[mlir] [VectorOps] Improve SIMD compares with narrower indices

[aartbik]

Summary

When allowed, use 32-bit indices rather than 64-bit indices in the
SIMD computation of masks. This runs up to 2x and 4x faster on
a number of AVX2 and AVX512 microbenchmarks.

[bkramer]

Looks good. Would it make sense to automatically enable this if the incoming memref is known to have fewer than 2^32 elements?

mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp

127

clang-tidy rightfully complains about the name not being camelBack :)

[aartbik]

rebased

[aartbik]

In D87116#2255567, @bkramer wrote:

Looks good. Would it make sense to automatically enable this if the incoming memref is known to have fewer than 2^32 elements?

I was pondering over this. Would a 32-bit index space in general suffice?
For now I made it an option we can play around with and keep this automatic optimization in mind.

mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp

127

you are right, I am mixing my styleguides

[aartbik]

fixed method case, thanks Ben!

[nicolasvasilache]

Thanks for tracking and fixing the perf bug Aart!
Maybe this also influences how aggressive we need to be in splitting in the codegen strategy to get to peak.
Will also be interesting to see the effects on mobile (cc @asaadaldien )

mlir/include/mlir/Conversion/Passes.td

 def ConvertVectorToLLVM : Pass<"convert-vector-to-llvm", "ModuleOp"> {
   let summary = "Lower the operations from the vector dialect into the LLVM "
                 "dialect";
   let constructor = "mlir::createConvertVectorToLLVMPass()";
   let dependentDialects = ["LLVM::LLVMDialect"];
   let options = [
     Option<"reassociateFPReductions", "reassociate-fp-reductions",
            "bool", /*default=*/"false",
-           "Allows llvm to reassociate floating-point reductions for speed">
+           "Allows llvm to reassociate floating-point reductions for speed">,
+    Option<"enableIndexOptimizations", "enable-index-optimizations",
+           "bool", /*default=*/"false",
+           "Allows compiler to assume indices fit in 32-bit if that yields faster code">
   ];
 }
 
 //===----------------------------------------------------------------------===//
 // 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"];
 }
 
 #endif // MLIR_CONVERSION_PASSES

mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h

 class OperationPass;
 
 /// Options to control Vector to LLVM lowering.
 ///
 /// This should kept in sync with VectorToLLVM options defined for the
 /// ConvertVectorToLLVM pass in include/mlir/Conversion/Passes.td
 struct LowerVectorToLLVMOptions {
   bool reassociateFPReductions = false;
-  LowerVectorToLLVMOptions &setReassociateFPReductions(bool r) {
-    reassociateFPReductions = r;
+  bool enableIndexOptimizations = false;
+  LowerVectorToLLVMOptions &setReassociateFPReductions(bool b) {
+    reassociateFPReductions = b;
+    return *this;
+  }
+  LowerVectorToLLVMOptions &setEnableIndexOptimizations(bool b) {
+    enableIndexOptimizations = b;
     return *this;
   }
 };
 
 /// Collect a set of patterns to convert from Vector contractions to LLVM Matrix
 /// Intrinsics. To lower to assembly, the LLVM flag -lower-matrix-intrinsics
 /// will be needed when invoking LLVM.
 void populateVectorToLLVMMatrixConversionPatterns(
     LLVMTypeConverter &converter, OwningRewritePatternList &patterns);
 
 /// Collect a set of patterns to convert from the Vector dialect to LLVM.
 void populateVectorToLLVMConversionPatterns(
     LLVMTypeConverter &converter, OwningRewritePatternList &patterns,
-    bool reassociateFPReductions = false);
+    bool reassociateFPReductions = false,
+    bool enableIndexOptimizations = false);
 
 /// Create a pass to convert vector operations to the LLVMIR dialect.
 std::unique_ptr<OperationPass<ModuleOp>> createConvertVectorToLLVMPass(
     const LowerVectorToLLVMOptions &options = LowerVectorToLLVMOptions());
 
 } // namespace mlir
 
 #endif // MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVM_H_

mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp

static SmallVector<int64_t, 4> getI64SubArray(ArrayAttr arrayAttr,
   SmallVector<int64_t, 4> res;
   res.reserve(arrayAttr.size() - dropFront - dropBack);
   for (auto it = range.begin() + dropFront, eit = range.end() - dropBack;
        it != eit; ++it)
     res.push_back((*it).getValue().getSExtValue());
   return res;
 }
 
