DI32Float: - Opc = NVPTX::TEX_2D_I32_F32; + case NVPTXISD::Tex2DU32S32: + Opc = NVPTX::TEX_2D_U32_S32; break; - case NVPTXISD::Tex2DI32FloatLevel: - Opc = NVPTX::TEX_2D_I32_F32_LEVEL; + case NVPTXISD::Tex2DU32Float: + Opc = NVPTX::TEX_2D_U32_F32; break; - case NVPTXISD::Tex2DI32FloatGrad: - Opc = NVPTX::TEX_2D_I32_F32_GRAD; + case NVPTXISD::Tex2DU32FloatLevel: + Opc = NVPTX::TEX_2D_U32_F32_LEVEL; break; - case NVPTXISD::Tex2DArrayFloatI32: - Opc = NVPTX::TEX_2D_ARRAY_F32_I32; + case NVPTXISD::Tex2DU32FloatGrad: + Opc = NVPTX::TEX_2D_U32_F32_GRAD; + break; + case NVPTXISD::Tex2DArrayFloatS32: + Opc = NVPTX::TEX_2D_ARRAY_F32_S32; break; case NVPTXISD::Tex2DArrayFloatFloat: Opc = NVPTX::TEX_2D_ARRAY_F32_F32; @@ -2873,20 +3141,32 @@ SDNode *NVPTXDAGToDAGISel::SelectTextureIntrinsic(SDNode *N) { case NVPTXISD::Tex2DArrayFloatFloatGrad: Opc = NVPTX::TEX_2D_ARRAY_F32_F32_GRAD; break; - case NVPTXISD::Tex2DArrayI32I32: - Opc = NVPTX::TEX_2D_ARRAY_I32_I32; + case NVPTXISD::Tex2DArrayS32S32: + Opc = NVPTX::TEX_2D_ARRAY_S32_S32; + break; + case NVPTXISD::Tex2DArrayS32Float: + Opc = NVPTX::TEX_2D_ARRAY_S32_F32; + break; + case NVPTXISD::Tex2DArrayS32FloatLevel: + Opc = NVPTX::TEX_2D_ARRAY_S32_F32_LEVEL; break; - case NVPTXISD::Tex2DArrayI32Float: - Opc = NVPTX::TEX_2D_ARRAY_I32_F32; + case NVPTXISD::Tex2DArrayS32FloatGrad: + Opc = NVPTX::TEX_2D_ARRAY_S32_F32_GRAD; break; - case NVPTXISD::Tex2DArrayI32FloatLevel: - Opc = NVPTX::TEX_2D_ARRAY_I32_F32_LEVEL; + case NVPTXISD::Tex2DArrayU32S32: + Opc = NVPTX::TEX_2D_ARRAY_U32_S32; break; - case NVPTXISD::Tex2DArrayI32FloatGrad: - Opc = NVPTX::TEX_2D_ARRAY_I32_F32_GRAD; + case NVPTXISD::Tex2DArrayU32Float: + Opc = NVPTX::TEX_2D_ARRAY_U32_F32; break; - case NVPTXISD::Tex3DFloatI32: - Opc = NVPTX::TEX_3D_F32_I32; + case NVPTXISD::Tex2DArrayU32FloatLevel: + Opc = NVPTX::TEX_2D_ARRAY_U32_F32_LEVEL; + break; + case NVPTXISD::Tex2DArrayU32FloatGrad: + Opc = NVPTX::TEX_2D_ARRAY_U32_F32_GRAD; + break; + case NVPTXISD::Tex3DFloatS32: + Opc = NVPTX::TEX_3D_F32_S32; break; case NVPTXISD::Tex3DFloatFloat: Opc = NVPTX::TEX_3D_F32_F32; @@ -2897,25 +3177,358 @@ SDNode *NVPTXDAGToDAGISel::SelectTextureIntrinsic(SDNode *N) { case NVPTXISD::Tex3DFloatFloatGrad: Opc = NVPTX::TEX_3D_F32_F32_GRAD; break; - case NVPTXISD::Tex3DI32I32: - Opc = NVPTX::TEX_3D_I32_I32; + case NVPTXISD::Tex3DS32S32: + Opc = NVPTX::TEX_3D_S32_S32; + break; + case NVPTXISD::Tex3DS32Float: + Opc = NVPTX::TEX_3D_S32_F32; + break; + case NVPTXISD::Tex3DS32FloatLevel: + Opc = NVPTX::TEX_3D_S32_F32_LEVEL; + break; + case NVPTXISD::Tex3DS32FloatGrad: + Opc = NVPTX::TEX_3D_S32_F32_GRAD; + break; + case NVPTXISD::Tex3DU32S32: + Opc = NVPTX::TEX_3D_U32_S32; + break; + case NVPTXISD::Tex3DU32Float: + Opc = NVPTX::TEX_3D_U32_F32; + break; + case NVPTXISD::Tex3DU32FloatLevel: + Opc = NVPTX::TEX_3D_U32_F32_LEVEL; + break; + case NVPTXISD::Tex3DU32FloatGrad: + Opc = NVPTX::TEX_3D_U32_F32_GRAD; + break; + case NVPTXISD::TexCubeFloatFloat: + Opc = NVPTX::TEX_CUBE_F32_F32; + break; + case NVPTXISD::TexCubeFloatFloatLevel: + Opc = NVPTX::TEX_CUBE_F32_F32_LEVEL; + break; + case NVPTXISD::TexCubeS32Float: + Opc = NVPTX::TEX_CUBE_S32_F32; + break; + case NVPTXISD::TexCubeS32FloatLevel: + Opc = NVPTX::TEX_CUBE_S32_F32_LEVEL; + break; + case NVPTXISD::TexCubeU32Float: + Opc = NVPTX::TEX_CUBE_U32_F32; + break; + case NVPTXISD::TexCubeU32FloatLevel: + Opc = NVPTX::TEX_CUBE_U32_F32_LEVEL; + break; + case NVPTXISD::TexCubeArrayFloatFloat: + Opc = NVPTX::TEX_CUBE_ARRAY_F32_F32; + break; + case NVPTXISD::TexCubeArrayFloatFloatLevel: + Opc = NVPTX::TEX_CUBE_ARRAY_F32_F32_LEVEL; + break; + case