Re-sort includes with sort-includes.py and insert raw_ostream.h where it's used.

[oota-llvm.git] / lib / Target / NVPTX / NVPTXISelDAGToDAG.cpp

diff --git a/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp b/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp
index cd308806c36a5e9a1ee493b0f97a9c8e2f481cb7..52c5e1b87886a39f9095820b0d60d132d0c769e3 100644 (file)
--- a/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp
+++ b/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp
@@ -24,16 +24,10 @@ using namespace llvm;
 
 #define DEBUG_TYPE "nvptx-isel"
 
-static cl::opt<int>
-FMAContractLevel("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::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>
@@ -56,21 +50,15 @@ FunctionPass *llvm::createNVPTXISelDag(NVPTXTargetMachine &TM,
 
 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);
-
+    : SelectionDAGISel(tm, OptLevel), TM(tm) {
   doMulWide = (OptLevel > 0);
 }
 
+bool NVPTXDAGToDAGISel::runOnMachineFunction(MachineFunction &MF) {
+    Subtarget = &static_cast<const NVPTXSubtarget &>(MF.getSubtarget());
+    return SelectionDAGISel::runOnMachineFunction(MF);
+}
+
 int NVPTXDAGToDAGISel::getDivF32Level() const {
   if (UsePrecDivF32.getNumOccurrences() > 0) {
     // If nvptx-prec-div32=N is used on the command-line, always honor it
@@ -90,10 +78,7 @@ bool NVPTXDAGToDAGISel::usePrecSqrtF32() const {
     return UsePrecSqrtF32;
   } else {
     // Otherwise, use sqrt.approx if fast math is enabled
-    if (TM.Options.UnsafeFPMath)
-      return false;
-    else
-      return true;
+    return !TM.Options.UnsafeFPMath;
   }
 }
 
@@ -105,14 +90,17 @@ bool NVPTXDAGToDAGISel::useF32FTZ() const {
     const Function *F = MF->getFunction();
     // Otherwise, check for an nvptx-f32ftz attribute on the function
     if (F->hasFnAttribute("nvptx-f32ftz"))
-      return (F->getAttributes().getAttribute(AttributeSet::FunctionIndex,
-                                              "nvptx-f32ftz")
-                                              .getValueAsString() == "true");
+      return F->getFnAttribute("nvptx-f32ftz").getValueAsString() == "true";
     else
       return false;
   }
 }
 
+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) {
@@ -138,7 +126,7 @@ SDNode *NVPTXDAGToDAGISel::Select(SDNode *N) {
   case NVPTXISD::LDGV4:
   case NVPTXISD::LDUV2:
   case NVPTXISD::LDUV4:
-    ResNode = SelectLDGLDUVector(N);
+    ResNode = SelectLDGLDU(N);
     break;
   case NVPTXISD::StoreV2:
   case NVPTXISD::StoreV4:
@@ -164,95 +152,352 @@ SDNode *NVPTXDAGToDAGISel::Select(SDNode *N) {
   case ISD::INTRINSIC_WO_CHAIN:
     ResNode = SelectIntrinsicNoChain(N);
     break;
-  case NVPTXISD::Tex1DFloatI32:
+  case ISD::INTRINSIC_W_CHAIN:
+    ResNode = SelectIntrinsicChain(N);
+    break;
+  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:
+  case ISD::SRA:
+  case ISD::SRL:
+    // Try to select BFE
+    ResNode = SelectBFE(N);
+    break;
   case ISD::ADDRSPACECAST:
     ResNode = SelectAddrSpaceCast(N);
     break;
@@ -264,8 +509,22 @@ SDNode *NVPTXDAGToDAGISel::Select(SDNode *N) {
   return SelectCode(N);
 }
 
-static unsigned int getCodeAddrSpace(MemSDNode *N,
-                                     const NVPTXSubtarget &Subtarget) {
+SDNode *NVPTXDAGToDAGISel::SelectIntrinsicChain(SDNode *N) {
+  unsigned IID = cast<ConstantSDNode>(N->getOperand(1))->getZExtValue();
+  switch (IID) {
+  default:
+    return NULL;
+  case Intrinsic::nvvm_ldg_global_f:
+  case Intrinsic::nvvm_ldg_global_i:
+  case Intrinsic::nvvm_ldg_global_p:
+  case Intrinsic::nvvm_ldu_global_f:
+  case Intrinsic::nvvm_ldu_global_i:
+  case Intrinsic::nvvm_ldu_global_p:
+    return SelectLDGLDU(N);
+  }
+}
+
+static unsigned int getCodeAddrSpace(MemSDNode *N) {
   const Value *Src = N->getMemOperand()->getValue();
 
   if (!Src)
@@ -318,20 +577,16 @@ SDNode *NVPTXDAGToDAGISel::SelectAddrSpaceCast(SDNode *N) {
     switch (SrcAddrSpace) {
     default: report_fatal_error("Bad address space in addrspacecast");
     case ADDRESS_SPACE_GLOBAL:
-      Opc = Subtarget.is64Bit() ? NVPTX::cvta_global_yes_64
-                                : NVPTX::cvta_global_yes;
+      Opc = TM.is64Bit() ? NVPTX::cvta_global_yes_64 : NVPTX::cvta_global_yes;
       break;
     case ADDRESS_SPACE_SHARED:
-      Opc = Subtarget.is64Bit() ? NVPTX::cvta_shared_yes_64
-                                : NVPTX::cvta_shared_yes;
+      Opc = TM.is64Bit() ? NVPTX::cvta_shared_yes_64 : NVPTX::cvta_shared_yes;
       break;
     case ADDRESS_SPACE_CONST:
-      Opc = Subtarget.is64Bit() ? NVPTX::cvta_const_yes_64
-                                : NVPTX::cvta_const_yes;
+      Opc = TM.is64Bit() ? NVPTX::cvta_const_yes_64 : NVPTX::cvta_const_yes;
       break;
     case ADDRESS_SPACE_LOCAL:
-      Opc = Subtarget.is64Bit() ? NVPTX::cvta_local_yes_64
-                                : NVPTX::cvta_local_yes;
+      Opc = TM.is64Bit() ? NVPTX::cvta_local_yes_64 : NVPTX::cvta_local_yes;
       break;
     }
     return CurDAG->getMachineNode(Opc, SDLoc(N), N->getValueType(0), Src);
@@ -343,20 +598,20 @@ SDNode *NVPTXDAGToDAGISel::SelectAddrSpaceCast(SDNode *N) {
     switch (DstAddrSpace) {
     default: report_fatal_error("Bad address space in addrspacecast");
     case ADDRESS_SPACE_GLOBAL:
-      Opc = Subtarget.is64Bit() ? NVPTX::cvta_to_global_yes_64
-                                : NVPTX::cvta_to_global_yes;
+      Opc = TM.is64Bit() ? NVPTX::cvta_to_global_yes_64
+                         : NVPTX::cvta_to_global_yes;
       break;
     case ADDRESS_SPACE_SHARED:
-      Opc = Subtarget.is64Bit() ? NVPTX::cvta_to_shared_yes_64
-                                : NVPTX::cvta_to_shared_yes;
+      Opc = TM.is64Bit() ? NVPTX::cvta_to_shared_yes_64
+                         : NVPTX::cvta_to_shared_yes;
       break;
     case ADDRESS_SPACE_CONST:
-      Opc = Subtarget.is64Bit() ? NVPTX::cvta_to_const_yes_64
-                                : NVPTX::cvta_to_const_yes;
+      Opc =
+          TM.is64Bit() ? NVPTX::cvta_to_const_yes_64 : NVPTX::cvta_to_const_yes;
       break;
     case ADDRESS_SPACE_LOCAL:
-      Opc = Subtarget.is64Bit() ? NVPTX::cvta_to_local_yes_64
-                                : NVPTX::cvta_to_local_yes;
+      Opc =
+          TM.is64Bit() ? NVPTX::cvta_to_local_yes_64 : NVPTX::cvta_to_local_yes;
       break;
     }
     return CurDAG->getMachineNode(Opc, SDLoc(N), N->getValueType(0), Src);
@@ -377,7 +632,7 @@ SDNode *NVPTXDAGToDAGISel::SelectLoad(SDNode *N) {
     return nullptr;
 
