[mlir] Remove VectorToROCDL

[krzysz00]

Summary

Between issues such as
https://github.com/llvm/llvm-project/issues/56323, the fact that this
lowering (unlike the code in amdgpu-to-rocdl) does not correctly set
up bounds checks (and thus will cause page faults on reads that might
need to be padded instead), and that fixing these problems would,
essentially, involve replicating amdgpu-to-rocdl, remove
--vector-to-rocdl for being broken. In addition, the lowering does not
support many aspects of transfer_{read,write}, like supervectors, and
may not work correctly in their presence.

We (the MLIR-based convolution generator at AMD) do not use this
conversion pass, nor are we aware of any other clients.

Migration strategies:

• Use VectorToLLVM
• If buffer ops are particularly needed in your application, use

amdgpu.raw_buffer_{load,store}

A VectorToAMDGPU pass may be introduced in the future.

clang/docs/ClangFormattedStatus.rst

* - mlir/include/mlir/Conversion/VectorToGPU
      - `1`
      - `0`
      - :good:`100%`
    * - mlir/include/mlir/Conversion/VectorToLLVM
      - `1`
      - `1`
      - `0`
      - :good:`100%`
-   * - mlir/include/mlir/Conversion/VectorToROCDL
-     - `1`
-     - `1`
-     - `0`
-     - :good:`100%`
    * - mlir/include/mlir/Conversion/VectorToSCF
      - `1`
      - `1`
      - `0`
      - :good:`100%`
    * - mlir/include/mlir/Conversion/VectorToSPIRV
      - `2`
      - `2`
* - mlir/lib/Conversion/VectorToGPU
      - `0`
      - `1`
      - :none:`0%`
    * - mlir/lib/Conversion/VectorToLLVM
      - `2`
      - `2`
      - `0`
      - :good:`100%`
-   * - mlir/lib/Conversion/VectorToROCDL
-     - `1`
-     - `1`
-     - `0`
-     - :good:`100%`
    * - mlir/lib/Conversion/VectorToSCF
      - `1`
      - `1`
      - `0`
      - :good:`100%`
    * - mlir/lib/Conversion/VectorToSPIRV
      - `2`
      - `1`

clang/docs/tools/clang-formatted-files.txt

 mlir/include/mlir/Conversion/StandardToLLVM/ConvertStandardToLLVMPass.h
 mlir/include/mlir/Conversion/TensorToSPIRV/TensorToSPIRV.h
 mlir/include/mlir/Conversion/TensorToSPIRV/TensorToSPIRVPass.h
 mlir/include/mlir/Conversion/TosaToLinalg/TosaToLinalg.h
 mlir/include/mlir/Conversion/TosaToSCF/TosaToSCF.h
 mlir/include/mlir/Conversion/TosaToStandard/TosaToStandard.h
 mlir/include/mlir/Conversion/VectorToGPU/VectorToGPU.h
 mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h
-mlir/include/mlir/Conversion/VectorToROCDL/VectorToROCDL.h
 mlir/include/mlir/Conversion/VectorToSCF/VectorToSCF.h
 mlir/include/mlir/Conversion/VectorToSPIRV/VectorToSPIRV.h
 mlir/include/mlir/Conversion/VectorToSPIRV/VectorToSPIRVPass.h
 mlir/include/mlir/Dialect/CommonFolders.h
 mlir/include/mlir/Dialect/Traits.h
 mlir/include/mlir/Dialect/Affine/LoopFusionUtils.h
 mlir/include/mlir/Dialect/Affine/LoopUtils.h
 mlir/include/mlir/Dialect/Affine/Passes.h