NVPTXISD::TexCubeArrayS32Float: + Opc = NVPTX::TEX_CUBE_ARRAY_S32_F32; + break; + case NVPTXISD::TexCubeArrayS32FloatLevel: + Opc = NVPTX::TEX_CUBE_ARRAY_S32_F32_LEVEL; + break; + case NVPTXISD::TexCubeArrayU32Float: + Opc = NVPTX::TEX_CUBE_ARRAY_U32_F32; + break; + case NVPTXISD::TexCubeArrayU32FloatLevel: + Opc = NVPTX::TEX_CUBE_ARRAY_U32_F32_LEVEL; + break; + case NVPTXISD::Tld4R2DFloatFloat: + Opc = NVPTX::TLD4_R_2D_F32_F32; + break; + case NVPTXISD::Tld4G2DFloatFloat: + Opc = NVPTX::TLD4_G_2D_F32_F32; + break; + case NVPTXISD::Tld4B2DFloatFloat: + Opc = NVPTX::TLD4_B_2D_F32_F32; + break; + case NVPTXISD::Tld4A2DFloatFloat: + Opc = NVPTX::TLD4_A_2D_F32_F32; + break; + case NVPTXISD::Tld4R2DS64Float: + Opc = NVPTX::TLD4_R_2D_S32_F32; + break; + case NVPTXISD::Tld4G2DS64Float: + Opc = NVPTX::TLD4_G_2D_S32_F32; + break; + case NVPTXISD::Tld4B2DS64Float: + Opc = NVPTX::TLD4_B_2D_S32_F32; + break; + case NVPTXISD::Tld4A2DS64Float: + Opc = NVPTX::TLD4_A_2D_S32_F32; + break; + case NVPTXISD::Tld4R2DU64Float: + Opc = NVPTX::TLD4_R_2D_U32_F32; + break; + case NVPTXISD::Tld4G2DU64Float: + Opc = NVPTX::TLD4_G_2D_U32_F32; + break; + case NVPTXISD::Tld4B2DU64Float: + Opc = NVPTX::TLD4_B_2D_U32_F32; + break; + case NVPTXISD::Tld4A2DU64Float: + Opc = NVPTX::TLD4_A_2D_U32_F32; + break; + case NVPTXISD::TexUnified1DFloatS32: + Opc = NVPTX::TEX_UNIFIED_1D_F32_S32; + break; + case NVPTXISD::TexUnified1DFloatFloat: + Opc = NVPTX::TEX_UNIFIED_1D_F32_F32; + break; + case NVPTXISD::TexUnified1DFloatFloatLevel: + Opc = NVPTX::TEX_UNIFIED_1D_F32_F32_LEVEL; + break; + case NVPTXISD::TexUnified1DFloatFloatGrad: + Opc = NVPTX::TEX_UNIFIED_1D_F32_F32_GRAD; + break; + case NVPTXISD::TexUnified1DS32S32: + Opc = NVPTX::TEX_UNIFIED_1D_S32_S32; + break; + case NVPTXISD::TexUnified1DS32Float: + Opc = NVPTX::TEX_UNIFIED_1D_S32_F32; + break; + case NVPTXISD::TexUnified1DS32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_1D_S32_F32_LEVEL; + break; + case NVPTXISD::TexUnified1DS32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_1D_S32_F32_GRAD; + break; + case NVPTXISD::TexUnified1DU32S32: + Opc = NVPTX::TEX_UNIFIED_1D_U32_S32; + break; + case NVPTXISD::TexUnified1DU32Float: + Opc = NVPTX::TEX_UNIFIED_1D_U32_F32; + break; + case NVPTXISD::TexUnified1DU32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_1D_U32_F32_LEVEL; + break; + case NVPTXISD::TexUnified1DU32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_1D_U32_F32_GRAD; + break; + case NVPTXISD::TexUnified1DArrayFloatS32: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_F32_S32; + break; + case NVPTXISD::TexUnified1DArrayFloatFloat: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_F32_F32; + break; + case NVPTXISD::TexUnified1DArrayFloatFloatLevel: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_F32_F32_LEVEL; + break; + case NVPTXISD::TexUnified1DArrayFloatFloatGrad: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_F32_F32_GRAD; + break; + case NVPTXISD::TexUnified1DArrayS32S32: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_S32_S32; + break; + case NVPTXISD::TexUnified1DArrayS32Float: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_S32_F32; + break; + case NVPTXISD::TexUnified1DArrayS32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_S32_F32_LEVEL; + break; + case NVPTXISD::TexUnified1DArrayS32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_S32_F32_GRAD; + break; + case NVPTXISD::TexUnified1DArrayU32S32: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_U32_S32; + break; + case