[AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling

[dstuttard]

Summary

A16 support for image instructions assembly/disassembly (gfx10) was missing

Also refactor MIMG op addr size calcs to common function

We'd got 3 places where the same operation was being done.

One test is now marked XFAIL until a related codegen patch is in place

[dstuttard]

This is a combination of the 2 previous patches - both asmparser and disassembler. I've also unified the code from SIInstrInfo.cpp into a single function since the code was duplicated 3 times before.

[dstuttard]

Fixed sense of NoG16 argument passed in from asmparser and disassembler

[dp]

llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp

159

I'd have renamed "NoG16" to "IsG16Supported" though this is a matter of taste.

169

Am I correct that A16 do affect gradient packing when G16 is not supported? In other words, A16 affects gradient packing on GFX9, right? (gfx9_shader_programming is silent about packing.)

As this is a badly documented feature, could you add a comment explaining how G16 and A16 affect gradient packing?

I compared the tests with your original patch and they are identical. Probably a few extra tests for _d/_cd opcodes with different combinations of A16 and G16 would be useful to have.

[dstuttard]

llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp

159

Yes, I think you are right. I'll make that change.

169

Yes, that's right. If G16 isn't supported, A16 implies G16. (Yes, for gfx9 a16 makes gradients 16 bit too).

Yes, I'll add a comment to that effect.

You're correct, there aren't any new tests. I'll enhance those too.

[dstuttard]

Made suggested changes

Increased test cases

[dp]

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.image.sample.a16.dim.ll

4

I may be mistaken but this comment hints that you are going to submit this patch independently of
https://reviews.llvm.org/D102066. This won't work because many tests you added in the last update will fail - they depend on changes in MIMG_Sampler_AddrSizes. (Maybe there are more dependencies - I did not dig deeper.)

[dstuttard]

Moved change from D102066 required for this change (which will be committed first)

[dstuttard]

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.image.sample.a16.dim.ll

4

You are correct.
I've moved the change into this patch.

I've checked them independently this time and they both pass lit testing.

Not sure that trying to keep these as two patches was worthwhile.

[dp]

Not sure that trying to keep these as two patches was worthwhile.

This definitely made review simpler for me, thanks!

[dp]

LGTM

llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp

public:
   bool hasPackedD16() const {
     return AMDGPU::hasPackedD16(getSTI());
   }
 
   bool hasGFX10A16() const {
     return AMDGPU::hasGFX10A16(getSTI());
   }
 
+  bool hasG16() const { return AMDGPU::hasG16(getSTI()); }
+
   bool isSI() const {
     return AMDGPU::isSI(getSTI());
   }
 
   bool isCI() const {
     return AMDGPU::isCI(getSTI());
   }
 
bool AMDGPUAsmParser::validateMIMGAddrSize(const MCInst &Inst) {
 
   const AMDGPU::MIMGInfo *Info = AMDGPU::getMIMGInfo(Opc);
 
   const AMDGPU::MIMGBaseOpcodeInfo *BaseOpcode =
       AMDGPU::getMIMGBaseOpcodeInfo(Info->BaseOpcode);
   int VAddr0Idx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::vaddr0);
   int SrsrcIdx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::srsrc);
   int DimIdx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::dim);
+  int A16Idx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::a16);
 
   assert(VAddr0Idx != -1);
   assert(SrsrcIdx != -1);
   assert(SrsrcIdx > VAddr0Idx);
 
   if (DimIdx == -1)
     return true; // intersect_ray
 
   unsigned Dim = Inst.getOperand(DimIdx).getImm();
   const AMDGPU::MIMGDimInfo *DimInfo = AMDGPU::getMIMGDimInfoByEncoding(Dim);
   bool IsNSA = SrsrcIdx - VAddr0Idx > 1;
   unsigned VAddrSize =
       IsNSA ? SrsrcIdx - VAddr0Idx
             : AMDGPU::getRegOperandSize(getMRI(), Desc, VAddr0Idx) / 4;
+  bool IsA16 = (A16Idx != -1 && Inst.getOperand(A16Idx).getImm());
+
+  unsigned AddrSize =
+      AMDGPU::getAddrSizeMIMGOp(BaseOpcode, DimInfo, IsA16, hasG16());
 