 mlir/lib/Conversion/TosaToLinalg/TosaToLinalgNamedPass.cpp
 mlir/lib/Conversion/TosaToLinalg/TosaToLinalgPass.cpp
 mlir/lib/Conversion/TosaToSCF/TosaToSCF.cpp
 mlir/lib/Conversion/TosaToSCF/TosaToSCFPass.cpp
 mlir/lib/Conversion/TosaToStandard/TosaToStandard.cpp
 mlir/lib/Conversion/TosaToStandard/TosaToStandardPass.cpp
 mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
 mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
-mlir/lib/Conversion/VectorToROCDL/VectorToROCDL.cpp
 mlir/lib/Conversion/VectorToSCF/VectorToSCF.cpp
 mlir/lib/Conversion/VectorToSPIRV/VectorToSPIRVPass.cpp
 mlir/lib/Dialect/Traits.cpp
 mlir/lib/Dialect/Affine/Analysis/AffineAnalysis.cpp
 mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
 mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
 mlir/lib/Dialect/Affine/Analysis/NestedMatcher.cpp
 mlir/lib/Dialect/Affine/Analysis/Utils.cpp

mlir/include/mlir/Conversion/Passes.h

 #include "mlir/Conversion/TensorToLinalg/TensorToLinalgPass.h"
 #include "mlir/Conversion/TensorToSPIRV/TensorToSPIRVPass.h"
 #include "mlir/Conversion/TosaToArith/TosaToArith.h"
 #include "mlir/Conversion/TosaToLinalg/TosaToLinalg.h"
 #include "mlir/Conversion/TosaToSCF/TosaToSCF.h"
 #include "mlir/Conversion/TosaToTensor/TosaToTensor.h"
 #include "mlir/Conversion/VectorToGPU/VectorToGPU.h"
 #include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
-#include "mlir/Conversion/VectorToROCDL/VectorToROCDL.h"
 #include "mlir/Conversion/VectorToSCF/VectorToSCF.h"
 #include "mlir/Conversion/VectorToSPIRV/VectorToSPIRVPass.h"
 
 namespace mlir {
 
 /// Generate the code for registering conversion passes.
 #define GEN_PASS_REGISTRATION
 #include "mlir/Conversion/Passes.h.inc"
 
 } // namespace mlir
 
 #endif // MLIR_CONVERSION_PASSES_H

mlir/include/mlir/Conversion/Passes.td

 // VectorToGPU
 //===----------------------------------------------------------------------===//
 
 def ConvertVectorToGPU : Pass<"convert-vector-to-gpu"> {
   let summary = "Lower the operations from the vector dialect into the GPU "
                 "dialect";
   let constructor = "mlir::createConvertVectorToGPUPass()";
   let dependentDialects = [
-    "memref::MemRefDialect", "gpu::GPUDialect", "AffineDialect", 
+    "memref::MemRefDialect", "gpu::GPUDialect", "AffineDialect",
     "vector::VectorDialect", "nvgpu::NVGPUDialect"
   ];
 
   let options = [
-    Option<"useNvGpu", "use-nvgpu", "bool", /*default=*/"false", 
+    Option<"useNvGpu", "use-nvgpu", "bool", /*default=*/"false",
       "convert to NvGPU ops instead of GPU dialect ops">
   ];
 }
 
 //===----------------------------------------------------------------------===//
 // VectorToSCF
 //===----------------------------------------------------------------------===//
 
let options = [
     Option<"x86Vector", "enable-x86vector",
            "bool", /*default=*/"false",
            "Enables the use of X86Vector dialect while lowering the vector "
      "dialect.">
   ];
 }
 
 //===----------------------------------------------------------------------===//
-// VectorToROCDL
-//===----------------------------------------------------------------------===//
-
-def ConvertVectorToROCDL : Pass<"convert-vector-to-rocdl", "ModuleOp"> {
-  let summary = "Lower the operations from the vector dialect into the ROCDL "
-                "dialect";
-  let constructor = "mlir::createConvertVectorToROCDLPass()";
-  let dependentDialects = ["ROCDL::ROCDLDialect"];
-}
-
-//===----------------------------------------------------------------------===//
 // VectorToSPIRV
 //===----------------------------------------------------------------------===//
 
 def ConvertVectorToSPIRV : Pass<"convert-vector-to-spirv", "ModuleOp"> {
   let summary = "Convert Vector dialect to SPIR-V dialect";
   let constructor = "mlir::createConvertVectorToSPIRVPass()";
   let dependentDialects = ["spirv::SPIRVDialect"];
 }
 