NVPTXISD::TexUnified1DArrayU32Float: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_U32_F32; + break; + case NVPTXISD::TexUnified1DArrayU32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_U32_F32_LEVEL; + break; + case NVPTXISD::TexUnified1DArrayU32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_1D_ARRAY_U32_F32_GRAD; + break; + case NVPTXISD::TexUnified2DFloatS32: + Opc = NVPTX::TEX_UNIFIED_2D_F32_S32; + break; + case NVPTXISD::TexUnified2DFloatFloat: + Opc = NVPTX::TEX_UNIFIED_2D_F32_F32; + break; + case NVPTXISD::TexUnified2DFloatFloatLevel: + Opc = NVPTX::TEX_UNIFIED_2D_F32_F32_LEVEL; + break; + case NVPTXISD::TexUnified2DFloatFloatGrad: + Opc = NVPTX::TEX_UNIFIED_2D_F32_F32_GRAD; + break; + case NVPTXISD::TexUnified2DS32S32: + Opc = NVPTX::TEX_UNIFIED_2D_S32_S32; + break; + case NVPTXISD::TexUnified2DS32Float: + Opc = NVPTX::TEX_UNIFIED_2D_S32_F32; + break; + case NVPTXISD::TexUnified2DS32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_2D_S32_F32_LEVEL; + break; + case NVPTXISD::TexUnified2DS32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_2D_S32_F32_GRAD; + break; + case NVPTXISD::TexUnified2DU32S32: + Opc = NVPTX::TEX_UNIFIED_2D_U32_S32; + break; + case NVPTXISD::TexUnified2DU32Float: + Opc = NVPTX::TEX_UNIFIED_2D_U32_F32; + break; + case NVPTXISD::TexUnified2DU32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_2D_U32_F32_LEVEL; + break; + case NVPTXISD::TexUnified2DU32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_2D_U32_F32_GRAD; + break; + case NVPTXISD::TexUnified2DArrayFloatS32: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_F32_S32; + break; + case NVPTXISD::TexUnified2DArrayFloatFloat: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_F32_F32; + break; + case NVPTXISD::TexUnified2DArrayFloatFloatLevel: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_F32_F32_LEVEL; + break; + case NVPTXISD::TexUnified2DArrayFloatFloatGrad: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_F32_F32_GRAD; + break; + case NVPTXISD::TexUnified2DArrayS32S32: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_S32_S32; + break; + case NVPTXISD::TexUnified2DArrayS32Float: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_S32_F32; break; - case NVPTXISD::Tex3DI32Float: - Opc = NVPTX::TEX_3D_I32_F32; + case NVPTXISD::TexUnified2DArrayS32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_S32_F32_LEVEL; break; - case NVPTXISD::Tex3DI32FloatLevel: - Opc = NVPTX::TEX_3D_I32_F32_LEVEL; + case NVPTXISD::TexUnified2DArrayS32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_S32_F32_GRAD; break; - case NVPTXISD::Tex3DI32FloatGrad: - Opc = NVPTX::TEX_3D_I32_F32_GRAD; + case NVPTXISD::TexUnified2DArrayU32S32: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_U32_S32; + break; + case NVPTXISD::TexUnified2DArrayU32Float: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_U32_F32; + break; + case NVPTXISD::TexUnified2DArrayU32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_U32_F32_LEVEL; + break; + case NVPTXISD::TexUnified2DArrayU32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_U32_F32_GRAD; + break; + case NVPTXISD::TexUnified3DFloatS32: + Opc = NVPTX::TEX_UNIFIED_3D_F32_S32; + break; + case NVPTXISD::TexUnified3DFloatFloat: + Opc = NVPTX::TEX_UNIFIED_3D_F32_F32; + break; + case NVPTXISD::TexUnified3DFloatFloatLevel: + Opc = NVPTX::TEX_UNIFIED_3D_F32_F32_LEVEL; + break; + case NVPTXISD::TexUnified3DFloatFloatGrad: + Opc = NVPTX::TEX_UNIFIED_3D_F32_F32_GRAD; + break; + case NVPTXISD::TexUnified3DS32S32: + Opc = NVPTX::TEX_UNIFIED_3D_S32_S32; + break; + case NVPTXISD::TexUnified3DS32Float: + Opc = NVPTX::TEX_UNIFIED_3D_S32_F32; + break; + case