-  unsigned AddrSize = BaseOpcode->NumExtraArgs +
-                      (BaseOpcode->Gradients ? DimInfo->NumGradients : 0) +
-                      (BaseOpcode->Coordinates ? DimInfo->NumCoords : 0) +
-                      (BaseOpcode->LodOrClampOrMip ? 1 : 0);
   if (!IsNSA) {
     if (AddrSize > 8)
       AddrSize = 16;
     else if (AddrSize > 4)
       AddrSize = 8;
   }
 
   return VAddrSize == AddrSize;

llvm/lib/Target/AMDGPU/Disassembler/AMDGPUDisassembler.cpp

DecodeStatus AMDGPUDisassembler::convertMIMGInst(MCInst &MI) const {
   bool IsGather4 = MCII->get(MI.getOpcode()).TSFlags & SIInstrFlags::Gather4;
 
   bool IsNSA = false;
   unsigned AddrSize = Info->VAddrDwords;
 
   if (STI.getFeatureBits()[AMDGPU::FeatureGFX10]) {
     unsigned DimIdx =
         AMDGPU::getNamedOperandIdx(MI.getOpcode(), AMDGPU::OpName::dim);
+    int A16Idx =
+        AMDGPU::getNamedOperandIdx(MI.getOpcode(), AMDGPU::OpName::a16);
     const AMDGPU::MIMGBaseOpcodeInfo *BaseOpcode =
         AMDGPU::getMIMGBaseOpcodeInfo(Info->BaseOpcode);
     const AMDGPU::MIMGDimInfo *Dim =
         AMDGPU::getMIMGDimInfoByEncoding(MI.getOperand(DimIdx).getImm());
+    const bool IsA16 = (A16Idx != -1 && MI.getOperand(A16Idx).getImm());
+
+    AddrSize =
+        AMDGPU::getAddrSizeMIMGOp(BaseOpcode, Dim, IsA16, AMDGPU::hasG16(STI));
 
-    AddrSize = BaseOpcode->NumExtraArgs +
-               (BaseOpcode->Gradients ? Dim->NumGradients : 0) +
-               (BaseOpcode->Coordinates ? Dim->NumCoords : 0) +
-               (BaseOpcode->LodOrClampOrMip ? 1 : 0);
     IsNSA = Info->MIMGEncoding == AMDGPU::MIMGEncGfx10NSA;
     if (!IsNSA) {
       if (AddrSize > 8)
         AddrSize = 16;
       else if (AddrSize > 4)
         AddrSize = 8;
     } else {
       if (AddrSize > Info->VAddrDwords) {

llvm/lib/Target/AMDGPU/MIMGInstructions.td

 
 class MIMG_Sampler_AddrSizes<AMDGPUSampleVariant sample> {
   // List of all possible numbers of address words, taking all combinations of
   // A16 and image dimension into account (note: no MSAA, since this is for
   // sample/gather ops).
   list<int> AllNumAddrWords =
     !foreach(dw, !if(sample.Gradients,
                      !if(!eq(sample.LodOrClamp, ""),
-                         [2, 3, 4, 5, 6, 7, 9],
-                         [2, 3, 4, 5, 7, 8, 10]),
+                         [2, 3, 4, 5, 6, 7, 8, 9],
+                         [2, 3, 4, 5, 6, 7, 8, 9, 10]),
                      !if(!eq(sample.LodOrClamp, ""),
                          [1, 2, 3],
                          [1, 2, 3, 4])),
              !add(dw, !size(sample.ExtraAddrArgs)));
 
   // Generate machine instructions based on possible register classes for the
   // required numbers of address words. The disassembler defaults to the
   // smallest register class.

llvm/lib/Target/AMDGPU/SIInstrInfo.cpp

if (DimOp) {
       if (ST.hasR128A16()) {
         const MachineOperand *R128A16 = getNamedOperand(MI, AMDGPU::OpName::r128);
         IsA16 = R128A16->getImm() != 0;
       } else if (ST.hasGFX10A16()) {
         const MachineOperand *A16 = getNamedOperand(MI, AMDGPU::OpName::a16);
         IsA16 = A16->getImm() != 0;
       }
 
-      bool PackDerivatives = IsA16 || BaseOpcode->G16;
       bool IsNSA = SRsrcIdx - VAddr0Idx > 1;
 
-      unsigned AddrWords = BaseOpcode->NumExtraArgs;
-      unsigned AddrComponents = (BaseOpcode->Coordinates ? Dim->NumCoords : 0) +
-                                (BaseOpcode->LodOrClampOrMip ? 1 : 0);
-      if (IsA16)
-        AddrWords += divideCeil(AddrComponents, 2);
-      else
-        AddrWords += AddrComponents;
-
-      if (BaseOpcode->Gradients) {
-        if (PackDerivatives)
-          // There are two gradients per coordinate, we pack them separately.
-          // For the 3d case, we get (dy/du, dx/du) (-, dz/du) (dy/dv, dx/dv) (-, dz/dv)
-          AddrWords += alignTo<2>(Dim->NumGradients / 2);
-        else
-          AddrWords += Dim->NumGradients;
-      }
+      unsigned AddrWords =
+          AMDGPU::getAddrSizeMIMGOp(BaseOpcode, Dim, IsA16, ST.hasG16());
 