   // Address Space Setting
-  unsigned int codeAddrSpace = getCodeAddrSpace(LD, Subtarget);
+  unsigned int codeAddrSpace = getCodeAddrSpace(LD);
 
   // Volatile Setting
   // - .volatile is only availalble for .global and .shared
@@ -452,9 +707,8 @@ SDNode *NVPTXDAGToDAGISel::SelectLoad(SDNode *N) {
                       getI32Imm(vecType), getI32Imm(fromType),
                       getI32Imm(fromTypeWidth), Addr, Chain };
     NVPTXLD = CurDAG->getMachineNode(Opcode, dl, TargetVT, MVT::Other, Ops);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRsi64(N1.getNode(), N1, Base, Offset)
-                 : SelectADDRsi(N1.getNode(), N1, Base, Offset)) {
+  } else if (TM.is64Bit() ? SelectADDRsi64(N1.getNode(), N1, Base, Offset)
+                          : SelectADDRsi(N1.getNode(), N1, Base, Offset)) {
     switch (TargetVT) {
     case MVT::i8:
       Opcode = NVPTX::LD_i8_asi;
@@ -481,10 +735,9 @@ SDNode *NVPTXDAGToDAGISel::SelectLoad(SDNode *N) {
                       getI32Imm(vecType), getI32Imm(fromType),
                       getI32Imm(fromTypeWidth), Base, Offset, Chain };
     NVPTXLD = CurDAG->getMachineNode(Opcode, dl, TargetVT, MVT::Other, Ops);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRri64(N1.getNode(), N1, Base, Offset)
-                 : SelectADDRri(N1.getNode(), N1, Base, Offset)) {
-    if (Subtarget.is64Bit()) {
+  } else if (TM.is64Bit() ? SelectADDRri64(N1.getNode(), N1, Base, Offset)
+                          : SelectADDRri(N1.getNode(), N1, Base, Offset)) {
+    if (TM.is64Bit()) {
       switch (TargetVT) {
       case MVT::i8:
         Opcode = NVPTX::LD_i8_ari_64;
@@ -536,7 +789,7 @@ SDNode *NVPTXDAGToDAGISel::SelectLoad(SDNode *N) {
                       getI32Imm(fromTypeWidth), Base, Offset, Chain };
     NVPTXLD = CurDAG->getMachineNode(Opcode, dl, TargetVT, MVT::Other, Ops);
   } else {
-    if (Subtarget.is64Bit()) {
+    if (TM.is64Bit()) {
       switch (TargetVT) {
       case MVT::i8:
         Opcode = NVPTX::LD_i8_areg_64;
@@ -613,7 +866,7 @@ SDNode *NVPTXDAGToDAGISel::SelectLoadVector(SDNode *N) {
     return nullptr;
 
   // Address Space Setting
-  unsigned int CodeAddrSpace = getCodeAddrSpace(MemSD, Subtarget);
+  unsigned int CodeAddrSpace = getCodeAddrSpace(MemSD);
 
   // Volatile Setting
   // - .volatile is only availalble for .global and .shared
@@ -713,9 +966,8 @@ SDNode *NVPTXDAGToDAGISel::SelectLoadVector(SDNode *N) {
                       getI32Imm(VecType), getI32Imm(FromType),
                       getI32Imm(FromTypeWidth), Addr, Chain };
     LD = CurDAG->getMachineNode(Opcode, DL, N->getVTList(), Ops);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRsi64(Op1.getNode(), Op1, Base, Offset)
-                 : SelectADDRsi(Op1.getNode(), Op1, Base, Offset)) {
+  } else if (TM.is64Bit() ? SelectADDRsi64(Op1.getNode(), Op1, Base, Offset)
+                          : SelectADDRsi(Op1.getNode(), Op1, Base, Offset)) {
     switch (N->getOpcode()) {
     default:
       return nullptr;
@@ -767,10 +1019,9 @@ SDNode *NVPTXDAGToDAGISel::SelectLoadVector(SDNode *N) {
                       getI32Imm(VecType), getI32Imm(FromType),
                       getI32Imm(FromTypeWidth), Base, Offset, Chain };
     LD = CurDAG->getMachineNode(Opcode, DL, N->getVTList(), Ops);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRri64(Op1.getNode(), Op1, Base, Offset)
-                 : SelectADDRri(Op1.getNode(), Op1, Base, Offset)) {
-    if (Subtarget.is64Bit()) {
+  } else if (TM.is64Bit() ? SelectADDRri64(Op1.getNode(), Op1, Base, Offset)
+                          : SelectADDRri(Op1.getNode(), Op1, Base, Offset)) {
+    if (TM.is64Bit()) {
       switch (N->getOpcode()) {
       default:
         return nullptr;
@@ -872,7 +1123,7 @@ SDNode *NVPTXDAGToDAGISel::SelectLoadVector(SDNode *N) {
 