 #endif // MLIR_CONVERSION_PASSES

mlir/include/mlir/Conversion/VectorToROCDL/VectorToROCDL.h

-//===- VectorToROCDL.h - Convert Vector to ROCDL dialect ---*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-#ifndef MLIR_CONVERSION_VECTORTOROCDL_VECTORTOROCDL_H_
-#define MLIR_CONVERSION_VECTORTOROCDL_VECTORTOROCDL_H_
-
-#include <memory>
-
-namespace mlir {
-class LLVMTypeConverter;
-class ModuleOp;
-template <typename OpT>
-class OperationPass;
-class RewritePatternSet;
-
-/// Collect a set of patterns to convert from the GPU dialect to ROCDL.
-void populateVectorToROCDLConversionPatterns(LLVMTypeConverter &converter,
-                                             RewritePatternSet &patterns);
-
-/// Create a pass to convert vector operations to the ROCDL dialect.
-std::unique_ptr<OperationPass<ModuleOp>> createConvertVectorToROCDLPass();
-
-} // namespace mlir
-#endif // MLIR_CONVERSION_VECTORTOROCDL_VECTORTOROCDL_H_

mlir/lib/Conversion/CMakeLists.txt

 add_subdirectory(ShapeToStandard)
 add_subdirectory(SPIRVToLLVM)
 add_subdirectory(TensorToLinalg)
 add_subdirectory(TensorToSPIRV)
 add_subdirectory(TosaToArith)
 add_subdirectory(TosaToLinalg)
 add_subdirectory(TosaToSCF)
 add_subdirectory(TosaToTensor)
-add_subdirectory(VectorToROCDL)
 add_subdirectory(VectorToLLVM)
 add_subdirectory(VectorToGPU)
 add_subdirectory(VectorToSCF)
 add_subdirectory(VectorToSPIRV)

mlir/lib/Conversion/GPUToROCDL/CMakeLists.txt

add_mlir_conversion_library(MLIRGPUToROCDLTransforms
   MLIRFuncToLLVM
   MLIRGPUOps
   MLIRGPUToGPURuntimeTransforms
   MLIRLLVMCommonConversion
   MLIRLLVMDialect
   MLIRMemRefToLLVM
   MLIRROCDLDialect
   MLIRPass
-  MLIRVectorToROCDL
   )

mlir/lib/Conversion/GPUToROCDL/LowerGpuOpsToROCDLOps.cpp

 #include "mlir/Conversion/ArithmeticToLLVM/ArithmeticToLLVM.h"
 #include "mlir/Conversion/FuncToLLVM/ConvertFuncToLLVM.h"
 #include "mlir/Conversion/LLVMCommon/ConversionTarget.h"
 #include "mlir/Conversion/LLVMCommon/LoweringOptions.h"
 #include "mlir/Conversion/LLVMCommon/Pattern.h"
 #include "mlir/Conversion/LLVMCommon/TypeConverter.h"
 #include "mlir/Conversion/MemRefToLLVM/MemRefToLLVM.h"
 #include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
-#include "mlir/Conversion/VectorToROCDL/VectorToROCDL.h"
 #include "mlir/Dialect/Func/IR/FuncOps.h"
 #include "mlir/Dialect/GPU/IR/GPUDialect.h"
 #include "mlir/Dialect/GPU/Transforms/Passes.h"
 #include "mlir/Dialect/LLVMIR/ROCDLDialect.h"
 #include "mlir/Dialect/Math/IR/Math.h"
 #include "mlir/Dialect/Vector/IR/VectorOps.h"
 #include "mlir/Pass/Pass.h"
 #include "mlir/Transforms/DialectConversion.h"
void runOnOperation() override {
     populateGpuRewritePatterns(patterns);
     (void)applyPatternsAndFoldGreedily(m, std::move(patterns));
 