       unsigned VAddrWords;
       if (IsNSA) {
         VAddrWords = SRsrcIdx - VAddr0Idx;
       } else {
         const TargetRegisterClass *RC = getOpRegClass(MI, VAddr0Idx);
         VAddrWords = MRI.getTargetRegisterInfo()->getRegSizeInBits(*RC) / 32;
         if (AddrWords > 8)

llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.h

 
 LLVM_READONLY
 int getMIMGOpcode(unsigned BaseOpcode, unsigned MIMGEncoding,
                   unsigned VDataDwords, unsigned VAddrDwords);
 
 LLVM_READONLY
 int getMaskedMIMGOp(unsigned Opc, unsigned NewChannels);
 
+LLVM_READONLY
+unsigned getAddrSizeMIMGOp(const MIMGBaseOpcodeInfo *BaseOpcode,
+                           const MIMGDimInfo *Dim, bool IsA16,
+                           bool IsG16Supported);
+
 struct MIMGInfo {
   uint16_t Opcode;
   uint16_t BaseOpcode;
   uint8_t MIMGEncoding;
   uint8_t VDataDwords;
   uint8_t VAddrDwords;
 };
 

llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp

 int getMaskedMIMGOp(unsigned Opc, unsigned NewChannels) {
   const MIMGInfo *OrigInfo = getMIMGInfo(Opc);
   const MIMGInfo *NewInfo =
       getMIMGOpcodeHelper(OrigInfo->BaseOpcode, OrigInfo->MIMGEncoding,
                           NewChannels, OrigInfo->VAddrDwords);
   return NewInfo ? NewInfo->Opcode : -1;
 }
 
+unsigned getAddrSizeMIMGOp(const MIMGBaseOpcodeInfo *BaseOpcode,
+                           const MIMGDimInfo *Dim, bool IsA16,
+                           bool IsG16Supported) {
+  unsigned AddrWords = BaseOpcode->NumExtraArgs;
+  unsigned AddrComponents = (BaseOpcode->Coordinates ? Dim->NumCoords : 0) +
+                            (BaseOpcode->LodOrClampOrMip ? 1 : 0);
+  if (IsA16)
+    AddrWords += divideCeil(AddrComponents, 2);
+  else
+    AddrWords += AddrComponents;
+
+  // Note: For subtargets that support A16 but not G16, enabling A16 also
+  // enables 16 bit gradients.
+  // For subtargets that support A16 (operand) and G16 (done with a different
+  // instruction encoding), they are independent.
+
+  if (BaseOpcode->Gradients) {
+    if ((IsA16 && !IsG16Supported) || BaseOpcode->G16)
+      // There are two gradients per coordinate, we pack them separately.
+      // For the 3d case,
+      // we get (dy/du, dx/du) (-, dz/du) (dy/dv, dx/dv) (-, dz/dv)
+      AddrWords += alignTo<2>(Dim->NumGradients / 2);
+    else
+      AddrWords += Dim->NumGradients;
+  }
+  return AddrWords;
+}
+
 struct MUBUFInfo {
   uint16_t Opcode;
   uint16_t BaseOpcode;
   uint8_t elements;
   bool has_vaddr;
   bool has_srsrc;
   bool has_soffset;
   bool IsBufferInv;

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.image.sample.a16.dim.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX9 %s
 ; RUN: llc -march=amdgcn -mcpu=gfx1010 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX10 %s
+; This test is expected to fail until the new a16/g16 codegen changes are in place
+; XFAIL: *
 
 define amdgpu_ps <4 x float> @sample_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %s) {
 ; GFX9-LABEL: sample_1d:
 ; GFX9:       ; %bb.0: ; %main_body
 ; GFX9-NEXT:    s_mov_b64 s[12:13], exec
 ; GFX9-NEXT:    s_wqm_b64 exec, exec
 ; GFX9-NEXT:    s_and_b64 exec, exec, s[12:13]
 ; GFX9-NEXT:    image_sample v[0:3], v0, s[0:7], s[8:11] dmask:0xf a16

llvm/test/MC/AMDGPU/gfx10_asm_mimg.s

 ; GFX10: image_sample_cd_cl_o v[64:66], [v32, v16, v0, v2, v1, v4, v5, v6, v7, v8, v9], s[4:11], s[100:103] dmask:0x7 dim:SQ_RSRC_IMG_3D ; encoding: [0x16,0x07,0xb4,0xf1,0x20,0x40,0x21,0x03,0x10,0x00,0x02,0x01,0x04,0x05,0x06,0x07,0x08,0x09,0x00,0x00]
 
