SILoadStoreOptimizer: add support for GFX10 image instructions

[foad]

Summary

GFX10 image instructions use one or more address operands starting at
vaddr0, instead of a single vaddr operand, to allow for NSA forms.

[foad]

This isn't ready to push yet but I'd like to get some feedback on it -- see comments inline.

llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp

116

Is this really true? I see opcodes like IMAGE_SAMPLE_D_CL_O_V4_V16 mentioned in some generated tables, but I'm not sure what the V16 there really means or whether these will ever occur in practice.

llvm/test/CodeGen/AMDGPU/merge-image-load.mir

2 ↗

(On Diff #270187)

These tests pass with -mcpu=gfx1010 but I don't think they're really testing anything useful. Do I need to regenerate them using gfx10-specific image sample opcodes or something?

[rampitec]

llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp

116

I think that's because we do not have all register classes, so it uses VReg_512 for 10, 11, or 12 dword address for example. I believe a longest one has 12 vaddrs, like this:

image_sample_c_d_cl_o v[16:19], [v8, v9, v10, v11, v12, v13, v14, v15, v16, v17, v18, v19], s[20:27], s[100:103] dmask:0xf dim:SQ_RSRC_IMG_3D ; encoding: [0x16,0x0f,0xec,0xf0,0x08,0x10,0x25,0x03,0x09,0x0a,0x0b,0x0c,0x0d,0x0e,0x0f,0x10,0x11,0x12,0x13,0x00]
0x16,0x0f,0xec,0xf0,0x08,0x10,0x25,0x03,0x09,0x0a,0x0b,0x0c,0x0d,0x0e,0x0f,0x10,0x11,0x12,0x13,0x14

You can check it in MIMGInstructions.td where MIMG_Sampler_AddrSizes is defined.

[arsenm]

llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp

116

I think 12 is what I remember was the maximum. I was confused by the missing register classes. We probably should add the missing ones now that globalisel at least would handle them (I had to go back and break the code to round to the instruction constraints)

[rampitec]

llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp

116

It will probably require not only register classes, but value types as well.

[nhaehnle]

Mostly LGTM, with some nitpicks.

llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp

108–113

false instead of 0 for bools.

But how about keeping RegisterEnum and making this two unsigned chars, one for the RegisterEnum flags and one for the NumVAddrs?

116

I remember the value types being the part where I got stuck when I tried to do this a very, very long time ago. And yes, 12 is the maximum number of vaddrs.

430–431

Remove the struct

[foad]

llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp

108–113

I think using separate fields is better than bit twiddling. The storage size of AddressRegs doesn't really matter because it isn't stored anywhere, only used to return info from getRegs. In retrospect I shouldn't have bothered squashing NumVAddrs into a single byte.

[foad]

Address review comments.

[piotr]

llvm/test/CodeGen/AMDGPU/merge-image-load.mir

2 ↗

(On Diff #270187)

Yes, it would be good if you could add tests for image instructions with NSA.

[foad]

Check dlc bit for gfx10 MIMG instructions. Add tests.

[piotr]

Excellent test coverage, thank you for doing that.

[piotr]

llvm/test/CodeGen/AMDGPU/merge-image-load-gfx10.mir

13

s/ =C/ = C/

[foad]

Fix formatting in MIR tests.

[foad]

Ping!

[nhaehnle]

LGTM

llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp

enum InstClassEnum {
   S_BUFFER_LOAD_IMM,
   BUFFER_LOAD,
   BUFFER_STORE,
   MIMG,
   TBUFFER_LOAD,
   TBUFFER_STORE,
 };
 
-enum RegisterEnum {
-  SBASE = 0x1,
-  SRSRC = 0x2,
-  SOFFSET = 0x4,
-  VADDR = 0x8,
-  ADDR = 0x10,
-  SSAMP = 0x20,
+struct AddressRegs {
+  unsigned char NumVAddrs = 0;
+  bool SBase = false;
+  bool SRsrc = false;
+  bool SOffset = false;
+  bool VAddr = false;
+  bool Addr = false;
+  bool SSamp = false;
 };
 