+// Helper that returns a vector comparison that constructs a mask:
+//     mask = [0,1,..,n-1] + [o,o,..,o] < [b,b,..,b]
+//
+// NOTE: The LLVM::GetActiveLaneMaskOp intrinsic would provide an alternative,
+//       much more compact, IR for this operation, but LLVM eventually
+//       generates more elaborate instructions for this intrinsic since it
+//       is very conservative on the boundary conditions.
+static Value BuildVectorComparison(ConversionPatternRewriter &rewriter,
+                                   Operation *op, bool enableIndexOptimizations,
+                                   int64_t dim, Value b, Value *off = nullptr) {
+  auto loc = op->getLoc();
+  // If we can assume all indices fit in 32-bit, we perform the vector
+  // comparison in 32-bit to get a higher degree of SIMD parallelism.
+  // Otherwise we perform the vector comparison using 64-bit indices.
+  Value indices;
+  Type idxType;
+  if (enableIndexOptimizations) {
+    SmallVector<int32_t, 4> values(dim);
+    for (int64_t d = 0; d < dim; d++)
+      values[d] = d;
+    indices =
+        rewriter.create<ConstantOp>(loc, rewriter.getI32VectorAttr(values));
+    idxType = rewriter.getI32Type();
+  } else {
+    SmallVector<int64_t, 4> values(dim);
+    for (int64_t d = 0; d < dim; d++)
+      values[d] = d;
+    indices =
+        rewriter.create<ConstantOp>(loc, rewriter.getI64VectorAttr(values));
+    idxType = rewriter.getI64Type();
+  }
+  // Add in an offset if requested.
+  if (off) {
+    Value o = rewriter.create<IndexCastOp>(loc, idxType, *off);
+    Value ov = rewriter.create<SplatOp>(loc, indices.getType(), o);
+    indices = rewriter.create<AddIOp>(loc, ov, indices);
+  }
+  // Construct the vector comparison.
+  Value bound = rewriter.create<IndexCastOp>(loc, idxType, b);
+  Value bounds = rewriter.create<SplatOp>(loc, indices.getType(), bound);
+  return rewriter.create<CmpIOp>(loc, CmpIPredicate::slt, indices, bounds);
+}
+
 // Helper that returns data layout alignment of an operation with memref.
 template <typename T>
 LogicalResult getMemRefAlignment(LLVMTypeConverter &typeConverter, T op,
                                  unsigned &align) {
   Type elementTy =
       typeConverter.convertType(op.getMemRefType().getElementType());
   if (!elementTy)
     return failure();
public:
   }
 };
 
 /// Conversion pattern for all vector reductions.
 class VectorReductionOpConversion : public ConvertToLLVMPattern {
 public:
   explicit VectorReductionOpConversion(MLIRContext *context,
                                        LLVMTypeConverter &typeConverter,
-                                       bool reassociateFP)
+                                       bool reassociateFPRed)
       : ConvertToLLVMPattern(vector::ReductionOp::getOperationName(), context,
                              typeConverter),
-        reassociateFPReductions(reassociateFP) {}
+        reassociateFPReductions(reassociateFPRed) {}
 
   LogicalResult
   matchAndRewrite(Operation *op, ArrayRef<Value> operands,
                   ConversionPatternRewriter &rewriter) const override {
     auto reductionOp = cast<vector::ReductionOp>(op);
     auto kind = reductionOp.kind();
     Type eltType = reductionOp.dest().getType();
     Type llvmType = typeConverter.convertType(eltType);
matchAndRewrite(Operation *op, ArrayRef<Value> operands,
     }
     return failure();
   }
 
 private:
   const bool reassociateFPReductions;
 };
 
+/// Conversion pattern for a vector.create_mask (1-D only).
+class VectorCreateMaskOpConversion : public ConvertToLLVMPattern {
+public:
+  explicit VectorCreateMaskOpConversion(MLIRContext *context,
+                                        LLVMTypeConverter &typeConverter,
+                                        bool enableIndexOpt)
+      : ConvertToLLVMPattern(vector::CreateMaskOp::getOperationName(), context,
+                             typeConverter),
+        enableIndexOptimizations(enableIndexOpt) {}
+
+  LogicalResult
+  matchAndRewrite(Operation *op, ArrayRef<Value> operands,
+                  ConversionPatternRewriter &rewriter) const override {
+    auto dstType = op->getResult(0).getType().cast<VectorType>();
+    int64_t rank = dstType.getRank();
+    if (rank == 1) {
+      rewriter.replaceOp(
+          op, BuildVectorComparison(rewriter, op, enableIndexOptimizations,
+                                    dstType.getDimSize(0), operands[0]));
+      return success();
+    }
+    return failure();
+  }
+
+private:
+  const bool enableIndexOptimizations;
+};
+
 class VectorShuffleOpConversion : public ConvertToLLVMPattern {
 public:
   explicit VectorShuffleOpConversion(MLIRContext *context,
                                      LLVMTypeConverter &typeConverter)
       : ConvertToLLVMPattern(vector::ShuffleOp::getOperationName(), context,
                              typeConverter) {}
 