 image_sample_c_cd_o v[64:66], [v32, v16, v0, v2, v1, v4, v5, v6, v7, v8, v9], s[4:11], s[100:103] dmask:0x7 dim:SQ_RSRC_IMG_3D
 ; GFX10: image_sample_c_cd_o v[64:66], [v32, v16, v0, v2, v1, v4, v5, v6, v7, v8, v9], s[4:11], s[100:103] dmask:0x7 dim:SQ_RSRC_IMG_3D ; encoding: [0x16,0x07,0xb8,0xf1,0x20,0x40,0x21,0x03,0x10,0x00,0x02,0x01,0x04,0x05,0x06,0x07,0x08,0x09,0x00,0x00]
 
 image_sample_c_cd_cl_o v[64:66], [v32, v16, v0, v2, v1, v4, v5, v6, v7, v8, v9, v10], s[4:11], s[100:103] dmask:0x7 dim:SQ_RSRC_IMG_3D
 ; GFX10: image_sample_c_cd_cl_o v[64:66], [v32, v16, v0, v2, v1, v4, v5, v6, v7, v8, v9, v10], s[4:11], s[100:103] dmask:0x7 dim:SQ_RSRC_IMG_3D ; encoding: [0x16,0x07,0xbc,0xf1,0x20,0x40,0x21,0x03,0x10,0x00,0x02,0x01,0x04,0x05,0x06,0x07,0x08,0x09,0x0a,0x00]
 
-image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16
-; GFX10: image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x08,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
+image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16
+; GFX10: image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x08,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
 
 image_load v[0:4], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm tfe
 ; GFX10: image_load v[0:4], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm tfe ; encoding: [0x08,0x1f,0x01,0xf0,0x00,0x00,0x00,0x00]
 