+// GFX10 image_sample instructions can have 12 vaddrs + srsrc + ssamp.
+const unsigned MaxAddressRegs = 12 + 1 + 1;
+
 class SILoadStoreOptimizer : public MachineFunctionPass {
   struct CombineInfo {
     MachineBasicBlock::iterator I;
     unsigned EltSize;
     unsigned Offset;
     unsigned Width;
     unsigned Format;
     unsigned BaseOff;
     unsigned DMask;
     InstClassEnum InstClass;
     bool GLC;
     bool SLC;
     bool DLC;
     bool UseST64;
-    int AddrIdx[5];
-    const MachineOperand *AddrReg[5];
+    int AddrIdx[MaxAddressRegs];
+    const MachineOperand *AddrReg[MaxAddressRegs];
     unsigned NumAddresses;
     unsigned Order;
 
     bool hasSameBaseAddress(const MachineInstr &MI) {
       for (unsigned i = 0; i < NumAddresses; i++) {
         const MachineOperand &AddrRegNext = MI.getOperand(AddrIdx[i]);
 
         if (AddrReg[i]->isImm() || AddrRegNext.isImm()) {
if (TII.isMUBUF(Opc)) {
       case AMDGPU::BUFFER_STORE_DWORD_OFFEN_exact:
       case AMDGPU::BUFFER_STORE_DWORD_OFFSET:
       case AMDGPU::BUFFER_STORE_DWORD_OFFSET_exact:
         return BUFFER_STORE;
       }
     }
     if (TII.isMIMG(Opc)) {
       // Ignore instructions encoded without vaddr.
-      if (AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::vaddr) == -1)
+      if (AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::vaddr) == -1 &&
+          AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::vaddr0) == -1)
         return UNKNOWN;
       // TODO: Support IMAGE_GET_RESINFO and IMAGE_GET_LOD.
       if (TII.get(Opc).mayStore() || !TII.get(Opc).mayLoad() ||
           TII.isGather4(Opc))
         return UNKNOWN;
       return MIMG;
     }
     if (TII.isMTBUF(Opc)) {
case AMDGPU::DS_WRITE_B64_gfx9:
     return Opc;
   case AMDGPU::S_BUFFER_LOAD_DWORD_IMM:
   case AMDGPU::S_BUFFER_LOAD_DWORDX2_IMM:
   case AMDGPU::S_BUFFER_LOAD_DWORDX4_IMM:
     return AMDGPU::S_BUFFER_LOAD_DWORD_IMM;
   }
 }
 
-static unsigned getRegs(unsigned Opc, const SIInstrInfo &TII) {
-  if (TII.isMUBUF(Opc)) {
-    unsigned result = 0;
+static AddressRegs getRegs(unsigned Opc, const SIInstrInfo &TII) {
+  AddressRegs Result;
 
-    if (AMDGPU::getMUBUFHasVAddr(Opc)) {
-      result |= VADDR;
-    }
-
-    if (AMDGPU::getMUBUFHasSrsrc(Opc)) {
-      result |= SRSRC;
-    }
-
-    if (AMDGPU::getMUBUFHasSoffset(Opc)) {
-      result |= SOFFSET;
-    }
+  if (TII.isMUBUF(Opc)) {
+    if (AMDGPU::getMUBUFHasVAddr(Opc))
+      Result.VAddr = true;
+    if (AMDGPU::getMUBUFHasSrsrc(Opc))
+      Result.SRsrc = true;
+    if (AMDGPU::getMUBUFHasSoffset(Opc))
+      Result.SOffset = true;
 
-    return result;
+    return Result;
   }
 
   if (TII.isMIMG(Opc)) {
-    unsigned result = VADDR | SRSRC;
+    int VAddr0Idx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::vaddr0);
+    if (VAddr0Idx >= 0) {
+      int SRsrcIdx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::srsrc);
+      Result.NumVAddrs = SRsrcIdx - VAddr0Idx;
+    } else {
+      Result.VAddr = true;
+    }
+    Result.SRsrc = true;
     const AMDGPU::MIMGInfo *Info = AMDGPU::getMIMGInfo(Opc);
     if (Info && AMDGPU::getMIMGBaseOpcodeInfo(Info->BaseOpcode)->Sampler)
-      result |= SSAMP;
+      Result.SSamp = true;
 
-    return result;
+    return Result;
   }
   if (TII.isMTBUF(Opc)) {
-    unsigned result = 0;
-
-    if (AMDGPU::getMTBUFHasVAddr(Opc)) {
-      result |= VADDR;
-    }
-
-    if (AMDGPU::getMTBUFHasSrsrc(Opc)) {
-      result |= SRSRC;
-    }
-
-    if (AMDGPU::getMTBUFHasSoffset(Opc)) {
-      result |= SOFFSET;
-    }
+    if (AMDGPU::getMTBUFHasVAddr(Opc))
+      Result.VAddr = true;
+    if (AMDGPU::getMTBUFHasSrsrc(Opc))
+      Result.SRsrc = true;
+    if (AMDGPU::getMTBUFHasSoffset(Opc))
+      Result.SOffset = true;
 