     mlir::arith::populateArithmeticToLLVMConversionPatterns(converter,
                                                             llvmPatterns);
     populateAMDGPUToROCDLConversionPatterns(converter, llvmPatterns,
                                             *maybeChipset);
     populateVectorToLLVMConversionPatterns(converter, llvmPatterns);
-    populateVectorToROCDLConversionPatterns(converter, llvmPatterns);
     cf::populateControlFlowToLLVMConversionPatterns(converter, llvmPatterns);
     populateFuncToLLVMConversionPatterns(converter, llvmPatterns);
     populateMemRefToLLVMConversionPatterns(converter, llvmPatterns);
     populateGpuToROCDLConversionPatterns(converter, llvmPatterns, runtime);
     LLVMConversionTarget target(getContext());
     configureGpuToROCDLConversionLegality(target);
     if (failed(applyPartialConversion(m, target, std::move(llvmPatterns))))
       signalPassFailure();

mlir/lib/Conversion/VectorToROCDL/CMakeLists.txt

-add_mlir_conversion_library(MLIRVectorToROCDL
-  VectorToROCDL.cpp
-
-  ADDITIONAL_HEADER_DIRS
-  ${MLIR_MAIN_INCLUDE_DIR}/mlir/Conversion/VectorToROCDL
-
-  DEPENDS
-  MLIRConversionPassIncGen
-  intrinsics_gen
-
-  LINK_COMPONENTS
-  Core
-
-  LINK_LIBS PUBLIC
-  MLIRFuncToLLVM
-  MLIRROCDLDialect
-  MLIRLLVMCommonConversion
-  MLIRMemRefToLLVM
-  MLIRVectorDialect
-  MLIRTransforms
-  )