-image_load v[0:4], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 tfe
-; GFX10: image_load v[0:4], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 tfe ; encoding: [0x08,0x1f,0x01,0xf0,0x00,0x00,0x00,0x40]
+image_load v[0:4], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 tfe
+; GFX10: image_load v[0:4], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 tfe ; encoding: [0x08,0x1f,0x01,0xf0,0x00,0x00,0x00,0x40]
+
+image_load v1, v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_load v1, v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x01,0x00,0xf0,0x01,0x01,0x04,0x40]
+
+image_load v[1:2], v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 tfe
+; GFX10: image_load v[1:2], v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 tfe ; encoding: [0x08,0x01,0x01,0xf0,0x01,0x01,0x04,0x40]
+
+image_load v1, v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 lwe
+; GFX10: image_load v1, v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 lwe ; encoding: [0x08,0x01,0x02,0xf0,0x01,0x01,0x04,0x40]
+
+image_load v[1:2], v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 tfe lwe
+; GFX10: image_load v[1:2], v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 tfe lwe ; encoding: [0x08,0x01,0x03,0xf0,0x01,0x01,0x04,0x40]
+
+image_load v[1:2], v1, s[16:23] dmask:0x3 dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_load v[1:2], v1, s[16:23] dmask:0x3 dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x03,0x00,0xf0,0x01,0x01,0x04,0x40]
+
+image_load v[1:4], v1, s[16:23] dmask:0x7 dim:SQ_RSRC_IMG_2D a16 tfe
+; GFX10: image_load v[1:4], v1, s[16:23] dmask:0x7 dim:SQ_RSRC_IMG_2D a16 tfe ; encoding: [0x08,0x07,0x01,0xf0,0x01,0x01,0x04,0x40]
+
+image_load v[1:4], v1, s[16:23] dmask:0xf dim:SQ_RSRC_IMG_2D a16 lwe
+; GFX10: image_load v[1:4], v1, s[16:23] dmask:0xf dim:SQ_RSRC_IMG_2D a16 lwe ; encoding: [0x08,0x0f,0x02,0xf0,0x01,0x01,0x04,0x40]
+
+image_load v[1:3], v1, s[16:23] dmask:0x5 dim:SQ_RSRC_IMG_2D a16 tfe lwe
+; GFX10: image_load v[1:3], v1, s[16:23] dmask:0x5 dim:SQ_RSRC_IMG_2D a16 tfe lwe ; encoding: [0x08,0x05,0x03,0xf0,0x01,0x01,0x04,0x40]
+
+image_sample_d v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16
+; GFX10: image_sample_d v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40]
+
+image_sample_d v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_sample_d v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40]
+
+image_sample_d v[0:3], [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D a16
+; GFX10: image_sample_d v[0:3], [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D a16 ; encoding: [0x14,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00]
+
+image_sample_c_d v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16
+; GFX10: image_sample_c_d v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x40]
+
+image_sample_c_d v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_sample_c_d v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x40]
+
+image_sample_d_cl v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16
+; GFX10: image_sample_d_cl v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x40]
+
+image_sample_d_cl v[0:3], [v0, v1, v2, v3, v4, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_sample_d_cl v[0:3], [v0, v1, v2, v3, v4, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x0c,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x06,0x00,0x00,0x00]
+
+image_sample_c_d_cl v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16
+; GFX10: image_sample_c_d_cl v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xac,0xf0,0x00,0x00,0x40,0x40]
+
+image_sample_c_d_cl v[0:3], [v0, v1, v2, v3, v4, v5, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_sample_c_d_cl v[0:3], [v0, v1, v2, v3, v4, v5, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x0c,0x0f,0xac,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x07,0x00,0x00]
+
+image_sample_cd v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16
+; GFX10: image_sample_cd v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x40]
+
+image_sample_cd v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_sample_cd v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x40]
+
+image_sample_c_cd v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16
+; GFX10: image_sample_c_cd v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x40]
+
+image_sample_c_cd v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_sample_c_cd v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x40]
+
+image_sample_cd_cl v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16
+; GFX10: image_sample_cd_cl v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x40]
+
+image_sample_cd_cl v[0:3], [v0, v1, v2, v3, v4, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_sample_cd_cl v[0:3], [v0, v1, v2, v3, v4, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x0c,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x06,0x00,0x00,0x00]
+
+image_sample_c_cd_cl v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16
+; GFX10: image_sample_c_cd_cl v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xac,0xf1,0x00,0x00,0x40,0x40]
+
+image_sample_c_cd_cl v[0:3], [v0, v1, v2, v3, v4, v5, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16
+; GFX10: image_sample_c_cd_cl v[0:3], [v0, v1, v2, v3, v4, v5, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x0c,0x0f,0xac,0xf1,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x07,0x00,0x00]
+
+image_sample_c_d_o v0, [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY a16
+; GFX10: image_sample_c_d_o v0, [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x2c,0x04,0xe8,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00]
+
+image_sample_c_d_o v[0:1], [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY a16
+; GFX10: image_sample_c_d_o v[0:1], [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x2c,0x06,0xe8,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00]
+
+image_sample_d_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
+; GFX10: image_sample_d_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00]
+
+image_sample_d_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
+; GFX10: image_sample_d_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x00]
+
+image_sample_d_g16 v[0:3], [v0, v2, v3, v5, v6, v7, v8], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D
+; GFX10: image_sample_d_g16 v[0:3], [v0, v2, v3, v5, v6, v7, v8], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D ; encoding: [0x15,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00,0x02,0x03,0x05,0x06,0x07,0x08,0x00,0x00]
+
+image_sample_c_d_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
+; GFX10: image_sample_c_d_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x00]
+
+image_sample_c_d_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
+; GFX10: image_sample_c_d_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06]
+
+image_sample_d_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
+; GFX10: image_sample_d_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x00]
+
+image_sample_d_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
+; GFX10: image_sample_d_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x06]
+
+image_sample_c_d_cl_g16 v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
+; GFX10: image_sample_c_d_cl_g16 v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xac,0xf0,0x00,0x00,0x40,0x00]
+
+image_sample_c_d_cl_g16 v[0:3], [v0, v1, v3, v5, v6, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
+; GFX10: image_sample_c_d_cl_g16 v[0:3], [v0, v1, v3, v5, v6, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0d,0x0f,0xac,0xf0,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06,0x07,0x00,0x00,0x00]
+
+image_sample_cd_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
+; GFX10: image_sample_cd_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x00]
+
+image_sample_cd_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
+; GFX10: image_sample_cd_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x00]
+
+image_sample_c_cd_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
+; GFX10: image_sample_c_cd_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x00]
+
+image_sample_c_cd_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
+; GFX10: image_sample_c_cd_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06]
+
+image_sample_cd_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
+; GFX10: image_sample_cd_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x00]
+
+image_sample_cd_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
+; GFX10: image_sample_cd_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x06]
+
+image_sample_c_cd_cl_g16 v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
+; GFX10: image_sample_c_cd_cl_g16 v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xac,0xf1,0x00,0x00,0x40,0x00]
+
+image_sample_c_cd_cl_g16 v[0:3], [v0, v1, v3, v5, v6, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
+; GFX10: image_sample_c_cd_cl_g16 v[0:3], [v0, v1, v3, v5, v6, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0d,0x0f,0xac,0xf1,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06,0x07,0x00,0x00,0x00]
+
+image_sample_c_d_o_g16 v0, [v0, v1, v2, v4, v6, v7, v8], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY
+; GFX10: image_sample_c_d_o_g16 v0, [v0, v1, v2, v4, v6, v7, v8], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x2d,0x04,0xe8,0xf0,0x00,0x00,0x40,0x00,0x01,0x02,0x04,0x06,0x07,0x08,0x00,0x00]
+
+image_sample_c_d_o_g16 v[0:1], [v0, v1, v2, v4, v6, v7, v8], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY
+; GFX10: image_sample_c_d_o_g16 v[0:1], [v0, v1, v2, v4, v6, v7, v8], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x2d,0x06,0xe8,0xf0,0x00,0x00,0x40,0x00,0x01,0x02,0x04,0x06,0x07,0x08,0x00,0x00]