-    return result;
+    return Result;
   }
 
   switch (Opc) {
   default:
-    return 0;
+    return Result;
   case AMDGPU::S_BUFFER_LOAD_DWORD_IMM:
   case AMDGPU::S_BUFFER_LOAD_DWORDX2_IMM:
   case AMDGPU::S_BUFFER_LOAD_DWORDX4_IMM:
-    return SBASE;
+    Result.SBase = true;
+    return Result;
   case AMDGPU::DS_READ_B32:
   case AMDGPU::DS_READ_B64:
   case AMDGPU::DS_READ_B32_gfx9:
   case AMDGPU::DS_READ_B64_gfx9:
   case AMDGPU::DS_WRITE_B32:
   case AMDGPU::DS_WRITE_B64:
   case AMDGPU::DS_WRITE_B32_gfx9:
   case AMDGPU::DS_WRITE_B64_gfx9:
-    return ADDR;
+    Result.Addr = true;
+    return Result;
   }
 }
 
 void SILoadStoreOptimizer::CombineInfo::setMI(MachineBasicBlock::iterator MI,
                                               const SIInstrInfo &TII,
                                               const GCNSubtarget &STM) {
   I = MI;
   unsigned Opc = MI->getOpcode();
void SILoadStoreOptimizer::CombineInfo::setMI(MachineBasicBlock::iterator MI,
   } else if (InstClass != MIMG) {
     GLC = TII.getNamedOperand(*I, AMDGPU::OpName::glc)->getImm();
     if (InstClass != S_BUFFER_LOAD_IMM) {
       SLC = TII.getNamedOperand(*I, AMDGPU::OpName::slc)->getImm();
     }
     DLC = TII.getNamedOperand(*I, AMDGPU::OpName::dlc)->getImm();
   }
 
-  unsigned AddrOpName[5] = {0};
-  NumAddresses = 0;
-  const unsigned Regs = getRegs(I->getOpcode(), TII);
-
-  if (Regs & ADDR) {
-    AddrOpName[NumAddresses++] = AMDGPU::OpName::addr;
-  }
-
-  if (Regs & SBASE) {
-    AddrOpName[NumAddresses++] = AMDGPU::OpName::sbase;
-  }
-
-  if (Regs & SRSRC) {
-    AddrOpName[NumAddresses++] = AMDGPU::OpName::srsrc;
-  }
-
-  if (Regs & SOFFSET) {
-    AddrOpName[NumAddresses++] = AMDGPU::OpName::soffset;
-  }
+  AddressRegs Regs = getRegs(Opc, TII);
 
-  if (Regs & VADDR) {
-    AddrOpName[NumAddresses++] = AMDGPU::OpName::vaddr;
-  }
-
-  if (Regs & SSAMP) {
-    AddrOpName[NumAddresses++] = AMDGPU::OpName::ssamp;
-  }
+  NumAddresses = 0;
+  for (unsigned J = 0; J < Regs.NumVAddrs; J++)
+    AddrIdx[NumAddresses++] =
+        AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::vaddr0) + J;
+  if (Regs.Addr)
+    AddrIdx[NumAddresses++] =
+        AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::addr);
+  if (Regs.SBase)
+    AddrIdx[NumAddresses++] =
+        AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::sbase);
+  if (Regs.SRsrc)
+    AddrIdx[NumAddresses++] =
+        AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::srsrc);
+  if (Regs.SOffset)
+    AddrIdx[NumAddresses++] =
+        AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::soffset);
+  if (Regs.VAddr)
+    AddrIdx[NumAddresses++] =
+        AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::vaddr);
+  if (Regs.SSamp)
+    AddrIdx[NumAddresses++] =
+        AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::ssamp);
+  assert(NumAddresses <= MaxAddressRegs);
 
-  for (unsigned i = 0; i < NumAddresses; i++) {
-    AddrIdx[i] = AMDGPU::getNamedOperandIdx(I->getOpcode(), AddrOpName[i]);
-    AddrReg[i] = &I->getOperand(AddrIdx[i]);
-  }
+  for (unsigned J = 0; J < NumAddresses; J++)
+    AddrReg[J] = &I->getOperand(AddrIdx[J]);
 }
 
 } // end anonymous namespace.
 