mlir/lib/Conversion/VectorToROCDL/VectorToROCDL.cpp

-//===- VectorToROCDL.cpp - Vector to ROCDL lowering passes ------===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-//
-// This file implements a pass to generate ROCDLIR operations for higher-level
-// Vector operations.
-//
-//===----------------------------------------------------------------------===//
-
-#include "mlir/Conversion/VectorToROCDL/VectorToROCDL.h"
-
-#include "../PassDetail.h"
-#include "mlir/Conversion/FuncToLLVM/ConvertFuncToLLVM.h"
-#include "mlir/Conversion/LLVMCommon/ConversionTarget.h"
-#include "mlir/Conversion/LLVMCommon/Pattern.h"
-#include "mlir/Conversion/MemRefToLLVM/MemRefToLLVM.h"
-#include "mlir/Dialect/GPU/IR/GPUDialect.h"
-#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
-#include "mlir/Dialect/LLVMIR/ROCDLDialect.h"
-#include "mlir/Dialect/Vector/IR/VectorOps.h"
-#include "mlir/Pass/Pass.h"
-#include "mlir/Transforms/DialectConversion.h"
-
-using namespace mlir;
-using namespace mlir::vector;
-
-static LogicalResult replaceTransferOpWithMubuf(
-    ConversionPatternRewriter &rewriter, ValueRange operands,
-    LLVMTypeConverter &typeConverter, Location loc, TransferReadOp xferOp,
-    Type &vecTy, Value &dwordConfig, Value &vindex, Value &offsetSizeInBytes,
-    Value &glc, Value &slc) {
-  rewriter.replaceOpWithNewOp<ROCDL::MubufLoadOp>(
-      xferOp, vecTy, dwordConfig, vindex, offsetSizeInBytes, glc, slc);
-  return success();
-}
-
-static LogicalResult replaceTransferOpWithMubuf(
-    ConversionPatternRewriter &rewriter, ValueRange operands,
-    LLVMTypeConverter &typeConverter, Location loc, TransferWriteOp xferOp,
-    Type &vecTy, Value &dwordConfig, Value &vindex, Value &offsetSizeInBytes,
-    Value &glc, Value &slc) {
-  auto adaptor = TransferWriteOpAdaptor(operands, xferOp->getAttrDictionary());
-  rewriter.replaceOpWithNewOp<ROCDL::MubufStoreOp>(xferOp, adaptor.getVector(),
-                                                   dwordConfig, vindex,
-                                                   offsetSizeInBytes, glc, slc);
-  return success();
-}
-
-namespace {
-/// Conversion pattern that converts a 1-D vector transfer read/write.
-/// Note that this conversion pass only converts vector x2 or x4 f32
-/// types. For unsupported cases, they will fall back to the vector to
-/// llvm conversion pattern.
-template <typename ConcreteOp>
-class VectorTransferConversion : public ConvertOpToLLVMPattern<ConcreteOp> {
-public:
-  using ConvertOpToLLVMPattern<ConcreteOp>::ConvertOpToLLVMPattern;
-
-  LogicalResult
-  matchAndRewrite(ConcreteOp xferOp, typename ConcreteOp::Adaptor adaptor,
-                  ConversionPatternRewriter &rewriter) const override {
-    // TODO: support 0-d corner case.
-    if (xferOp.getTransferRank() == 0)
-      return failure();
-
-    if (xferOp.getVectorType().getRank() > 1 ||
-        llvm::size(xferOp.getIndices()) == 0)
-      return failure();
-
-    if (!xferOp.getPermutationMap().isMinorIdentity())
-      return failure();
-
-    // Have it handled in vector->llvm conversion pass.
-    if (xferOp.isDimInBounds(0))
-      return failure();
-
-    auto toLLVMTy = [&](Type t) {
-      return this->getTypeConverter()->convertType(t);
-    };
-    auto vecTy = toLLVMTy(xferOp.getVectorType());
-    unsigned vecWidth = LLVM::getVectorNumElements(vecTy).getFixedValue();
-    Location loc = xferOp->getLoc();
-
-    // The backend result vector scalarization have trouble scalarize
-    // <1 x ty> result, exclude the x1 width from the lowering.
-    if (vecWidth != 2 && vecWidth != 4)
-      return failure();
-
-    // Obtain dataPtr and elementType from the memref.
-    auto memRefType = xferOp.getShapedType().template dyn_cast<MemRefType>();
-    if (!memRefType)
-      return failure();
-    // MUBUF instruction operate only on addresspace 0(unified) or 1(global)
-    // In case of 3(LDS): fall back to vector->llvm pass
-    // In case of 5(VGPR): wrong
-    if ((memRefType.getMemorySpaceAsInt() != 0) &&
-        (memRefType.getMemorySpaceAsInt() != 1))
-      return failure();
-
-    // Note that the dataPtr starts at the offset address specified by
-    // indices, so no need to calculate offset size in bytes again in
-    // the MUBUF instruction.
-    Value dataPtr = this->getStridedElementPtr(
-        loc, memRefType, adaptor.getSource(), adaptor.getIndices(), rewriter);
-
-    // 1. Create and fill a <4 x i32> dwordConfig with:
-    //    1st two elements holding the address of dataPtr.
-    //    3rd element: -1.
-    //    4th element: 0x27000.
-    SmallVector<int32_t, 4> constConfigAttr{0, 0, -1, 0x27000};
-    Type i32Ty = rewriter.getIntegerType(32);
-    VectorType i32Vecx4 = VectorType::get(4, i32Ty);
-    Value constConfig = rewriter.create<LLVM::ConstantOp>(
-        loc, toLLVMTy(i32Vecx4),
-        DenseElementsAttr::get(i32Vecx4, ArrayRef<int32_t>(constConfigAttr)));
-
-    // Treat first two element of <4 x i32> as i64, and save the dataPtr
-    // to it.
-    Type i64Ty = rewriter.getIntegerType(64);
-    Value i64x2Ty = rewriter.create<LLVM::BitcastOp>(
-        loc, LLVM::getFixedVectorType(toLLVMTy(i64Ty), 2), constConfig);
-    Value dataPtrAsI64 = rewriter.create<LLVM::PtrToIntOp>(
-        loc, toLLVMTy(i64Ty).template cast<Type>(), dataPtr);
-    Value zero = this->createIndexConstant(rewriter, loc, 0);
-    Value dwordConfig = rewriter.create<LLVM::InsertElementOp>(
-        loc, LLVM::getFixedVectorType(toLLVMTy(i64Ty), 2), i64x2Ty,
-        dataPtrAsI64, zero);
-    dwordConfig =
-        rewriter.create<LLVM::BitcastOp>(loc, toLLVMTy(i32Vecx4), dwordConfig);
-
-    // 2. Rewrite op as a buffer read or write.
-    Value int1False = rewriter.create<LLVM::ConstantOp>(
-        loc, toLLVMTy(rewriter.getIntegerType(1)),
-        rewriter.getIntegerAttr(rewriter.getIntegerType(1), 0));
-    Value int32Zero = rewriter.create<LLVM::ConstantOp>(
-        loc, toLLVMTy(i32Ty),
-        rewriter.getIntegerAttr(rewriter.getIntegerType(32), 0));
-    return replaceTransferOpWithMubuf(
-        rewriter, adaptor.getOperands(), *this->getTypeConverter(), loc, xferOp,
-        vecTy, dwordConfig, int32Zero, int32Zero, int1False, int1False);
-  }
-};
-} // namespace
-
-void mlir::populateVectorToROCDLConversionPatterns(
-    LLVMTypeConverter &converter, RewritePatternSet &patterns) {
-  patterns.add<VectorTransferConversion<TransferReadOp>,
-               VectorTransferConversion<TransferWriteOp>>(converter);
-}
-
-namespace {
-struct LowerVectorToROCDLPass
-    : public ConvertVectorToROCDLBase<LowerVectorToROCDLPass> {
-  void runOnOperation() override;
-};
-} // namespace
-
-void LowerVectorToROCDLPass::runOnOperation() {
-  LLVMTypeConverter converter(&getContext());
-  RewritePatternSet patterns(&getContext());
-
-  populateVectorToROCDLConversionPatterns(converter, patterns);
-  populateMemRefToLLVMConversionPatterns(converter, patterns);
-  populateFuncToLLVMConversionPatterns(converter, patterns);
-
-  LLVMConversionTarget target(getContext());
-  target.addLegalDialect<ROCDL::ROCDLDialect>();
-
-  if (failed(
-          applyPartialConversion(getOperation(), target, std::move(patterns))))
-    signalPassFailure();
-}
-
-std::unique_ptr<OperationPass<ModuleOp>>
-mlir::createConvertVectorToROCDLPass() {
-  return std::make_unique<LowerVectorToROCDLPass>();
-}