llvm/test/MC/Disassembler/AMDGPU/mimg_gfx10.txt

+# RUN: llvm-mc -arch=amdgcn -mcpu=gfx1010 -disassemble -show-encoding < %s | FileCheck %s -check-prefix=GFX10
+
+# GFX10: image_load v[4:6], v238, s[28:35] dmask:0x7 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x17,0x00,0xf0,0xee,0x04,0x07,0x00]
+0x00,0x17,0x00,0xf0,0xee,0x04,0x07,0x00
+
+# GFX10: image_load_pck v5, v0, s[8:15] dmask:0x1 dim:SQ_RSRC_IMG_1D glc ; encoding: [0x00,0x21,0x08,0xf0,0x00,0x05,0x02,0x00]
+0x00,0x21,0x08,0xf0,0x00,0x05,0x02,0x00
+
+# GFX10: image_load_pck_sgn v5, v0, s[8:15] dmask:0x1 dim:SQ_RSRC_IMG_1D lwe ; encoding: [0x00,0x01,0x0e,0xf0,0x00,0x05,0x02,0x00]
+0x00,0x01,0x0e,0xf0,0x00,0x05,0x02,0x00
+
+# GFX10: image_load_mip v5, v[0:1], s[8:15] dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x00,0x04,0xf0,0x00,0x05,0x02,0x00]
+0x00,0x00,0x04,0xf0,0x00,0x05,0x02,0x00
+
+# GFX10: image_load_mip_pck v5, v[1:2], s[8:15] dmask:0x1 dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x01,0x10,0xf0,0x01,0x05,0x02,0x00]
+0x00,0x01,0x10,0xf0,0x01,0x05,0x02,0x00
+
+# GFX10: image_load_mip_pck_sgn v[4:5], v[0:1], s[8:15] dmask:0x5 dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x05,0x14,0xf0,0x00,0x04,0x02,0x00]
+0x00,0x05,0x14,0xf0,0x00,0x04,0x02,0x00
+
+# GFX10: image_store v[192:194], v238, s[28:35] dmask:0x7 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x17,0x20,0xf0,0xee,0xc0,0x07,0x00]
+0x00,0x17,0x20,0xf0,0xee,0xc0,0x07,0x00
+
+# GFX10: image_store_pck v1, v2, s[12:19] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x28,0xf0,0x02,0x01,0x03,0x00]
+0x00,0x51,0x28,0xf0,0x02,0x01,0x03,0x00
+
+# GFX10: image_store_mip v1, v[2:3], s[12:19] dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x00,0x24,0xf0,0x02,0x01,0x03,0x00]
+0x00,0x00,0x24,0xf0,0x02,0x01,0x03,0x00
+
+# GFX10: image_store_mip_pck v252, v[2:3], s[12:19] dmask:0x1 dim:SQ_RSRC_IMG_1D r128 ; encoding: [0x00,0x81,0x2c,0xf0,0x02,0xfc,0x03,0x00]
+0x00,0x81,0x2c,0xf0,0x02,0xfc,0x03,0x00
+
+# GFX10: image_atomic_sub v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc ; encoding: [0x00,0x31,0x48,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x31,0x48,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_and v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x60,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x60,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_cmpswap v[4:5], v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc ; encoding: [0x00,0x31,0x40,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x31,0x40,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_add v[4:5], v192, s[28:35] dmask:0x3 dim:SQ_RSRC_IMG_1D unorm glc ; encoding: [0x00,0x33,0x44,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x33,0x44,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_or v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x64,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x64,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_xor v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x68,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x68,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_sub v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x48,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x48,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_smin v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x50,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x50,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_smax v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x58,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x58,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_umin v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x54,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x54,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_umax v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x5c,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x5c,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_inc v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x6c,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x6c,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_atomic_dec v4, v192, s[28:35] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x70,0xf0,0xc0,0x04,0x07,0x00]
+0x00,0x11,0x70,0xf0,0xc0,0x04,0x07,0x00
+
+# GFX10: image_get_resinfo v5, v1, s[8:15] dmask:0x1 dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x01,0x38,0xf0,0x01,0x05,0x02,0x00]
+0x00,0x01,0x38,0xf0,0x01,0x05,0x02,0x00
+
+# GFX10: image_sample v5, v0, s[8:15], s[12:15] dmask:0x1 dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x01,0x80,0xf0,0x00,0x05,0x62,0x00]