     LD = CurDAG->getMachineNode(Opcode, DL, N->getVTList(), Ops);
   } else {
-    if (Subtarget.is64Bit()) {
+    if (TM.is64Bit()) {
       switch (N->getOpcode()) {
       default:
         return nullptr;
@@ -981,22 +1232,101 @@ SDNode *NVPTXDAGToDAGISel::SelectLoadVector(SDNode *N) {
   return LD;
 }
 
-SDNode *NVPTXDAGToDAGISel::SelectLDGLDUVector(SDNode *N) {
+SDNode *NVPTXDAGToDAGISel::SelectLDGLDU(SDNode *N) {
 
   SDValue Chain = N->getOperand(0);
-  SDValue Op1 = N->getOperand(1);
+  SDValue Op1;
+  MemSDNode *Mem;
+  bool IsLDG = true;
+
+  // If this is an LDG intrinsic, the address is the third operand. Its its an
+  // LDG/LDU SD node (from custom vector handling), then its the second operand
+  if (N->getOpcode() == ISD::INTRINSIC_W_CHAIN) {
+    Op1 = N->getOperand(2);
+    Mem = cast<MemIntrinsicSDNode>(N);
+    unsigned IID = cast<ConstantSDNode>(N->getOperand(1))->getZExtValue();
+    switch (IID) {
+    default:
+      return NULL;
+    case Intrinsic::nvvm_ldg_global_f:
+    case Intrinsic::nvvm_ldg_global_i:
+    case Intrinsic::nvvm_ldg_global_p:
+      IsLDG = true;
+      break;
+    case Intrinsic::nvvm_ldu_global_f:
+    case Intrinsic::nvvm_ldu_global_i:
+    case Intrinsic::nvvm_ldu_global_p:
+      IsLDG = false;
+      break;
+    }
+  } else {
+    Op1 = N->getOperand(1);
+    Mem = cast<MemSDNode>(N);
+  }
+
   unsigned Opcode;
   SDLoc DL(N);
   SDNode *LD;
-  MemSDNode *Mem = cast<MemSDNode>(N);
   SDValue Base, Offset, Addr;
 
-  EVT EltVT = Mem->getMemoryVT().getVectorElementType();
+  EVT EltVT = Mem->getMemoryVT();
+  if (EltVT.isVector()) {
+    EltVT = EltVT.getVectorElementType();
+  }
 
   if (SelectDirectAddr(Op1, Addr)) {
     switch (N->getOpcode()) {
     default:
       return nullptr;
+    case ISD::INTRINSIC_W_CHAIN:
+      if (IsLDG) {
+        switch (EltVT.getSimpleVT().SimpleTy) {
+        default:
+          return nullptr;
+        case MVT::i8:
+          Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i8avar;
+          break;
+        case MVT::i16:
+          Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i16avar;
+          break;
+        case MVT::i32:
+          Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i32avar;
+          break;
+        case MVT::i64:
+          Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i64avar;
+          break;
+        case MVT::f32:
+          Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f32avar;
+          break;
+        case MVT::f64:
+          Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f64avar;
+          break;
+        }
+      } else {
+        switch (EltVT.getSimpleVT().SimpleTy) {
+        default:
+          return nullptr;
+        case MVT::i8:
+          Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i8avar;
+          break;
+        case MVT::i16:
+          Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i16avar;
+          break;
+        case MVT::i32:
+          Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i32avar;
+          break;
+        case MVT::i64:
+          Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i64avar;
+          break;
+        case MVT::f32:
+          Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f32avar;
+          break;
+        case MVT::f64:
+          Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f64avar;
+          break;
+        }
+      }
+      break;
     case NVPTXISD::LDGV2:
       switch (EltVT.getSimpleVT().SimpleTy) {
       default:
@@ -1085,13 +1415,61 @@ SDNode *NVPTXDAGToDAGISel::SelectLDGLDUVector(SDNode *N) {
 
     SDValue Ops[] = { Addr, Chain };
     LD = CurDAG->getMachineNode(Opcode, DL, N->getVTList(), Ops);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRri64(Op1.getNode(), Op1, Base, Offset)
-                 : SelectADDRri(Op1.getNode(), Op1, Base, Offset)) {
-    if (Subtarget.is64Bit()) {
+  } else if (TM.is64Bit() ? SelectADDRri64(Op1.getNode(), Op1, Base, Offset)
+                          : SelectADDRri(Op1.getNode(), Op1, Base, Offset)) {
+    if (TM.is64Bit()) {
       switch (N->getOpcode()) {
       default:
         return nullptr;
+      case ISD::INTRINSIC_W_CHAIN:
+        if (IsLDG) {
+          switch (EltVT.getSimpleVT().SimpleTy) {
+          default:
+            return nullptr;
+          case MVT::i8:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i8ari64;
+            break;
+          case MVT::i16:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i16ari64;
+            break;
+          case MVT::i32:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i32ari64;
+            break;
+          case MVT::i64:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i64ari64;
+            break;
+          case MVT::f32:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f32ari64;
+            break;
+          case MVT::f64:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f64ari64;
+            break;
+          }
+        } else {
+          switch (EltVT.getSimpleVT().SimpleTy) {
+          default:
+            return nullptr;
+          case MVT::i8:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i8ari64;
+            break;
+          case MVT::i16:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i16ari64;
+            break;
+          case MVT::i32:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i32ari64;
+            break;
+          case MVT::i64:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i64ari64;
+            break;
+          case MVT::f32:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f32ari64;
+            break;
+          case MVT::f64:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f64ari64;
+            break;
+          }
+        }
+        break;
       case NVPTXISD::LDGV2:
         switch (EltVT.getSimpleVT().SimpleTy) {
         default:
@@ -1181,6 +1559,55 @@ SDNode *NVPTXDAGToDAGISel::SelectLDGLDUVector(SDNode *N) {
       switch (N->getOpcode()) {
       default:
         return nullptr;
+      case ISD::INTRINSIC_W_CHAIN:
+        if (IsLDG) {
+          switch (EltVT.getSimpleVT().SimpleTy) {
+          default:
+            return nullptr;
+          case MVT::i8:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i8ari;
+            break;
+          case MVT::i16:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i16ari;
+            break;
+          case MVT::i32:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i32ari;
+            break;
+          case MVT::i64:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i64ari;
+            break;
+          case MVT::f32:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f32ari;
+            break;
+          case MVT::f64:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f64ari;
+            break;
+          }
+        } else {
+          switch (EltVT.getSimpleVT().SimpleTy) {
+          default:
+            return nullptr;
+          case MVT::i8:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i8ari;
+            break;
+          case MVT::i16:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i16ari;
+            break;
+          case MVT::i32:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i32ari;
+            break;
+          case MVT::i64:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i64ari;
+            break;
+          case MVT::f32:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f32ari;
+            break;
+          case MVT::f64:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f64ari;
+            break;
+          }
+        }
+        break;
       case NVPTXISD::LDGV2:
         switch (EltVT.getSimpleVT().SimpleTy) {
         default:
@@ -1272,10 +1699,59 @@ SDNode *NVPTXDAGToDAGISel::SelectLDGLDUVector(SDNode *N) {
 