mlir/test/Conversion/VectorToROCDL/vector-to-rocdl.mlir

-// RUN: mlir-opt %s -convert-vector-to-rocdl | FileCheck %s
-
-gpu.module @test_read{
-func.func @transfer_readx2(%A : memref<?xf32>, %base: index) -> vector<2xf32> {
-  %f0 = arith.constant 0.0: f32
-  %f = vector.transfer_read %A[%base], %f0
-      {permutation_map = affine_map<(d0) -> (d0)>} :
-    memref<?xf32>, vector<2xf32>
-  return %f: vector<2xf32>
-}
-// CHECK-LABEL: @transfer_readx2
-// CHECK: rocdl.buffer.load {{.*}} vector<2xf32>
-
-func.func @transfer_readx4(%A : memref<?xf32>, %base: index) -> vector<4xf32> {
-  %f0 = arith.constant 0.0: f32
-  %f = vector.transfer_read %A[%base], %f0
-      {permutation_map = affine_map<(d0) -> (d0)>} :
-    memref<?xf32>, vector<4xf32>
-  return %f: vector<4xf32>
-}
-// CHECK-LABEL: @transfer_readx4
-// CHECK: rocdl.buffer.load {{.*}} vector<4xf32>
-
-func.func @transfer_read_dwordConfig(%A : memref<?xf32>, %base: index) -> vector<4xf32> {
-  %f0 = arith.constant 0.0: f32
-  %f = vector.transfer_read %A[%base], %f0
-      {permutation_map = affine_map<(d0) -> (d0)>} :
-    memref<?xf32>, vector<4xf32>
-  return %f: vector<4xf32>
-}
-// CHECK-LABEL: @transfer_read_dwordConfig
-// CHECK: %[[gep:.*]] = llvm.getelementptr {{.*}}
-// CHECK: [0, 0, -1, 159744]
-// CHECK: %[[i64:.*]] = llvm.ptrtoint %[[gep]]
-// CHECK: llvm.insertelement %[[i64]]
-}
-
-gpu.module @test_write{
-func.func @transfer_writex2(%A : memref<?xf32>, %B : vector<2xf32>, %base: index) {
-  vector.transfer_write %B, %A[%base]
-      {permutation_map = affine_map<(d0) -> (d0)>} :
-    vector<2xf32>, memref<?xf32>
-  return
-}
-// CHECK-LABEL: @transfer_writex2
-// CHECK: rocdl.buffer.store {{.*}} vector<2xf32>
-
-func.func @transfer_writex4(%A : memref<?xf32>, %B : vector<4xf32>, %base: index) {
-  vector.transfer_write %B, %A[%base]
-      {permutation_map = affine_map<(d0) -> (d0)>} :
-    vector<4xf32>, memref<?xf32>
-  return
-}
-// CHECK-LABEL: @transfer_writex4
-// CHECK: rocdl.buffer.store {{.*}} vector<4xf32>
-
-func.func @transfer_write_dwordConfig(%A : memref<?xf32>, %B : vector<2xf32>, %base: index) {
-  vector.transfer_write %B, %A[%base]
-      {permutation_map = affine_map<(d0) -> (d0)>} :
-    vector<2xf32>, memref<?xf32>
-  return
-}
-// CHECK-LABEL: @transfer_write_dwordConfig
-// CHECK: %[[gep:.*]] = llvm.getelementptr {{.*}}
-// CHECK: [0, 0, -1, 159744]
-// CHECK: %[[i64:.*]] = llvm.ptrtoint %[[gep]]
-// CHECK: llvm.insertelement %[[i64]]
-}