NVPTXISD::TexUnified3DS32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_3D_S32_F32_LEVEL; + break; + case NVPTXISD::TexUnified3DS32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_3D_S32_F32_GRAD; + break; + case NVPTXISD::TexUnified3DU32S32: + Opc = NVPTX::TEX_UNIFIED_3D_U32_S32; + break; + case NVPTXISD::TexUnified3DU32Float: + Opc = NVPTX::TEX_UNIFIED_3D_U32_F32; + break; + case NVPTXISD::TexUnified3DU32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_3D_U32_F32_LEVEL; + break; + case NVPTXISD::TexUnified3DU32FloatGrad: + Opc = NVPTX::TEX_UNIFIED_3D_U32_F32_GRAD; + break; + case NVPTXISD::TexUnifiedCubeFloatFloat: + Opc = NVPTX::TEX_UNIFIED_CUBE_F32_F32; + break; + case NVPTXISD::TexUnifiedCubeFloatFloatLevel: + Opc = NVPTX::TEX_UNIFIED_CUBE_F32_F32_LEVEL; + break; + case NVPTXISD::TexUnifiedCubeS32Float: + Opc = NVPTX::TEX_UNIFIED_CUBE_S32_F32; + break; + case NVPTXISD::TexUnifiedCubeS32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_CUBE_S32_F32_LEVEL; + break; + case NVPTXISD::TexUnifiedCubeU32Float: + Opc = NVPTX::TEX_UNIFIED_CUBE_U32_F32; + break; + case NVPTXISD::TexUnifiedCubeU32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_CUBE_U32_F32_LEVEL; + break; + case NVPTXISD::TexUnifiedCubeArrayFloatFloat: + Opc = NVPTX::TEX_UNIFIED_CUBE_ARRAY_F32_F32; + break; + case NVPTXISD::TexUnifiedCubeArrayFloatFloatLevel: + Opc = NVPTX::TEX_UNIFIED_CUBE_ARRAY_F32_F32_LEVEL; + break; + case NVPTXISD::TexUnifiedCubeArrayS32Float: + Opc = NVPTX::TEX_UNIFIED_CUBE_ARRAY_S32_F32; + break; + case NVPTXISD::TexUnifiedCubeArrayS32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_CUBE_ARRAY_S32_F32_LEVEL; + break; + case NVPTXISD::TexUnifiedCubeArrayU32Float: + Opc = NVPTX::TEX_UNIFIED_CUBE_ARRAY_U32_F32; + break; + case NVPTXISD::TexUnifiedCubeArrayU32FloatLevel: + Opc = NVPTX::TEX_UNIFIED_CUBE_ARRAY_U32_F32_LEVEL; + break; + case NVPTXISD::Tld4UnifiedR2DFloatFloat: + Opc = NVPTX::TLD4_UNIFIED_R_2D_F32_F32; + break; + case NVPTXISD::Tld4UnifiedG2DFloatFloat: + Opc = NVPTX::TLD4_UNIFIED_G_2D_F32_F32; + break; + case NVPTXISD::Tld4UnifiedB2DFloatFloat: + Opc = NVPTX::TLD4_UNIFIED_B_2D_F32_F32; + break; + case NVPTXISD::Tld4UnifiedA2DFloatFloat: + Opc = NVPTX::TLD4_UNIFIED_A_2D_F32_F32; + break; + case NVPTXISD::Tld4UnifiedR2DS64Float: + Opc = NVPTX::TLD4_UNIFIED_R_2D_S32_F32; + break; + case NVPTXISD::Tld4UnifiedG2DS64Float: + Opc = NVPTX::TLD4_UNIFIED_G_2D_S32_F32; + break; + case NVPTXISD::Tld4UnifiedB2DS64Float: + Opc = NVPTX::TLD4_UNIFIED_B_2D_S32_F32; + break; + case NVPTXISD::Tld4UnifiedA2DS64Float: + Opc = NVPTX::TLD4_UNIFIED_A_2D_S32_F32; + break; + case NVPTXISD::Tld4UnifiedR2DU64Float: + Opc = NVPTX::TLD4_UNIFIED_R_2D_U32_F32; + break; + case NVPTXISD::Tld4UnifiedG2DU64Float: + Opc = NVPTX::TLD4_UNIFIED_G_2D_U32_F32; + break; + case NVPTXISD::Tld4UnifiedB2DU64Float: + Opc = NVPTX::TLD4_UNIFIED_B_2D_U32_F32; + break; + case NVPTXISD::Tld4UnifiedA2DU64Float: + Opc = NVPTX::TLD4_UNIFIED_A_2D_U32_F32; break; } - Ops.push_back(TexRef); - Ops.push_back(SampRef); - - // Copy over indices - for (unsigned i = 3; i < N->getNumOperands(); ++i) { + // Copy over operands + for (unsigned i = 1; i < N->getNumOperands(); ++i) { Ops.push_back(N->getOperand(i)); } @@ -2932,6 +3545,402 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { SmallVector<SDValue, 8> Ops; switch (N->getOpcode()) { default: return nullptr; + case NVPTXISD::Suld1DI8Clamp: + Opc = NVPTX::SULD_1D_I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DI16Clamp: + Opc = NVPTX::SULD_1D_I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DI32Clamp: + Opc = NVPTX::SULD_1D_I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DI64Clamp: + Opc = NVPTX::SULD_1D_I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV2I8Clamp: + Opc = NVPTX::SULD_1D_V2I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV2I16Clamp: + Opc = NVPTX::SULD_1D_V2I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV2I32Clamp: + Opc = NVPTX::SULD_1D_V2I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV2I64Clamp: + Opc = NVPTX::SULD_1D_V2I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV4I8Clamp: + Opc = NVPTX::SULD_1D_V4I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV4I16Clamp: + Opc = NVPTX::SULD_1D_V4I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV4I32Clamp: + Opc = NVPTX::SULD_1D_V4I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayI8Clamp: + Opc = NVPTX::SULD_1D_ARRAY_I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayI16Clamp: + Opc = NVPTX::SULD_1D_ARRAY_I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayI32Clamp: + Opc = NVPTX::SULD_1D_ARRAY_I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayI64Clamp: + Opc = NVPTX::SULD_1D_ARRAY_I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV2I8Clamp: + Opc = NVPTX::SULD_1D_ARRAY_V2I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV2I16Clamp: + Opc = NVPTX::SULD_1D_ARRAY_V2I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV2I32Clamp: + Opc = NVPTX::SULD_1D_ARRAY_V2I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV2I64Clamp: + Opc = NVPTX::SULD_1D_ARRAY_V2I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV4I8Clamp: + Opc = NVPTX::SULD_1D_ARRAY_V4I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV4I16Clamp: + Opc = NVPTX::SULD_1D_ARRAY_V4I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV4I32Clamp: + Opc = NVPTX::SULD_1D_ARRAY_V4I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DI8Clamp: + Opc = NVPTX::SULD_2D_I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DI16Clamp: + Opc = NVPTX::SULD_2D_I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DI32Clamp: + Opc = NVPTX::SULD_2D_I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DI64Clamp: + Opc = NVPTX::SULD_2D_I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV2I8Clamp: + Opc = NVPTX::SULD_2D_V2I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV2I16Clamp: + Opc = NVPTX::SULD_2D_V2I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV2I32Clamp: + Opc = NVPTX::SULD_2D_V2I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV2I64Clamp: + Opc = NVPTX::SULD_2D_V2I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV4I8Clamp: + Opc = NVPTX::SULD_2D_V4I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV4I16Clamp: + Opc = NVPTX::SULD_2D_V4I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV4I32Clamp: + Opc = NVPTX::SULD_2D_V4I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayI8Clamp: + Opc = NVPTX::SULD_2D_ARRAY_I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayI16Clamp: + Opc = NVPTX::SULD_2D_ARRAY_I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayI32Clamp: + Opc = NVPTX::SULD_2D_ARRAY_I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayI64Clamp: + Opc = NVPTX::SULD_2D_ARRAY_I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV2I8Clamp: + Opc = NVPTX::SULD_2D_ARRAY_V2I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV2I16Clamp: + Opc = NVPTX::SULD_2D_ARRAY_V2I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV2I32Clamp: + Opc = NVPTX::SULD_2D_ARRAY_V2I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV2I64Clamp: + Opc = NVPTX::SULD_2D_ARRAY_V2I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV4I8Clamp: + Opc = NVPTX::SULD_2D_ARRAY_V4I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV4I16Clamp: + Opc = NVPTX::SULD_2D_ARRAY_V4I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV4I32Clamp: + Opc = NVPTX::SULD_2D_ARRAY_V4I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DI8Clamp: + Opc = NVPTX::SULD_3D_I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DI16Clamp: + Opc = NVPTX::SULD_3D_I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DI32Clamp: + Opc = NVPTX::SULD_3D_I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DI64Clamp: + Opc = NVPTX::SULD_3D_I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV2I8Clamp: + Opc = NVPTX::SULD_3D_V2I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV2I16Clamp: + Opc = NVPTX::SULD_3D_V2I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV2I32Clamp: + Opc = NVPTX::SULD_3D_V2I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV2I64Clamp: + Opc = NVPTX::SULD_3D_V2I64_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV4I8Clamp: + Opc = NVPTX::SULD_3D_V4I8_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV4I16Clamp: + Opc = NVPTX::SULD_3D_V4I16_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV4I32Clamp: + Opc = NVPTX::SULD_3D_V4I32_CLAMP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld1DI8Trap: Opc = NVPTX::SULD_1D_I8_TRAP; Ops.push_back(TexHandle); @@ -2950,6 +3959,12 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(2)); Ops.push_back(Chain); break; + case NVPTXISD::Suld1DI64Trap: + Opc = NVPTX::SULD_1D_I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld1DV2I8Trap: Opc = NVPTX::SULD_1D_V2I8_TRAP; Ops.push_back(TexHandle); @@ -2968,6 +3983,12 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(2)); Ops.push_back(Chain); break; + case NVPTXISD::Suld1DV2I64Trap: + Opc = NVPTX::SULD_1D_V2I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld1DV4I8Trap: Opc = NVPTX::SULD_1D_V4I8_TRAP; Ops.push_back(TexHandle); @@ -3007,6 +4028,13 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(3)); Ops.push_back(Chain); break; + case NVPTXISD::Suld1DArrayI64Trap: + Opc = NVPTX::SULD_1D_ARRAY_I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld1DArrayV2I8Trap: Opc = NVPTX::SULD_1D_ARRAY_V2I8_TRAP; Ops.push_back(TexHandle); @@ -3028,6 +4056,13 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(3)); Ops.push_back(Chain); break; + case NVPTXISD::Suld1DArrayV2I64Trap: + Opc = NVPTX::SULD_1D_ARRAY_V2I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld1DArrayV4I8Trap: Opc = NVPTX::SULD_1D_ARRAY_V4I8_TRAP; Ops.push_back(TexHandle); @@ -3070,6 +4105,13 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(3)); Ops.push_back(Chain); break; + case NVPTXISD::Suld2DI64Trap: + Opc = NVPTX::SULD_2D_I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld2DV2I8Trap: Opc = NVPTX::SULD_2D_V2I8_TRAP; Ops.push_back(TexHandle); @@ -3091,6 +4133,13 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(3)); Ops.push_back(Chain); break; + case NVPTXISD::Suld2DV2I64Trap: + Opc = NVPTX::SULD_2D_V2I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld2DV4I8Trap: Opc = NVPTX::SULD_2D_V4I8_TRAP; Ops.push_back(TexHandle); @@ -3136,6 +4185,14 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(4)); Ops.push_back(Chain); break; + case NVPTXISD::Suld2DArrayI64Trap: + Opc = NVPTX::SULD_2D_ARRAY_I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld2DArrayV2I8Trap: Opc = NVPTX::SULD_2D_ARRAY_V2I8_TRAP; Ops.push_back(TexHandle); @@ -3160,6 +4217,14 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(4)); Ops.push_back(Chain); break; + case