     LD = CurDAG->getMachineNode(Opcode, DL, N->getVTList(), Ops);
   } else {
-    if (Subtarget.is64Bit()) {
+    if (TM.is64Bit()) {
       switch (N->getOpcode()) {
       default:
         return nullptr;
+      case ISD::INTRINSIC_W_CHAIN:
+        if (IsLDG) {
+          switch (EltVT.getSimpleVT().SimpleTy) {
+          default:
+            return nullptr;
+          case MVT::i8:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i8areg64;
+            break;
+          case MVT::i16:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i16areg64;
+            break;
+          case MVT::i32:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i32areg64;
+            break;
+          case MVT::i64:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i64areg64;
+            break;
+          case MVT::f32:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f32areg64;
+            break;
+          case MVT::f64:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f64areg64;
+            break;
+          }
+        } else {
+          switch (EltVT.getSimpleVT().SimpleTy) {
+          default:
+            return nullptr;
+          case MVT::i8:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i8areg64;
+            break;
+          case MVT::i16:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i16areg64;
+            break;
+          case MVT::i32:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i32areg64;
+            break;
+          case MVT::i64:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i64areg64;
+            break;
+          case MVT::f32:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f32areg64;
+            break;
+          case MVT::f64:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f64areg64;
+            break;
+          }
+        }
+        break;
       case NVPTXISD::LDGV2:
         switch (EltVT.getSimpleVT().SimpleTy) {
         default:
@@ -1365,6 +1841,55 @@ SDNode *NVPTXDAGToDAGISel::SelectLDGLDUVector(SDNode *N) {
       switch (N->getOpcode()) {
       default:
         return nullptr;
+      case ISD::INTRINSIC_W_CHAIN:
+        if (IsLDG) {
+          switch (EltVT.getSimpleVT().SimpleTy) {
+          default:
+            return nullptr;
+          case MVT::i8:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i8areg;
+            break;
+          case MVT::i16:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i16areg;
+            break;
+          case MVT::i32:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i32areg;
+            break;
+          case MVT::i64:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_i64areg;
+            break;
+          case MVT::f32:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f32areg;
+            break;
+          case MVT::f64:
+            Opcode = NVPTX::INT_PTX_LDG_GLOBAL_f64areg;
+            break;
+          }
+        } else {
+          switch (EltVT.getSimpleVT().SimpleTy) {
+          default:
+            return nullptr;
+          case MVT::i8:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i8areg;
+            break;
+          case MVT::i16:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i16areg;
+            break;
+          case MVT::i32:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i32areg;
+            break;
+          case MVT::i64:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_i64areg;
+            break;
+          case MVT::f32:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f32areg;
+            break;
+          case MVT::f64:
+            Opcode = NVPTX::INT_PTX_LDU_GLOBAL_f64areg;
+            break;
+          }
+        }
+        break;
       case NVPTXISD::LDGV2:
         switch (EltVT.getSimpleVT().SimpleTy) {
         default:
@@ -1457,7 +1982,7 @@ SDNode *NVPTXDAGToDAGISel::SelectLDGLDUVector(SDNode *N) {
   }
 
   MachineSDNode::mmo_iterator MemRefs0 = MF->allocateMemRefsArray(1);
-  MemRefs0[0] = cast<MemSDNode>(N)->getMemOperand();
+  MemRefs0[0] = Mem->getMemOperand();
   cast<MachineSDNode>(LD)->setMemRefs(MemRefs0, MemRefs0 + 1);
 
   return LD;
@@ -1477,7 +2002,7 @@ SDNode *NVPTXDAGToDAGISel::SelectStore(SDNode *N) {
     return nullptr;
 
   // Address Space Setting
-  unsigned int codeAddrSpace = getCodeAddrSpace(ST, Subtarget);
+  unsigned int codeAddrSpace = getCodeAddrSpace(ST);
 
   // Volatile Setting
   // - .volatile is only availalble for .global and .shared
@@ -1547,9 +2072,8 @@ SDNode *NVPTXDAGToDAGISel::SelectStore(SDNode *N) {
                       getI32Imm(vecType), getI32Imm(toType),
                       getI32Imm(toTypeWidth), Addr, Chain };
     NVPTXST = CurDAG->getMachineNode(Opcode, dl, MVT::Other, Ops);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRsi64(N2.getNode(), N2, Base, Offset)
-                 : SelectADDRsi(N2.getNode(), N2, Base, Offset)) {
+  } else if (TM.is64Bit() ? SelectADDRsi64(N2.getNode(), N2, Base, Offset)
+                          : SelectADDRsi(N2.getNode(), N2, Base, Offset)) {
     switch (SourceVT) {
     case MVT::i8:
       Opcode = NVPTX::ST_i8_asi;
@@ -1576,10 +2100,9 @@ SDNode *NVPTXDAGToDAGISel::SelectStore(SDNode *N) {
                       getI32Imm(vecType), getI32Imm(toType),
                       getI32Imm(toTypeWidth), Base, Offset, Chain };
     NVPTXST = CurDAG->getMachineNode(Opcode, dl, MVT::Other, Ops);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRri64(N2.getNode(), N2, Base, Offset)
-                 : SelectADDRri(N2.getNode(), N2, Base, Offset)) {
-    if (Subtarget.is64Bit()) {
+  } else if (TM.is64Bit() ? SelectADDRri64(N2.getNode(), N2, Base, Offset)
+                          : SelectADDRri(N2.getNode(), N2, Base, Offset)) {
+    if (TM.is64Bit()) {
       switch (SourceVT) {
       case MVT::i8:
         Opcode = NVPTX::ST_i8_ari_64;
@@ -1631,7 +2154,7 @@ SDNode *NVPTXDAGToDAGISel::SelectStore(SDNode *N) {
                       getI32Imm(toTypeWidth), Base, Offset, Chain };
     NVPTXST = CurDAG->getMachineNode(Opcode, dl, MVT::Other, Ops);
   } else {
-    if (Subtarget.is64Bit()) {
+    if (TM.is64Bit()) {
       switch (SourceVT) {
       case MVT::i8:
         Opcode = NVPTX::ST_i8_areg_64;
@@ -1705,7 +2228,7 @@ SDNode *NVPTXDAGToDAGISel::SelectStoreVector(SDNode *N) {
   EVT StoreVT = MemSD->getMemoryVT();
 