 INITIALIZE_PASS_BEGIN(SILoadStoreOptimizer, DEBUG_TYPE,
                       "SI Load Store Optimizer", false, false)
 INITIALIZE_PASS_DEPENDENCY(AAResultsWrapperPass)
 INITIALIZE_PASS_END(SILoadStoreOptimizer, DEBUG_TYPE, "SI Load Store Optimizer",
bool SILoadStoreOptimizer::dmasksCanBeCombined(const CombineInfo &CI,
 
   if ((TFEOp && TFEOp->getImm()) || (LWEOp && LWEOp->getImm()))
     return false;
 
   // Check other optional immediate operands for equality.
   unsigned OperandsToMatch[] = {AMDGPU::OpName::glc, AMDGPU::OpName::slc,
                                 AMDGPU::OpName::d16, AMDGPU::OpName::unorm,
                                 AMDGPU::OpName::da,  AMDGPU::OpName::r128,
-                                AMDGPU::OpName::a16};
+                                AMDGPU::OpName::a16, AMDGPU::OpName::dlc};
 
   for (auto op : OperandsToMatch) {
     int Idx = AMDGPU::getNamedOperandIdx(CI.I->getOpcode(), op);
     if (AMDGPU::getNamedOperandIdx(Paired.I->getOpcode(), op) != Idx)
       return false;
     if (Idx != -1 &&
         CI.I->getOperand(Idx).getImm() != Paired.I->getOperand(Idx).getImm())
       return false;
MachineBasicBlock::iterator SILoadStoreOptimizer::mergeBufferLoadPair(
   const TargetRegisterClass *SuperRC = getTargetRegisterClass(CI, Paired);
 
   // Copy to the new source register.
   Register DestReg = MRI->createVirtualRegister(SuperRC);
   unsigned MergedOffset = std::min(CI.Offset, Paired.Offset);
 
   auto MIB = BuildMI(*MBB, Paired.I, DL, TII->get(Opcode), DestReg);
 
-  const unsigned Regs = getRegs(Opcode, *TII);
+  AddressRegs Regs = getRegs(Opcode, *TII);
 
-  if (Regs & VADDR)
+  if (Regs.VAddr)
     MIB.add(*TII->getNamedOperand(*CI.I, AMDGPU::OpName::vaddr));
 
   // It shouldn't be possible to get this far if the two instructions
   // don't have a single memoperand, because MachineInstr::mayAlias()
   // will return true if this is the case.
   assert(CI.I->hasOneMemOperand() && Paired.I->hasOneMemOperand());
 
   const MachineMemOperand *MMOa = *CI.I->memoperands_begin();
MachineBasicBlock::iterator SILoadStoreOptimizer::mergeTBufferLoadPair(
   const TargetRegisterClass *SuperRC = getTargetRegisterClass(CI, Paired);
 
   // Copy to the new source register.
   Register DestReg = MRI->createVirtualRegister(SuperRC);
   unsigned MergedOffset = std::min(CI.Offset, Paired.Offset);
 
   auto MIB = BuildMI(*MBB, Paired.I, DL, TII->get(Opcode), DestReg);
 
-  const unsigned Regs = getRegs(Opcode, *TII);
+  AddressRegs Regs = getRegs(Opcode, *TII);
 
-  if (Regs & VADDR)
+  if (Regs.VAddr)
     MIB.add(*TII->getNamedOperand(*CI.I, AMDGPU::OpName::vaddr));
 
   unsigned JoinedFormat =
       getBufferFormatWithCompCount(CI.Format, CI.Width + Paired.Width, *STM);
 
   // It shouldn't be possible to get this far if the two instructions
   // don't have a single memoperand, because MachineInstr::mayAlias()
   // will return true if this is the case.
BuildMI(*MBB, Paired.I, DL, TII->get(AMDGPU::REG_SEQUENCE), SrcReg)
       .add(*Src0)
       .addImm(SubRegIdx0)
       .add(*Src1)
       .addImm(SubRegIdx1);
 
   auto MIB = BuildMI(*MBB, Paired.I, DL, TII->get(Opcode))
                  .addReg(SrcReg, RegState::Kill);
 
-  const unsigned Regs = getRegs(Opcode, *TII);
+  AddressRegs Regs = getRegs(Opcode, *TII);
 