   LogicalResult
matchAndRewrite(Operation *op, ArrayRef<Value> operands,
 
     rewriter.replaceOp(op, {desc});
     return success();
   }
 };
 
 /// Conversion pattern that converts a 1-D vector transfer read/write op in a
 /// sequence of:
-/// 1. Bitcast or addrspacecast to vector form.
-/// 2. Create an offsetVector = [ offset + 0 .. offset + vector_length - 1 ].
-/// 3. Create a mask where offsetVector is compared against memref upper bound.
-/// 4. Rewrite op as a masked read or write.
+/// 1. Get the source/dst address as an LLVM vector pointer.
+/// 2. Create a vector with linear indices [ 0 .. vector_length - 1 ].
+/// 3. Create an offsetVector = [ offset + 0 .. offset + vector_length - 1 ].
+/// 4. Create a mask where offsetVector is compared against memref upper bound.
+/// 5. Rewrite op as a masked read or write.
 template <typename ConcreteOp>
 class VectorTransferConversion : public ConvertToLLVMPattern {
 public:
   explicit VectorTransferConversion(MLIRContext *context,
-                                    LLVMTypeConverter &typeConv)
-      : ConvertToLLVMPattern(ConcreteOp::getOperationName(), context,
-                             typeConv) {}
+                                    LLVMTypeConverter &typeConv,
+                                    bool enableIndexOpt)
+      : ConvertToLLVMPattern(ConcreteOp::getOperationName(), context, typeConv),
+        enableIndexOptimizations(enableIndexOpt) {}
 
   LogicalResult
   matchAndRewrite(Operation *op, ArrayRef<Value> operands,
                   ConversionPatternRewriter &rewriter) const override {
     auto xferOp = cast<ConcreteOp>(op);
     auto adaptor = getTransferOpAdapter(xferOp, operands);
 
     if (xferOp.getVectorType().getRank() > 1 ||
         llvm::size(xferOp.indices()) == 0)
       return failure();
     if (xferOp.permutation_map() !=
         AffineMap::getMinorIdentityMap(xferOp.permutation_map().getNumInputs(),
                                        xferOp.getVectorType().getRank(),
                                        op->getContext()))
       return failure();
     // Only contiguous source tensors supported atm.
     SmallVector<int64_t, 4> strides;
     if (!isContiguous(xferOp.getMemRefType(), strides))
       return failure();
 
     auto toLLVMTy = [&](Type t) { return typeConverter.convertType(t); };
 
     Location loc = op->getLoc();
-    Type i64Type = rewriter.getIntegerType(64);
     MemRefType memRefType = xferOp.getMemRefType();
 
     if (auto memrefVectorElementType =
             memRefType.getElementType().dyn_cast<VectorType>()) {
       // Memref has vector element type.
       if (memrefVectorElementType.getElementType() !=
           xferOp.getVectorType().getElementType())
         return failure();
else
       vectorDataPtr = rewriter.create<LLVM::AddrSpaceCastOp>(
           loc, vecTy.getPointerTo(), dataPtr);
 
     if (!xferOp.isMaskedDim(0))
       return replaceTransferOpWithLoadOrStore(rewriter, typeConverter, loc,
                                               xferOp, operands, vectorDataPtr);
 