mlir/test/Integration/GPU/ROCM/vector-transferops.mlir

 // RUN: mlir-opt %s \
 // RUN:   -convert-scf-to-cf \
 // RUN:   -gpu-kernel-outlining \
-// RUN:   -pass-pipeline='gpu.module(strip-debuginfo,convert-gpu-to-rocdl,gpu-to-hsaco{chip=%chip})' \
+// RUN:   -pass-pipeline='gpu.module(strip-debuginfo,convert-gpu-to-rocdl{chipset=%chip index-bitwidth=32},gpu-to-hsaco{chip=%chip})' \
 // RUN:   -gpu-to-llvm \
 // RUN: | mlir-cpu-runner \
 // RUN:   --shared-libs=%linalg_test_lib_dir/libmlir_rocm_runtime%shlibext \
 // RUN:   --shared-libs=%linalg_test_lib_dir/libmlir_runner_utils%shlibext \
 // RUN:   --entry-point-result=void \
 // RUN: | FileCheck %s
 
+// TODO: swap for vector transfer reads if we ever create a --vector-to-amdgpu
 func.func @vectransferx2(%arg0 : memref<?xf32>, %arg1 : memref<?xf32>) {
   %cst = arith.constant 1 : index
   gpu.launch blocks(%bx, %by, %bz) in (%grid_x = %cst, %grid_y = %cst, %grid_z = %cst)
              threads(%tx, %ty, %tz) in (%block_x = %cst, %block_y = %cst, %block_z = %cst) {
     %f0 = arith.constant 0.0: f32
-    %base = arith.constant 0 : index
-    %f = vector.transfer_read %arg0[%base], %f0
-        {permutation_map = affine_map<(d0) -> (d0)>} :
-      memref<?xf32>, vector<2xf32>
+    %base = arith.constant 0 : i32
+    %f = amdgpu.raw_buffer_load {boundsCheck = true } %arg0[%base]
+      : memref<?xf32>, i32 -> vector<2xf32>
 