   // Address Space Setting
-  unsigned CodeAddrSpace = getCodeAddrSpace(MemSD, Subtarget);
+  unsigned CodeAddrSpace = getCodeAddrSpace(MemSD);
 
   if (CodeAddrSpace == NVPTX::PTXLdStInstCode::CONSTANT) {
     report_fatal_error("Cannot store to pointer that points to constant "
@@ -1808,9 +2331,8 @@ SDNode *NVPTXDAGToDAGISel::SelectStoreVector(SDNode *N) {
       break;
     }
     StOps.push_back(Addr);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRsi64(N2.getNode(), N2, Base, Offset)
-                 : SelectADDRsi(N2.getNode(), N2, Base, Offset)) {
+  } else if (TM.is64Bit() ? SelectADDRsi64(N2.getNode(), N2, Base, Offset)
+                          : SelectADDRsi(N2.getNode(), N2, Base, Offset)) {
     switch (N->getOpcode()) {
     default:
       return nullptr;
@@ -1859,10 +2381,9 @@ SDNode *NVPTXDAGToDAGISel::SelectStoreVector(SDNode *N) {
     }
     StOps.push_back(Base);
     StOps.push_back(Offset);
-  } else if (Subtarget.is64Bit()
-                 ? SelectADDRri64(N2.getNode(), N2, Base, Offset)
-                 : SelectADDRri(N2.getNode(), N2, Base, Offset)) {
-    if (Subtarget.is64Bit()) {
+  } else if (TM.is64Bit() ? SelectADDRri64(N2.getNode(), N2, Base, Offset)
+                          : SelectADDRri(N2.getNode(), N2, Base, Offset)) {
+    if (TM.is64Bit()) {
       switch (N->getOpcode()) {
       default:
         return nullptr;
@@ -1960,7 +2481,7 @@ SDNode *NVPTXDAGToDAGISel::SelectStoreVector(SDNode *N) {
     StOps.push_back(Base);
     StOps.push_back(Offset);
   } else {
-    if (Subtarget.is64Bit()) {
+    if (TM.is64Bit()) {
       switch (N->getOpcode()) {
       default:
         return nullptr;
@@ -2479,16 +3000,14 @@ 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;
@@ -2499,20 +3018,32 @@ 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::Tex1DS32FloatGrad:
+    Opc = NVPTX::TEX_1D_S32_F32_GRAD;
     break;
-  case NVPTXISD::Tex1DI32Float:
-    Opc = NVPTX::TEX_1D_I32_F32;
+  case NVPTXISD::Tex1DU32S32:
+    Opc = NVPTX::TEX_1D_U32_S32;
     break;
-  case NVPTXISD::Tex1DI32FloatLevel:
-    Opc = NVPTX::TEX_1D_I32_F32_LEVEL;
+  case NVPTXISD::Tex1DU32Float:
+    Opc = NVPTX::TEX_1D_U32_F32;
     break;
-  case NVPTXISD::Tex1DI32FloatGrad:
-    Opc = NVPTX::TEX_1D_I32_F32_GRAD;
+  case NVPTXISD::Tex1DU32FloatLevel:
+    Opc = NVPTX::TEX_1D_U32_F32_LEVEL;
     break;
-  case NVPTXISD::Tex1DArrayFloatI32:
-    Opc = NVPTX::TEX_1D_ARRAY_F32_I32;
+  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;
@@ -2523,20 +3054,32 @@ 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::Tex1DArrayS32FloatLevel:
+    Opc = NVPTX::TEX_1D_ARRAY_S32_F32_LEVEL;
+    break;
+  case NVPTXISD::Tex1DArrayS32FloatGrad:
+    Opc = NVPTX::TEX_1D_ARRAY_S32_F32_GRAD;
     break;
-  case NVPTXISD::Tex1DArrayI32Float:
-    Opc = NVPTX::TEX_1D_ARRAY_I32_F32;
+  case NVPTXISD::Tex1DArrayU32S32:
+    Opc = NVPTX::TEX_1D_ARRAY_U32_S32;
     break;
-  case NVPTXISD::Tex1DArrayI32FloatLevel:
-    Opc = NVPTX::TEX_1D_ARRAY_I32_F32_LEVEL;
+  case NVPTXISD::Tex1DArrayU32Float:
+    Opc = NVPTX::TEX_1D_ARRAY_U32_F32;
     break;
-  case NVPTXISD::Tex1DArrayI32FloatGrad:
-    Opc = NVPTX::TEX_1D_ARRAY_I32_F32_GRAD;
+  case NVPTXISD::Tex1DArrayU32FloatLevel:
+    Opc = NVPTX::TEX_1D_ARRAY_U32_F32_LEVEL;
     break;
-  case NVPTXISD::Tex2DFloatI32:
-    Opc = NVPTX::TEX_2D_F32_I32;
+  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;
@@ -2547,20 +3090,32 @@ 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;
@@ -2571,20 +3126,32 @@ 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::Tex2DArrayS32FloatGrad:
+    Opc = NVPTX::TEX_2D_ARRAY_S32_F32_GRAD;
     break;
-  case NVPTXISD::Tex2DArrayI32Float:
-    Opc = NVPTX::TEX_2D_ARRAY_I32_F32;
+  case NVPTXISD::Tex2DArrayU32S32:
+    Opc = NVPTX::TEX_2D_ARRAY_U32_S32;
     break;
-  case NVPTXISD::Tex2DArrayI32FloatLevel:
-    Opc = NVPTX::TEX_2D_ARRAY_I32_F32_LEVEL;
+  case NVPTXISD::Tex2DArrayU32Float:
+    Opc = NVPTX::TEX_2D_ARRAY_U32_F32;
     break;
-  case NVPTXISD::Tex2DArrayI32FloatGrad:
-    Opc = NVPTX::TEX_2D_ARRAY_I32_F32_GRAD;
+  case NVPTXISD::Tex2DArrayU32FloatLevel:
+    Opc = NVPTX::TEX_2D_ARRAY_U32_F32_LEVEL;
     break;
-  case NVPTXISD::Tex3DFloatI32:
-    Opc = NVPTX::TEX_3D_F32_I32;
+  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;
@@ -2595,25 +3162,358 @@ 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::Tex3DI32Float:
-    Opc = NVPTX::TEX_3D_I32_F32;
+  case NVPTXISD::TexUnified1DU32S32:
+    Opc = NVPTX::TEX_UNIFIED_1D_U32_S32;
     break;
-  case NVPTXISD::Tex3DI32FloatLevel:
-    Opc = NVPTX::TEX_3D_I32_F32_LEVEL;
+  case NVPTXISD::TexUnified1DU32Float:
+    Opc = NVPTX::TEX_UNIFIED_1D_U32_F32;
     break;
-  case NVPTXISD::Tex3DI32FloatGrad:
-    Opc = NVPTX::TEX_3D_I32_F32_GRAD;
+  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::TexUnified2DArrayS32FloatLevel:
+    Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_S32_F32_LEVEL;
+    break;
+  case NVPTXISD::TexUnified2DArrayS32FloatGrad:
+    Opc = NVPTX::TEX_UNIFIED_2D_ARRAY_S32_F32_GRAD;
+    break;
+  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));
   }
 