+0x00,0x01,0x80,0xf0,0x00,0x05,0x62,0x00
+
+# GFX10: image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x08,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
+0x08,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40
+
+# GFX10: image_load v[0:4], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm tfe ; encoding: [0x08,0x1f,0x01,0xf0,0x00,0x00,0x00,0x00]
+0x08,0x1f,0x01,0xf0,0x00,0x00,0x00,0x00
+
+# GFX10: image_load v[0:4], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 tfe ; encoding: [0x08,0x1f,0x01,0xf0,0x00,0x00,0x00,0x40]
+0x08,0x1f,0x01,0xf0,0x00,0x00,0x00,0x40
+
+# GFX10: image_load v1, v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x01,0x00,0xf0,0x01,0x01,0x04,0x40]
+0x08,0x01,0x00,0xf0,0x01,0x01,0x04,0x40
+
+# GFX10: image_load v[1:2], v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 tfe ; encoding: [0x08,0x01,0x01,0xf0,0x01,0x01,0x04,0x40]
+0x08,0x01,0x01,0xf0,0x01,0x01,0x04,0x40
+
+# GFX10: image_load v1, v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 lwe ; encoding: [0x08,0x01,0x02,0xf0,0x01,0x01,0x04,0x40]
+0x08,0x01,0x02,0xf0,0x01,0x01,0x04,0x40
+
+# GFX10: image_load v[1:2], v1, s[16:23] dmask:0x1 dim:SQ_RSRC_IMG_2D a16 tfe lwe ; encoding: [0x08,0x01,0x03,0xf0,0x01,0x01,0x04,0x40]
+0x08,0x01,0x03,0xf0,0x01,0x01,0x04,0x40
+
+# GFX10: image_load v[1:2], v1, s[16:23] dmask:0x3 dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x03,0x00,0xf0,0x01,0x01,0x04,0x40]
+0x08,0x03,0x00,0xf0,0x01,0x01,0x04,0x40
+
+# GFX10: image_load v[1:4], v1, s[16:23] dmask:0x7 dim:SQ_RSRC_IMG_2D a16 tfe ; encoding: [0x08,0x07,0x01,0xf0,0x01,0x01,0x04,0x40]
+0x08,0x07,0x01,0xf0,0x01,0x01,0x04,0x40
+
+# GFX10: image_load v[1:4], v1, s[16:23] dmask:0xf dim:SQ_RSRC_IMG_2D a16 lwe ; encoding: [0x08,0x0f,0x02,0xf0,0x01,0x01,0x04,0x40]
+0x08,0x0f,0x02,0xf0,0x01,0x01,0x04,0x40
+
+# GFX10: image_load v[1:3], v1, s[16:23] dmask:0x5 dim:SQ_RSRC_IMG_2D a16 tfe lwe ; encoding: [0x08,0x05,0x03,0xf0,0x01,0x01,0x04,0x40]
+0x08,0x05,0x03,0xf0,0x01,0x01,0x04,0x40
+
+# GFX10: image_sample_d v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40]
+0x00,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_d v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40]
+0x08,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_d v[0:3], [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D a16 ; encoding: [0x14,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00]
+0x14,0x0f,0x88,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00
+
+# GFX10: image_sample_c_d v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x40]
+0x00,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_c_d v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x40]
+0x08,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_d_cl v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x40]
+0x00,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_d_cl v[0:3], [v0, v1, v2, v3, v4, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x0c,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x06,0x00,0x00,0x00]
+0x0c,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x06,0x00,0x00,0x00
+
+# GFX10: image_sample_c_d_cl v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xac,0xf0,0x00,0x00,0x40,0x40]
+0x00,0x0f,0xac,0xf0,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_c_d_cl v[0:3], [v0, v1, v2, v3, v4, v5, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x0c,0x0f,0xac,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x07,0x00,0x00]
+0x0c,0x0f,0xac,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x07,0x00,0x00
+
+# GFX10: image_sample_cd v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x40]
+0x00,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_cd v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x40]
+0x08,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_c_cd v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x40]
+0x00,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_c_cd v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x08,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x40]
+0x08,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_cd_cl v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x40]
+0x00,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_cd_cl v[0:3], [v0, v1, v2, v3, v4, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x0c,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x06,0x00,0x00,0x00]
+0x0c,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x06,0x00,0x00,0x00
+