NVPTXISD::Suld2DArrayV2I64Trap: + Opc = NVPTX::SULD_2D_ARRAY_V2I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld2DArrayV4I8Trap: Opc = NVPTX::SULD_2D_ARRAY_V4I8_TRAP; Ops.push_back(TexHandle); @@ -3208,6 +4273,14 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(4)); Ops.push_back(Chain); break; + case NVPTXISD::Suld3DI64Trap: + Opc = NVPTX::SULD_3D_I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld3DV2I8Trap: Opc = NVPTX::SULD_3D_V2I8_TRAP; Ops.push_back(TexHandle); @@ -3232,6 +4305,14 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(4)); Ops.push_back(Chain); break; + case NVPTXISD::Suld3DV2I64Trap: + Opc = NVPTX::SULD_3D_V2I64_TRAP; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; case NVPTXISD::Suld3DV4I8Trap: Opc = NVPTX::SULD_3D_V4I8_TRAP; Ops.push_back(TexHandle); @@ -3256,11 +4337,408 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) { Ops.push_back(N->getOperand(4)); Ops.push_back(Chain); break; + case NVPTXISD::Suld1DI8Zero: + Opc = NVPTX::SULD_1D_I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DI16Zero: + Opc = NVPTX::SULD_1D_I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DI32Zero: + Opc = NVPTX::SULD_1D_I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DI64Zero: + Opc = NVPTX::SULD_1D_I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV2I8Zero: + Opc = NVPTX::SULD_1D_V2I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV2I16Zero: + Opc = NVPTX::SULD_1D_V2I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV2I32Zero: + Opc = NVPTX::SULD_1D_V2I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV2I64Zero: + Opc = NVPTX::SULD_1D_V2I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV4I8Zero: + Opc = NVPTX::SULD_1D_V4I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV4I16Zero: + Opc = NVPTX::SULD_1D_V4I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DV4I32Zero: + Opc = NVPTX::SULD_1D_V4I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayI8Zero: + Opc = NVPTX::SULD_1D_ARRAY_I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayI16Zero: + Opc = NVPTX::SULD_1D_ARRAY_I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayI32Zero: + Opc = NVPTX::SULD_1D_ARRAY_I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayI64Zero: + Opc = NVPTX::SULD_1D_ARRAY_I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV2I8Zero: + Opc = NVPTX::SULD_1D_ARRAY_V2I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV2I16Zero: + Opc = NVPTX::SULD_1D_ARRAY_V2I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV2I32Zero: + Opc = NVPTX::SULD_1D_ARRAY_V2I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV2I64Zero: + Opc = NVPTX::SULD_1D_ARRAY_V2I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV4I8Zero: + Opc = NVPTX::SULD_1D_ARRAY_V4I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV4I16Zero: + Opc = NVPTX::SULD_1D_ARRAY_V4I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld1DArrayV4I32Zero: + Opc = NVPTX::SULD_1D_ARRAY_V4I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DI8Zero: + Opc = NVPTX::SULD_2D_I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DI16Zero: + Opc = NVPTX::SULD_2D_I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DI32Zero: + Opc = NVPTX::SULD_2D_I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DI64Zero: + Opc = NVPTX::SULD_2D_I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV2I8Zero: + Opc = NVPTX::SULD_2D_V2I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV2I16Zero: + Opc = NVPTX::SULD_2D_V2I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV2I32Zero: + Opc = NVPTX::SULD_2D_V2I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV2I64Zero: + Opc = NVPTX::SULD_2D_V2I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV4I8Zero: + Opc = NVPTX::SULD_2D_V4I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV4I16Zero: + Opc = NVPTX::SULD_2D_V4I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DV4I32Zero: + Opc = NVPTX::SULD_2D_V4I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayI8Zero: + Opc = NVPTX::SULD_2D_ARRAY_I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayI16Zero: + Opc = NVPTX::SULD_2D_ARRAY_I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayI32Zero: + Opc = NVPTX::SULD_2D_ARRAY_I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayI64Zero: + Opc = NVPTX::SULD_2D_ARRAY_I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV2I8Zero: + Opc = NVPTX::SULD_2D_ARRAY_V2I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV2I16Zero: + Opc = NVPTX::SULD_2D_ARRAY_V2I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV2I32Zero: + Opc = NVPTX::SULD_2D_ARRAY_V2I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV2I64Zero: + Opc = NVPTX::SULD_2D_ARRAY_V2I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV4I8Zero: + Opc = NVPTX::SULD_2D_ARRAY_V4I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV4I16Zero: + Opc = NVPTX::SULD_2D_ARRAY_V4I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld2DArrayV4I32Zero: + Opc = NVPTX::SULD_2D_ARRAY_V4I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DI8Zero: + Opc = NVPTX::SULD_3D_I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DI16Zero: + Opc = NVPTX::SULD_3D_I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DI32Zero: + Opc = NVPTX::SULD_3D_I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DI64Zero: + Opc = NVPTX::SULD_3D_I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV2I8Zero: + Opc = NVPTX::SULD_3D_V2I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV2I16Zero: + Opc = NVPTX::SULD_3D_V2I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV2I32Zero: + Opc = NVPTX::SULD_3D_V2I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV2I64Zero: + Opc = NVPTX::SULD_3D_V2I64_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV4I8Zero: + Opc = NVPTX::SULD_3D_V4I8_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV4I16Zero: + Opc = NVPTX::SULD_3D_V4I16_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; + case NVPTXISD::Suld3DV4I32Zero: + Opc = NVPTX::SULD_3D_V4I32_ZERO; + Ops.push_back(TexHandle); + Ops.push_back(N->getOperand(2)); + Ops.push_back(N->getOperand(3)); + Ops.push_back(N->getOperand(4)); + Ops.push_back(Chain); + break; } Ret = CurDAG->getMachineNode(Opc, SDLoc(N), N->getVTList(), Ops); return Ret; } + /// SelectBFE - Look for instruction sequences that can be made more efficient /// by using the 'bfe' (bit-field extract) PTX instruction SDNode *NVPTXDAGToDAGISel::SelectBFE(SDNode *N) { @@ -3563,17 +5041,10 @@ bool NVPTXDAGToDAGISel::SelectADDRri64(SDNode *OpNode, SDValue Addr, bool NVPTXDAGToDAGISel::ChkMemSDNodeAddressSpace(SDNode *N, unsigned int spN) const { const Value *Src = nullptr; - // Even though MemIntrinsicSDNode is a subclas of MemSDNode, - // the classof() for MemSDNode does not include MemIntrinsicSDNode - // (See SelectionDAGNodes.h). So we need to check for both. if (MemSDNode *mN = dyn_cast<MemSDNode>(N)) { if (spN == 0 && mN->getMemOperand()->getPseudoValue()) return true; Src = mN->getMemOperand()->getValue(); - } else if (MemSDNode *mN = dyn_cast<MemIntrinsicSDNode>(N)) { - if (spN == 0 && mN->getMemOperand()->getPseudoValue()) - return true; - Src = mN->getMemOperand()->getValue(); } if (!Src) return false; |