-  if (Regs & VADDR)
+  if (Regs.VAddr)
     MIB.add(*TII->getNamedOperand(*CI.I, AMDGPU::OpName::vaddr));
 
   unsigned JoinedFormat =
       getBufferFormatWithCompCount(CI.Format, CI.Width + Paired.Width, *STM);
 
   // It shouldn't be possible to get this far if the two instructions
   // don't have a single memoperand, because MachineInstr::mayAlias()
   // will return true if this is the case.
BuildMI(*MBB, Paired.I, DL, TII->get(AMDGPU::REG_SEQUENCE), SrcReg)
       .add(*Src0)
       .addImm(SubRegIdx0)
       .add(*Src1)
       .addImm(SubRegIdx1);
 
   auto MIB = BuildMI(*MBB, Paired.I, DL, TII->get(Opcode))
                  .addReg(SrcReg, RegState::Kill);
 
-  const unsigned Regs = getRegs(Opcode, *TII);
+  AddressRegs Regs = getRegs(Opcode, *TII);
 
-  if (Regs & VADDR)
+  if (Regs.VAddr)
     MIB.add(*TII->getNamedOperand(*CI.I, AMDGPU::OpName::vaddr));
 
 
   // It shouldn't be possible to get this far if the two instructions
   // don't have a single memoperand, because MachineInstr::mayAlias()
   // will return true if this is the case.
   assert(CI.I->hasOneMemOperand() && Paired.I->hasOneMemOperand());
 
for (; BlockI != End; ++BlockI) {
 
     CombineInfo CI;
     CI.setMI(MI, *TII, *STM);
     CI.Order = Order++;
 
     if (!CI.hasMergeableAddress(*MRI))
       continue;
 
+    LLVM_DEBUG(dbgs() << "Mergeable: " << MI);
+
     addInstToMergeableList(CI, MergeableInsts);
   }
 
   // At this point we have lists of Mergeable instructions.
   //
   // Part 2: Sort lists by offset and then for each CombineInfo object in the
   // list try to find an instruction that can be merged with I.  If an instruction
   // is found, it is stored in the Paired field.  If no instructions are found, then
for (auto I = MergeList.begin(), Next = std::next(I); Next != MergeList.end();
     SmallVector<MachineInstr *, 8> InstsToMove;
     if (!checkAndPrepareMerge(CI, Paired, InstsToMove)) {
       ++I;
       continue;
     }
 
     Modified = true;
 
+    LLVM_DEBUG(dbgs() << "Merging: " << *CI.I << "   with: " << *Paired.I);
+
     switch (CI.InstClass) {
     default:
       llvm_unreachable("unknown InstClass");
       break;
     case DS_READ: {
       MachineBasicBlock::iterator NewMI =
           mergeRead2Pair(CI, Paired, InstsToMove);
       CI.setMI(NewMI, *TII, *STM);