@@ -2630,174 +3530,610 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) {
   SmallVector<SDValue, 8> Ops;
   switch (N->getOpcode()) {
   default: return nullptr;
-  case NVPTXISD::Suld1DI8Trap:
-    Opc = NVPTX::SULD_1D_I8_TRAP;
+  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::Suld1DI16Trap:
-    Opc = NVPTX::SULD_1D_I16_TRAP;
+  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::Suld1DI32Trap:
-    Opc = NVPTX::SULD_1D_I32_TRAP;
+  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::Suld1DV2I8Trap:
-    Opc = NVPTX::SULD_1D_V2I8_TRAP;
+  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::Suld1DV2I16Trap:
-    Opc = NVPTX::SULD_1D_V2I16_TRAP;
+  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::Suld1DV2I32Trap:
-    Opc = NVPTX::SULD_1D_V2I32_TRAP;
+  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::Suld1DV4I8Trap:
-    Opc = NVPTX::SULD_1D_V4I8_TRAP;
+  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::Suld1DV4I16Trap:
-    Opc = NVPTX::SULD_1D_V4I16_TRAP;
+  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::Suld1DV4I32Trap:
-    Opc = NVPTX::SULD_1D_V4I32_TRAP;
+  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::Suld1DArrayI8Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_I8_TRAP;
+  case NVPTXISD::Suld1DV4I16Clamp:
+    Opc = NVPTX::SULD_1D_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::Suld1DArrayI16Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_I16_TRAP;
+  case NVPTXISD::Suld1DV4I32Clamp:
+    Opc = NVPTX::SULD_1D_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::Suld1DArrayI32Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_I32_TRAP;
+  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::Suld1DArrayV2I8Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_V2I8_TRAP;
+  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::Suld1DArrayV2I16Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_V2I16_TRAP;
+  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::Suld1DArrayV2I32Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_V2I32_TRAP;
+  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::Suld1DArrayV4I8Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_V4I8_TRAP;
+  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::Suld1DArrayV4I16Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_V4I16_TRAP;
+  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::Suld1DArrayV4I32Trap:
-    Opc = NVPTX::SULD_1D_ARRAY_V4I32_TRAP;
+  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::Suld2DI8Trap:
-    Opc = NVPTX::SULD_2D_I8_TRAP;
+  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::Suld2DI16Trap:
-    Opc = NVPTX::SULD_2D_I16_TRAP;
+  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::Suld2DI32Trap:
-    Opc = NVPTX::SULD_2D_I32_TRAP;
+  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::Suld2DV2I8Trap:
-    Opc = NVPTX::SULD_2D_V2I8_TRAP;
+  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::Suld2DV2I16Trap:
-    Opc = NVPTX::SULD_2D_V2I16_TRAP;
+  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::Suld2DV2I32Trap:
-    Opc = NVPTX::SULD_2D_V2I32_TRAP;
+  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::Suld2DV4I8Trap:
-    Opc = NVPTX::SULD_2D_V4I8_TRAP;
+  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::Suld2DV4I16Trap:
-    Opc = NVPTX::SULD_2D_V4I16_TRAP;
+  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);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DI16Trap:
+    Opc = NVPTX::SULD_1D_I16_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DI32Trap:
+    Opc = NVPTX::SULD_1D_I32_TRAP;
+    Ops.push_back(TexHandle);
+    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);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DV2I16Trap:
+    Opc = NVPTX::SULD_1D_V2I16_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DV2I32Trap:
+    Opc = NVPTX::SULD_1D_V2I32_TRAP;
+    Ops.push_back(TexHandle);
+    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);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DV4I16Trap:
+    Opc = NVPTX::SULD_1D_V4I16_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DV4I32Trap:
+    Opc = NVPTX::SULD_1D_V4I32_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DArrayI8Trap:
+    Opc = NVPTX::SULD_1D_ARRAY_I8_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DArrayI16Trap:
+    Opc = NVPTX::SULD_1D_ARRAY_I16_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DArrayI32Trap:
+    Opc = NVPTX::SULD_1D_ARRAY_I32_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    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);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DArrayV2I16Trap:
+    Opc = NVPTX::SULD_1D_ARRAY_V2I16_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DArrayV2I32Trap:
+    Opc = NVPTX::SULD_1D_ARRAY_V2I32_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    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);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DArrayV4I16Trap:
+    Opc = NVPTX::SULD_1D_ARRAY_V4I16_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld1DArrayV4I32Trap:
+    Opc = NVPTX::SULD_1D_ARRAY_V4I32_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld2DI8Trap:
+    Opc = NVPTX::SULD_2D_I8_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld2DI16Trap:
+    Opc = NVPTX::SULD_2D_I16_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld2DI32Trap:
+    Opc = NVPTX::SULD_2D_I32_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    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);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld2DV2I16Trap:
+    Opc = NVPTX::SULD_2D_V2I16_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld2DV2I32Trap:
+    Opc = NVPTX::SULD_2D_V2I32_TRAP;
+    Ops.push_back(TexHandle);
+    Ops.push_back(N->getOperand(2));
+    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);
+    Ops.push_back(N->getOperand(2));
+    Ops.push_back(N->getOperand(3));
+    Ops.push_back(Chain);
+    break;
+  case NVPTXISD::Suld2DV4I16Trap:
+    Opc = NVPTX::SULD_2D_V4I16_TRAP;
     Ops.push_back(TexHandle);
     Ops.push_back(N->getOperand(2));
     Ops.push_back(N->getOperand(3));