+# GFX10: image_sample_c_cd_cl v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x00,0x0f,0xac,0xf1,0x00,0x00,0x40,0x40]
+0x00,0x0f,0xac,0xf1,0x00,0x00,0x40,0x40
+
+# GFX10: image_sample_c_cd_cl v[0:3], [v0, v1, v2, v3, v4, v5, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x0c,0x0f,0xac,0xf1,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x07,0x00,0x00]
+0x0c,0x0f,0xac,0xf1,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x07,0x00,0x00
+
+# GFX10: image_sample_c_d_o v0, [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x2c,0x04,0xe8,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00]
+0x2c,0x04,0xe8,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00
+
+# GFX10: image_sample_c_d_o v[0:1], [v0, v1, v2, v3, v4, v5, v6, v8], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x2c,0x06,0xe8,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00]
+0x2c,0x06,0xe8,0xf0,0x00,0x00,0x40,0x40,0x01,0x02,0x03,0x04,0x05,0x06,0x08,0x00
+
+# GFX10: image_sample_d_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00]
+0x01,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00
+
+# GFX10: image_sample_d_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x00]
+0x0b,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x00
+
+# GFX10: image_sample_d_g16 v[0:3], [v0, v2, v3, v5, v6, v7, v8], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D ; encoding: [0x15,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00,0x02,0x03,0x05,0x06,0x07,0x08,0x00,0x00]
+0x15,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00,0x02,0x03,0x05,0x06,0x07,0x08,0x00,0x00
+
+# GFX10: image_sample_c_d_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x00]
+0x01,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x00
+
+# GFX10: image_sample_c_d_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06]
+0x0b,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06
+
+# GFX10: image_sample_d_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x00]
+0x01,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x00
+
+# GFX10: image_sample_d_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x06]
+0x0b,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x06
+
+# GFX10: image_sample_c_d_cl_g16 v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xac,0xf0,0x00,0x00,0x40,0x00]
+0x01,0x0f,0xac,0xf0,0x00,0x00,0x40,0x00
+
+# GFX10: image_sample_c_d_cl_g16 v[0:3], [v0, v1, v3, v5, v6, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0d,0x0f,0xac,0xf0,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06,0x07,0x00,0x00,0x00]
+0x0d,0x0f,0xac,0xf0,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06,0x07,0x00,0x00,0x00
+
+# GFX10: image_sample_cd_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x00]
+0x01,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x00
+
+# GFX10: image_sample_cd_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x00]
+0x0b,0x0f,0xa0,0xf1,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x00
+
+# GFX10: image_sample_c_cd_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x00]
+0x01,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x00
+
+# GFX10: image_sample_c_cd_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06]
+0x0b,0x0f,0xa8,0xf1,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06
+
+# GFX10: image_sample_cd_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x00]
+0x01,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x00
+
+# GFX10: image_sample_cd_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x06]
+0x0b,0x0f,0xa4,0xf1,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x06
+
+# GFX10: image_sample_c_cd_cl_g16 v[0:3], v[0:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xac,0xf1,0x00,0x00,0x40,0x00]
+0x01,0x0f,0xac,0xf1,0x00,0x00,0x40,0x00
+
+# GFX10: image_sample_c_cd_cl_g16 v[0:3], [v0, v1, v3, v5, v6, v7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0d,0x0f,0xac,0xf1,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06,0x07,0x00,0x00,0x00]
+0x0d,0x0f,0xac,0xf1,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06,0x07,0x00,0x00,0x00
+
+# GFX10: image_sample_c_d_o_g16 v0, [v0, v1, v2, v4, v6, v7, v8], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x2d,0x04,0xe8,0xf0,0x00,0x00,0x40,0x00,0x01,0x02,0x04,0x06,0x07,0x08,0x00,0x00]
+0x2d,0x04,0xe8,0xf0,0x00,0x00,0x40,0x00,0x01,0x02,0x04,0x06,0x07,0x08,0x00,0x00
+
+# GFX10: image_sample_c_d_o_g16 v[0:1], [v0, v1, v2, v4, v6, v7, v8], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x2d,0x06,0xe8,0xf0,0x00,0x00,0x40,0x00,0x01,0x02,0x04,0x06,0x07,0x08,0x00,0x00]
+0x2d,0x06,0xe8,0xf0,0x00,0x00,0x40,0x00,0x01,0x02,0x04,0x06,0x07,0x08,0x00,0x00