llvm/test/CodeGen/AMDGPU/merge-image-load-gfx10.mir

+# RUN: llc -march=amdgcn -mcpu=gfx1010 -verify-machineinstrs -run-pass si-load-store-opt -o - %s | FileCheck -check-prefix=GFX10 %s
+
+# GFX10-LABEL: name: image_load_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_load_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+# GFX10-LABEL: name: image_load_merged_v1v3_reversed
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub3
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub0_sub1_sub2
+
+name: image_load_merged_v1v3_reversed
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_merged_v2v2
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_64 = COPY %8.sub0_sub1
+# GFX10: %{{[0-9]+}}:vreg_64 = COPY killed %8.sub2_sub3
+
+name: image_load_merged_v2v2
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = IMAGE_LOAD_V2_V2_gfx10 %5:vreg_64, %3:sgpr_256, 3, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+    %7:vreg_64 = IMAGE_LOAD_V2_V2_gfx10 %5:vreg_64, %3:sgpr_256, 12, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_merged_v2v2_reversed
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_64 = COPY %8.sub2_sub3
+# GFX10: %{{[0-9]+}}:vreg_64 = COPY killed %8.sub0_sub1
+
+name: image_load_merged_v2v2_reversed
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = IMAGE_LOAD_V2_V2_gfx10 %5:vreg_64, %3:sgpr_256, 12, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+    %7:vreg_64 = IMAGE_LOAD_V2_V2_gfx10 %5:vreg_64, %3:sgpr_256, 3, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_merged_v3v1
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY %8.sub0_sub1_sub2
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY killed %8.sub3
+
+name: image_load_merged_v3v1
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+    %7:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_merged_v3v1_reversed
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY %8.sub1_sub2_sub3
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY killed %8.sub0
+
+name: image_load_merged_v3v1_reversed
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+    %7:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_divided_merged
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+
+name: image_load_divided_merged
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %8:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %9:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %7:vreg_64, %3:sgpr_256, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+    %10:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %11:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_divided_not_merged
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_divided_not_merged
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vreg_128 = COPY %2
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    IMAGE_STORE_V4_V2 %4:vreg_128, %5:vreg_64, %3:sgpr_256, 15, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (store 16)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_dmask_overlapped_not_merged
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 4, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_dmask_overlapped_not_merged
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 4, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_dmask_not_disjoint_not_merged
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 4, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 11, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_dmask_not_disjoint_not_merged
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 4, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 11, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_0
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %6, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_0
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 1, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %7:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %8:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %6, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_1
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %6, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %6, %4, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_1
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %5:vgpr_32 = COPY %2.sub3
+    %6:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %7:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %6, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %8:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %6, %4, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_3
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_4
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_4
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_5
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 1, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_5
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 1, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_6
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 1, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_6
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 1, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_7
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 1, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_7
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 1, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_8
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V1_gfx10 %6, %3, 8, 1, -1, 0, 0, 0, 0, 1, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_8
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = COPY %5.sub0
+    %7:vgpr_32 = IMAGE_LOAD_V1_V1_gfx10 %6, %3, 8, 1, -1, 0, 0, 0, 0, 1, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+    %8:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_9
+# GFX10: %{{[0-9]+}}:vreg_64 = IMAGE_LOAD_V2_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 1, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_9
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = IMAGE_LOAD_V2_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 1, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_10
+# GFX10: %{{[0-9]+}}:vreg_64 = IMAGE_LOAD_V2_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 1, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_10
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = IMAGE_LOAD_V2_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 1, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_not_merged_11
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 1, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_load_not_merged_11
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_V1_V2_gfx10 %5, %3, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 1, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_V3_V2_gfx10 %5, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_mip_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_MIP_V4_V3_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_load_mip_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_96 = BUFFER_LOAD_DWORDX3_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_MIP_V1_V3_gfx10 %5:vreg_96, %3:sgpr_256, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_MIP_V3_V3_gfx10 %5:vreg_96, %3:sgpr_256, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+
+# GFX10-LABEL: name: image_load_mip_pck_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_MIP_PCK_V4_V3_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_load_mip_pck_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_96 = BUFFER_LOAD_DWORDX3_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_MIP_PCK_V1_V3_gfx10 %5:vreg_96, %3:sgpr_256, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_MIP_PCK_V3_V3_gfx10 %5:vreg_96, %3:sgpr_256, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+
+# GFX10-LABEL: name: image_load_mip_pck_sgn_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_MIP_PCK_SGN_V4_V3_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_load_mip_pck_sgn_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_96 = BUFFER_LOAD_DWORDX3_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_MIP_PCK_SGN_V1_V3_gfx10 %5:vreg_96, %3:sgpr_256, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_MIP_PCK_SGN_V3_V3_gfx10 %5:vreg_96, %3:sgpr_256, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_pck_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_PCK_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_load_pck_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_PCK_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_PCK_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_load_pck_sgn_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_PCK_SGN_V4_V2_gfx10 %5, %3, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_load_pck_sgn_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 =COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vreg_64 = BUFFER_LOAD_DWORDX2_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_LOAD_PCK_SGN_V1_V2_gfx10 %5:vreg_64, %3:sgpr_256, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_LOAD_PCK_SGN_V3_V2_gfx10 %5:vreg_64, %3:sgpr_256, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---