@@ -2810,144 +4146,572 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) {
     Ops.push_back(N->getOperand(3));
     Ops.push_back(Chain);
     break;
-  case NVPTXISD::Suld2DArrayI8Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_I8_TRAP;
+  case NVPTXISD::Suld2DArrayI8Trap:
+    Opc = NVPTX::SULD_2D_ARRAY_I8_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::Suld2DArrayI16Trap:
+    Opc = NVPTX::SULD_2D_ARRAY_I16_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::Suld2DArrayI32Trap:
+    Opc = NVPTX::SULD_2D_ARRAY_I32_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::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);
+    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::Suld2DArrayV2I16Trap:
+    Opc = NVPTX::SULD_2D_ARRAY_V2I16_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::Suld2DArrayV2I32Trap:
+    Opc = NVPTX::SULD_2D_ARRAY_V2I32_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::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);
+    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::Suld2DArrayV4I16Trap:
+    Opc = NVPTX::SULD_2D_ARRAY_V4I16_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::Suld2DArrayV4I32Trap:
+    Opc = NVPTX::SULD_2D_ARRAY_V4I32_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::Suld3DI8Trap:
+    Opc = NVPTX::SULD_3D_I8_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::Suld3DI16Trap:
+    Opc = NVPTX::SULD_3D_I16_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::Suld3DI32Trap:
+    Opc = NVPTX::SULD_3D_I32_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::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);
+    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::Suld3DV2I16Trap:
+    Opc = NVPTX::SULD_3D_V2I16_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::Suld3DV2I32Trap:
+    Opc = NVPTX::SULD_3D_V2I32_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::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);
+    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::Suld3DV4I16Trap:
+    Opc = NVPTX::SULD_3D_V4I16_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::Suld3DV4I32Trap:
+    Opc = NVPTX::SULD_3D_V4I32_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::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::Suld2DArrayI16Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_I16_TRAP;
+  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::Suld2DArrayI32Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_I32_TRAP;
+  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::Suld2DArrayV2I8Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_V2I8_TRAP;
+  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::Suld2DArrayV2I16Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_V2I16_TRAP;
+  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::Suld2DArrayV2I32Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_V2I32_TRAP;
+  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::Suld2DArrayV4I8Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_V4I8_TRAP;
+  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::Suld2DArrayV4I16Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_V4I16_TRAP;
+  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::Suld2DArrayV4I32Trap:
-    Opc = NVPTX::SULD_2D_ARRAY_V4I32_TRAP;
+  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::Suld3DI8Trap:
-    Opc = NVPTX::SULD_3D_I8_TRAP;
+  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::Suld3DI16Trap:
-    Opc = NVPTX::SULD_3D_I16_TRAP;
+  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::Suld3DI32Trap:
-    Opc = NVPTX::SULD_3D_I32_TRAP;
+  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::Suld3DV2I8Trap:
-    Opc = NVPTX::SULD_3D_V2I8_TRAP;
+  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::Suld3DV2I16Trap:
-    Opc = NVPTX::SULD_3D_V2I16_TRAP;
+  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::Suld3DV2I32Trap:
-    Opc = NVPTX::SULD_3D_V2I32_TRAP;
+  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::Suld3DV4I8Trap:
-    Opc = NVPTX::SULD_3D_V4I8_TRAP;
+  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::Suld3DV4I16Trap:
-    Opc = NVPTX::SULD_3D_V4I16_TRAP;
+  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::Suld3DV4I32Trap:
-    Opc = NVPTX::SULD_3D_V4I32_TRAP;
+  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));
@@ -2959,6 +4723,215 @@ SDNode *NVPTXDAGToDAGISel::SelectSurfaceIntrinsic(SDNode *N) {
   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) {
+  SDValue LHS = N->getOperand(0);
+  SDValue RHS = N->getOperand(1);
+  SDValue Len;
+  SDValue Start;
+  SDValue Val;
+  bool IsSigned = false;
+
+  if (N->getOpcode() == ISD::AND) {
+    // Canonicalize the operands
+    // We want 'and %val, %mask'
+    if (isa<ConstantSDNode>(LHS) && !isa<ConstantSDNode>(RHS)) {
+      std::swap(LHS, RHS);
+    }
+
+    ConstantSDNode *Mask = dyn_cast<ConstantSDNode>(RHS);
+    if (!Mask) {
+      // We need a constant mask on the RHS of the AND
+      return NULL;
+    }
+
+    // Extract the mask bits
+    uint64_t MaskVal = Mask->getZExtValue();
+    if (!isMask_64(MaskVal)) {
+      // We *could* handle shifted masks here, but doing so would require an
+      // 'and' operation to fix up the low-order bits so we would trade
+      // shr+and for bfe+and, which has the same throughput
+      return NULL;
+    }
+
+    // How many bits are in our mask?
+    uint64_t NumBits = countTrailingOnes(MaskVal);
+    Len = CurDAG->getTargetConstant(NumBits, MVT::i32);
+
+    if (LHS.getOpcode() == ISD::SRL || LHS.getOpcode() == ISD::SRA) {
+      // We have a 'srl/and' pair, extract the effective start bit and length
+      Val = LHS.getNode()->getOperand(0);
+      Start = LHS.getNode()->getOperand(1);
+      ConstantSDNode *StartConst = dyn_cast<ConstantSDNode>(Start);
+      if (StartConst) {
+        uint64_t StartVal = StartConst->getZExtValue();
+        // How many "good" bits do we have left?  "good" is defined here as bits
+        // that exist in the original value, not shifted in.
+        uint64_t GoodBits = Start.getValueType().getSizeInBits() - StartVal;
+        if (NumBits > GoodBits) {
+          // Do not handle the case where bits have been shifted in. In theory