     %c = arith.addf %f, %f : vector<2xf32>
 
-    %base1 = arith.constant 1 : index
-    vector.transfer_write %c, %arg1[%base1]
-        {permutation_map = affine_map<(d0) -> (d0)>} :
-      vector<2xf32>, memref<?xf32>
+    %base1 = arith.constant 1 : i32
+    amdgpu.raw_buffer_store { boundsCheck = false } %c -> %arg1[%base1]
+      : vector<2xf32> -> memref<?xf32>, i32
 
     gpu.terminator
   }
   return
 }
 
 func.func @vectransferx4(%arg0 : memref<?xf32>, %arg1 : memref<?xf32>) {
   %cst = arith.constant 1 : index
   gpu.launch blocks(%bx, %by, %bz) in (%grid_x = %cst, %grid_y = %cst, %grid_z = %cst)
              threads(%tx, %ty, %tz) in (%block_x = %cst, %block_y = %cst, %block_z = %cst) {
     %f0 = arith.constant 0.0: f32
-    %base = arith.constant 0 : index
-    %f = vector.transfer_read %arg0[%base], %f0
-        {permutation_map = affine_map<(d0) -> (d0)>} :
-      memref<?xf32>, vector<4xf32>
+    %base = arith.constant 0 : i32
+    %f = amdgpu.raw_buffer_load { boundsCheck = false } %arg0[%base]
+      : memref<?xf32>, i32 -> vector<4xf32>
 
     %c = arith.addf %f, %f : vector<4xf32>
 
-    vector.transfer_write %c, %arg1[%base]
-        {permutation_map = affine_map<(d0) -> (d0)>} :
-      vector<4xf32>, memref<?xf32>
+    amdgpu.raw_buffer_store { boundsCheck = false } %c -> %arg1[%base]
+      : vector<4xf32> -> memref<?xf32>, i32
 
     gpu.terminator
   }
   return
 }
 
 func.func @main() {
   %c0 = arith.constant 0 : index

utils/bazel/llvm-project-overlay/mlir/BUILD.bazel

deps = [
         ":TensorToLinalg",
         ":TensorToSPIRV",
         ":TosaToArith",
         ":TosaToLinalg",
         ":TosaToSCF",
         ":TosaToTensor",
         ":VectorToGPU",
         ":VectorToLLVM",
-        ":VectorToROCDL",
         ":VectorToSCF",
         ":VectorToSPIRV",
     ],
 )
 
 cc_library(
     name = "AsyncToLLVM",
     srcs = glob([
deps = [
         ":NVVMDialect",
         ":Pass",
         ":Transforms",
         "//llvm:Support",
     ],
 )
 
 cc_library(
-    name = "VectorToROCDL",
-    srcs = [
-        "lib/Conversion/VectorToROCDL/VectorToROCDL.cpp",
-        ":ConversionPassDetail",
-    ],
-    hdrs = ["include/mlir/Conversion/VectorToROCDL/VectorToROCDL.h"],
-    includes = ["include"],
-    deps = [
-        ":ConversionPassIncGen",
-        ":FuncDialect",
-        ":FuncToLLVM",
-        ":GPUDialect",
-        ":IR",
-        ":LLVMCommonConversion",
-        ":LLVMDialect",
-        ":MemRefToLLVM",
-        ":Pass",
-        ":ROCDLDialect",
-        ":Transforms",
-        ":VectorDialect",
-    ],
-)
-
-cc_library(
     name = "VectorToSPIRV",
     srcs = glob([
         "lib/Conversion/VectorToSPIRV/*.cpp",
         "lib/Conversion/VectorToSPIRV/*.h",
     ]) + [":ConversionPassDetail"],
     hdrs = glob([
         "include/mlir/Conversion/VectorToSPIRV/*.h",
     ]),
deps = [
         ":LLVMCommonConversion",
         ":MathDialect",
         ":MemRefToLLVM",
         ":Pass",
         ":ROCDLDialect",
         ":Transforms",
         ":VectorDialect",
         ":VectorToLLVM",
-        ":VectorToROCDL",
         ":VectorToSCF",
         "//llvm:Support",
     ],
 )
 
 cc_library(
     name = "GPUToVulkanTransforms",
     srcs = [
deps = [
         ":TosaDialect",
         ":TosaToLinalg",
         ":TransformDialect",
         ":TransformDialectTransforms",
         ":Transforms",
         ":TransformsPassIncGen",
         ":VectorDialect",
         ":VectorToLLVM",
-        ":VectorToROCDL",
         ":VectorToSCF",
         ":VectorToSPIRV",
         ":VectorTransforms",
         ":X86VectorDialect",
         ":X86VectorTransforms",
     ],
 )