llvm/test/CodeGen/AMDGPU/merge-image-sample-gfx10.mir

+# RUN: llc -march=amdgcn -mcpu=gfx1010 -verify-machineinstrs -run-pass si-load-store-opt -o - %s | FileCheck -check-prefix=GFX10 %s
+
+# GFX10-LABEL: name: image_sample_l_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_L_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_l_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+# GFX10-LABEL: name: image_sample_l_merged_v1v3_reversed
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_L_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub3
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub0_sub1_sub2
+
+name: image_sample_l_merged_v1v3_reversed
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_merged_v2v2
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_L_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_64 = COPY %8.sub0_sub1
+# GFX10: %{{[0-9]+}}:vreg_64 = COPY killed %8.sub2_sub3
+
+name: image_sample_l_merged_v2v2
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = IMAGE_SAMPLE_L_V2_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 3, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+    %7:vreg_64 = IMAGE_SAMPLE_L_V2_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 12, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_merged_v2v2_reversed
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_L_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_64 = COPY %8.sub2_sub3
+# GFX10: %{{[0-9]+}}:vreg_64 = COPY killed %8.sub0_sub1
+
+name: image_sample_l_merged_v2v2_reversed
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = IMAGE_SAMPLE_L_V2_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 12, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+    %7:vreg_64 = IMAGE_SAMPLE_L_V2_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 3, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 8, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_merged_v3v1
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_L_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY %8.sub0_sub1_sub2
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY killed %8.sub3
+
+name: image_sample_l_merged_v3v1
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+    %7:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_merged_v3v1_reversed
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_L_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY %8.sub1_sub2_sub3
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY killed %8.sub0
+
+name: image_sample_l_merged_v3v1_reversed
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+    %7:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_divided_merged
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_L_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+
+name: image_sample_l_divided_merged
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %8:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %9:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %7:vgpr_32, %7:vgpr_32, %7:vgpr_32, %3:sgpr_256, %2:sgpr_128, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+    %10:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %11:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_divided_not_merged
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_divided_not_merged
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vreg_128 = COPY %2
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    IMAGE_STORE_V4_V2_nsa_gfx10 %4:vreg_128, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (store 16)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_dmask_overlapped_not_merged
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 4, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_dmask_overlapped_not_merged
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 4, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_dmask_not_disjoint_not_merged
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 4, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 11, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_dmask_not_disjoint_not_merged
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 4, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 11, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_0
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %6, %6, %6, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_0
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 1, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %7:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %8:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %6, %6, %6, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_1
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %6, %6, %6, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %6, %6, %6, %4, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_1
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %5:vgpr_32 = COPY %2.sub3
+    %6:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %7:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %6, %6, %6, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %8:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %6, %6, %6, %4, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_2
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %6, %6, %6, %4, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %6, %6, %6, %4, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_2
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_128 = COPY $sgpr92_sgpr93_sgpr94_sgpr95
+    %4:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %5:vgpr_32 = COPY %2.sub3
+    %6:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %7:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %6, %6, %6, %4, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %8:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %6, %6, %6, %4, %3, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_3
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_4
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 1, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_4
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 1, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_5
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 1, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_5
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 1, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_6
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 1, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_6
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 1, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_7
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V2_nsa_gfx10 %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 1, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_7
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V2_nsa_gfx10 %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 1, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_8
+# GFX10: %{{[0-9]+}}:vreg_64 = IMAGE_SAMPLE_L_V2_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 1, 0, 0, implicit $exec :: (dereferenceable load 8, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_8
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = IMAGE_SAMPLE_L_V2_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 1, 0, 0, implicit $exec :: (dereferenceable load 8, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_9
+# GFX10: %{{[0-9]+}}:vreg_64 = IMAGE_SAMPLE_L_V2_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 1, 0, implicit $exec :: (dereferenceable load 8, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_9
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vreg_64 = IMAGE_SAMPLE_L_V2_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 1, 0, implicit $exec :: (dereferenceable load 8, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+# GFX10-LABEL: name: image_sample_l_not_merged_10
+# GFX10: %{{[0-9]+}}:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 1, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+
+name: image_sample_l_not_merged_10
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_V1_V3_nsa_gfx10 %5, %5, %5, %3, %2, 8, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_V3_V3_nsa_gfx10 %5, %5, %5, %3, %2, 7, 1, -1, 0, 0, 0, 0, 0, 0, 0, 1, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+
+
+# GFX10-LABEL: name: image_sample_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_V4_V2_nsa_gfx10 %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_V1_V2_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_V3_V2_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_b_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_B_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_b_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_B_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_B_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_b_cl_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_B_CL_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_b_cl_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_B_CL_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_B_CL_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_b_cl_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_B_CL_O_V4_V5_nsa_gfx10 %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_b_cl_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_B_CL_O_V1_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_B_CL_O_V3_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_b_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_B_O_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_b_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_B_O_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_B_O_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_cd_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_CD_V4_V6_nsa_gfx10 %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_cd_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_CD_V1_V6_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_CD_V3_V6_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_cd_cl_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_CD_CL_V4_V7_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_cd_cl_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_CD_CL_V1_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_CD_CL_V3_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_cd_cl_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_CD_CL_O_V4_V8_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_cd_cl_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_CD_CL_O_V1_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_CD_CL_O_V3_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_cd_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_CD_O_V4_V7_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_cd_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_CD_O_V1_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_CD_O_V3_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_cl_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_CL_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_cl_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_CL_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_CL_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_cl_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_CL_O_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_cl_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_CL_O_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_CL_O_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_b_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_B_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_b_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_B_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_B_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_b_cl_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_B_CL_V4_V5_nsa_gfx10 %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_b_cl_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_B_CL_V1_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_B_CL_V3_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_b_cl_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_B_CL_O_V4_V6_nsa_gfx10 %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_b_cl_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_B_CL_O_V1_V6_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_B_CL_O_V3_V6_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_b_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_B_O_V4_V5_nsa_gfx10 %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_b_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_B_O_V1_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_B_O_V3_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_cd_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_CD_V4_V7_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_cd_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_CD_V1_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_CD_V3_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_cd_cl_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_CD_CL_V4_V8_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_cd_cl_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_CD_CL_V1_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_CD_CL_V3_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_cd_cl_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_CD_CL_O_V4_V9_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_cd_cl_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_CD_CL_O_V1_V9_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_CD_CL_O_V3_V9_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_cd_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_CD_O_V4_V8_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_cd_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_CD_O_V1_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_CD_O_V3_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_cl_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_CL_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_cl_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_CL_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_CL_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_cl_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_CL_O_V4_V5_nsa_gfx10 %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_cl_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_CL_O_V1_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_CL_O_V3_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_d_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_D_V4_V7_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_d_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_D_V1_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_D_V3_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_d_cl_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_D_CL_V4_V8_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_d_cl_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_D_CL_V1_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_D_CL_V3_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_d_cl_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_D_CL_O_V4_V9_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_d_cl_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_D_CL_O_V1_V9_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_D_CL_O_V3_V9_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_d_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_D_O_V4_V8_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_d_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_D_O_V1_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_D_O_V3_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_l_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_L_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_l_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_L_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_L_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_lz_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_LZ_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_lz_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_LZ_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_LZ_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_lz_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_LZ_O_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_lz_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_LZ_O_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_LZ_O_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_l_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_L_O_V4_V5_nsa_gfx10 %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_l_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_L_O_V1_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_L_O_V3_V5_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_c_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_C_O_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_c_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_C_O_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_C_O_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_d_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_D_V4_V6_nsa_gfx10 %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_d_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_D_V1_V6_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_D_V3_V6_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_d_cl_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_D_CL_V4_V7_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_d_cl_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_D_CL_V1_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_D_CL_V3_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_d_cl_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_D_CL_O_V4_V8_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_d_cl_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_D_CL_O_V1_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_D_CL_O_V3_V8_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_d_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_D_O_V4_V7_nsa_gfx10 %5, %5, %5, %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_d_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_D_O_V1_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_D_O_V3_V7_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_lz_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_LZ_V4_V2_nsa_gfx10 %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_lz_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_LZ_V1_V2_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_LZ_V3_V2_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_lz_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_LZ_O_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_lz_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_LZ_O_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_LZ_O_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_l_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_L_O_V4_V4_nsa_gfx10 %5, %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_l_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_L_O_V1_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_L_O_V3_V4_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---
+
+
+# GFX10-LABEL: name: image_sample_o_merged_v1v3
+# GFX10: %{{[0-9]+}}:vreg_128 = IMAGE_SAMPLE_O_V4_V3_nsa_gfx10 %5, %5, %5, %3, %2, 15, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec, implicit $exec :: (dereferenceable load 16, align 4, addrspace 4)
+# GFX10: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
+# GFX10: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
+
+name: image_sample_o_merged_v1v3
+body:             |
+  bb.0.entry:
+    %0:sgpr_64 = COPY $sgpr0_sgpr1
+    %1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0, 0
+    %2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
+    %3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0, 0
+    %4:vgpr_32 = COPY %2.sub3
+    %5:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET %2:sgpr_128, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16)
+    %6:vgpr_32 = IMAGE_SAMPLE_O_V1_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 1, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 4, addrspace 4)
+    %7:vreg_96 = IMAGE_SAMPLE_O_V3_V3_nsa_gfx10 %5:vgpr_32, %5:vgpr_32, %5:vgpr_32, %3:sgpr_256, %2:sgpr_128, 14, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load 12, align 16, addrspace 4)
+...
+---