+          // we could handle this, but the cost is likely higher than just
+          // emitting the srl/and pair.
+          return NULL;
+        }
+        Start = CurDAG->getTargetConstant(StartVal, MVT::i32);
+      } else {
+        // Do not handle the case where the shift amount (can be zero if no srl
+        // was found) is not constant. We could handle this case, but it would
+        // require run-time logic that would be more expensive than just
+        // emitting the srl/and pair.
+        return NULL;
+      }
+    } else {
+      // Do not handle the case where the LHS of the and is not a shift. While
+      // it would be trivial to handle this case, it would just transform
+      // 'and' -> 'bfe', but 'and' has higher-throughput.
+      return NULL;
+    }
+  } else if (N->getOpcode() == ISD::SRL || N->getOpcode() == ISD::SRA) {
+    if (LHS->getOpcode() == ISD::AND) {
+      ConstantSDNode *ShiftCnst = dyn_cast<ConstantSDNode>(RHS);
+      if (!ShiftCnst) {
+        // Shift amount must be constant
+        return NULL;
+      }
+
+      uint64_t ShiftAmt = ShiftCnst->getZExtValue();
+
+      SDValue AndLHS = LHS->getOperand(0);
+      SDValue AndRHS = LHS->getOperand(1);
+
+      // Canonicalize the AND to have the mask on the RHS
+      if (isa<ConstantSDNode>(AndLHS)) {
+        std::swap(AndLHS, AndRHS);
+      }
+
+      ConstantSDNode *MaskCnst = dyn_cast<ConstantSDNode>(AndRHS);
+      if (!MaskCnst) {
+        // Mask must be constant
+        return NULL;
+      }
+
+      uint64_t MaskVal = MaskCnst->getZExtValue();
+      uint64_t NumZeros;
+      uint64_t NumBits;
+      if (isMask_64(MaskVal)) {
+        NumZeros = 0;
+        // The number of bits in the result bitfield will be the number of
+        // trailing ones (the AND) minus the number of bits we shift off
+        NumBits = countTrailingOnes(MaskVal) - ShiftAmt;
+      } else if (isShiftedMask_64(MaskVal)) {
+        NumZeros = countTrailingZeros(MaskVal);
+        unsigned NumOnes = countTrailingOnes(MaskVal >> NumZeros);
+        // The number of bits in the result bitfield will be the number of
+        // trailing zeros plus the number of set bits in the mask minus the
+        // number of bits we shift off
+        NumBits = NumZeros + NumOnes - ShiftAmt;
+      } else {
+        // This is not a mask we can handle
+        return NULL;
+      }
+
+      if (ShiftAmt < NumZeros) {
+        // Handling this case would require extra logic that would make this
+        // transformation non-profitable
+        return NULL;
+      }
+
+      Val = AndLHS;
+      Start = CurDAG->getTargetConstant(ShiftAmt, MVT::i32);
+      Len = CurDAG->getTargetConstant(NumBits, MVT::i32);
+    } else if (LHS->getOpcode() == ISD::SHL) {
+      // Here, we have a pattern like:
+      //
+      // (sra (shl val, NN), MM)
+      // or
+      // (srl (shl val, NN), MM)
+      //
+      // If MM >= NN, we can efficiently optimize this with bfe
+      Val = LHS->getOperand(0);
+
+      SDValue ShlRHS = LHS->getOperand(1);
+      ConstantSDNode *ShlCnst = dyn_cast<ConstantSDNode>(ShlRHS);
+      if (!ShlCnst) {
+        // Shift amount must be constant
+        return NULL;
+      }
+      uint64_t InnerShiftAmt = ShlCnst->getZExtValue();
+
+      SDValue ShrRHS = RHS;
+      ConstantSDNode *ShrCnst = dyn_cast<ConstantSDNode>(ShrRHS);
+      if (!ShrCnst) {
+        // Shift amount must be constant
+        return NULL;
+      }
+      uint64_t OuterShiftAmt = ShrCnst->getZExtValue();
+
+      // To avoid extra codegen and be profitable, we need Outer >= Inner
+      if (OuterShiftAmt < InnerShiftAmt) {
+        return NULL;
+      }
+
+      // If the outer shift is more than the type size, we have no bitfield to
+      // extract (since we also check that the inner shift is <= the outer shift
+      // then this also implies that the inner shift is < the type size)
+      if (OuterShiftAmt >= Val.getValueType().getSizeInBits()) {
+        return NULL;
+      }
+
+      Start =
+        CurDAG->getTargetConstant(OuterShiftAmt - InnerShiftAmt, MVT::i32);
+      Len =
+        CurDAG->getTargetConstant(Val.getValueType().getSizeInBits() -
+                                  OuterShiftAmt, MVT::i32);
+
+      if (N->getOpcode() == ISD::SRA) {
+        // If we have a arithmetic right shift, we need to use the signed bfe
+        // variant
+        IsSigned = true;
+      }
+    } else {
+      // No can do...
+      return NULL;
+    }
+  } else {
+    // No can do...
+    return NULL;
+  }
+
+
+  unsigned Opc;
+  // For the BFE operations we form here from "and" and "srl", always use the
+  // unsigned variants.
+  if (Val.getValueType() == MVT::i32) {
+    if (IsSigned) {
+      Opc = NVPTX::BFE_S32rii;
+    } else {
+      Opc = NVPTX::BFE_U32rii;
+    }
+  } else if (Val.getValueType() == MVT::i64) {
+    if (IsSigned) {
+      Opc = NVPTX::BFE_S64rii;
+    } else {
+      Opc = NVPTX::BFE_U64rii;
+    }
+  } else {
+    // We cannot handle this type
+    return NULL;
+  }
+
+  SDValue Ops[] = {
+    Val, Start, Len
+  };
+
+  SDNode *Ret =
+    CurDAG->getMachineNode(Opc, SDLoc(N), N->getVTList(), Ops);
+
+  return Ret;
+}
+
 // SelectDirectAddr - Match a direct address for DAG.
 // A direct address could be a globaladdress or externalsymbol.
 bool NVPTXDAGToDAGISel::SelectDirectAddr(SDValue N, SDValue &Address) {
@@ -3053,17 +5026,10 @@ 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;
@@ -3075,12 +5041,12 @@ bool NVPTXDAGToDAGISel::ChkMemSDNodeAddressSpace(SDNode *N,
 /// SelectInlineAsmMemoryOperand - Implement addressing mode selection for
 /// inline asm expressions.
 bool NVPTXDAGToDAGISel::SelectInlineAsmMemoryOperand(
-    const SDValue &Op, char ConstraintCode, std::vector<SDValue> &OutOps) {
+    const SDValue &Op, unsigned ConstraintID, std::vector<SDValue> &OutOps) {
   SDValue Op0, Op1;
-  switch (ConstraintCode) {
+  switch (ConstraintID) {
   default:
     return true;
-  case 'm': // memory
+  case InlineAsm::Constraint_m: // memory
     if (SelectDirectAddr(Op, Op0)) {
       OutOps.push_back(Op0);
       OutOps.push_back(CurDAG->getTargetConstant(0, MVT::i32));