     // 2. Create a vector with linear indices [ 0 .. vector_length - 1 ].
-    unsigned vecWidth = vecTy.getVectorNumElements();
-    VectorType vectorCmpType = VectorType::get(vecWidth, i64Type);
-    SmallVector<int64_t, 8> indices;
-    indices.reserve(vecWidth);
-    for (unsigned i = 0; i < vecWidth; ++i)
-      indices.push_back(i);
-    Value linearIndices = rewriter.create<ConstantOp>(
-        loc, vectorCmpType,
-        DenseElementsAttr::get(vectorCmpType, ArrayRef<int64_t>(indices)));
-    linearIndices = rewriter.create<LLVM::DialectCastOp>(
-        loc, toLLVMTy(vectorCmpType), linearIndices);
-
     // 3. Create offsetVector = [ offset + 0 .. offset + vector_length - 1 ].
-    // TODO: when the leaf transfer rank is k > 1 we need the last
-    // `k` dimensions here.
-    unsigned lastIndex = llvm::size(xferOp.indices()) - 1;
-    Value offsetIndex = *(xferOp.indices().begin() + lastIndex);
-    offsetIndex = rewriter.create<IndexCastOp>(loc, i64Type, offsetIndex);
-    Value base = rewriter.create<SplatOp>(loc, vectorCmpType, offsetIndex);
-    Value offsetVector = rewriter.create<AddIOp>(loc, base, linearIndices);
-
     // 4. Let dim the memref dimension, compute the vector comparison mask:
     //   [ offset + 0 .. offset + vector_length - 1 ] < [ dim .. dim ]
+    //
+    // TODO: when the leaf transfer rank is k > 1, we need the last `k`
+    //       dimensions here.
+    unsigned vecWidth = vecTy.getVectorNumElements();
+    unsigned lastIndex = llvm::size(xferOp.indices()) - 1;
+    Value off = *(xferOp.indices().begin() + lastIndex);
     Value dim = rewriter.create<DimOp>(loc, xferOp.memref(), lastIndex);
-    dim = rewriter.create<IndexCastOp>(loc, i64Type, dim);
-    dim = rewriter.create<SplatOp>(loc, vectorCmpType, dim);
-    Value mask =
-        rewriter.create<CmpIOp>(loc, CmpIPredicate::slt, offsetVector, dim);
-    mask = rewriter.create<LLVM::DialectCastOp>(loc, toLLVMTy(mask.getType()),
-                                                mask);
+    Value mask = BuildVectorComparison(rewriter, op, enableIndexOptimizations,
+                                       vecWidth, dim, &off);
 
     // 5. Rewrite as a masked read / write.
     return replaceTransferOpWithMasked(rewriter, typeConverter, loc, xferOp,
                                        operands, vectorDataPtr, mask);
   }
+
+private:
+  const bool enableIndexOptimizations;
 };
 
 class VectorPrintOpConversion : public ConvertToLLVMPattern {
 public:
   explicit VectorPrintOpConversion(MLIRContext *context,
                                    LLVMTypeConverter &typeConverter)
       : ConvertToLLVMPattern(vector::PrintOp::getOperationName(), context,
                              typeConverter) {}
public:
   bool hasBoundedRewriteRecursion() const final { return true; }
 };
 
 } // namespace
 
 /// Populate the given list with patterns that convert from Vector to LLVM.
 void mlir::populateVectorToLLVMConversionPatterns(
     LLVMTypeConverter &converter, OwningRewritePatternList &patterns,
-    bool reassociateFPReductions) {
+    bool reassociateFPReductions, bool enableIndexOptimizations) {
   MLIRContext *ctx = converter.getDialect()->getContext();
   // clang-format off
   patterns.insert<VectorFMAOpNDRewritePattern,
                   VectorInsertStridedSliceOpDifferentRankRewritePattern,
                   VectorInsertStridedSliceOpSameRankRewritePattern,
                   VectorExtractStridedSliceOpConversion>(ctx);
   patterns.insert<VectorReductionOpConversion>(
       ctx, converter, reassociateFPReductions);
+  patterns.insert<VectorCreateMaskOpConversion,
+                  VectorTransferConversion<TransferReadOp>,
+                  VectorTransferConversion<TransferWriteOp>>(
+      ctx, converter, enableIndexOptimizations);
   patterns
       .insert<VectorShuffleOpConversion,
               VectorExtractElementOpConversion,
               VectorExtractOpConversion,
               VectorFMAOp1DConversion,
               VectorInsertElementOpConversion,
               VectorInsertOpConversion,
               VectorPrintOpConversion,
-              VectorTransferConversion<TransferReadOp>,
-              VectorTransferConversion<TransferWriteOp>,
               VectorTypeCastOpConversion,
               VectorMaskedLoadOpConversion,
               VectorMaskedStoreOpConversion,
               VectorGatherOpConversion,
               VectorScatterOpConversion,
               VectorExpandLoadOpConversion,
               VectorCompressStoreOpConversion>(ctx, converter);
   // clang-format on
 }
 
 void mlir::populateVectorToLLVMMatrixConversionPatterns(
     LLVMTypeConverter &converter, OwningRewritePatternList &patterns) {
   MLIRContext *ctx = converter.getDialect()->getContext();
   patterns.insert<VectorMatmulOpConversion>(ctx, converter);
   patterns.insert<VectorFlatTransposeOpConversion>(ctx, converter);
 }
 
 namespace {
 struct LowerVectorToLLVMPass
     : public ConvertVectorToLLVMBase<LowerVectorToLLVMPass> {
   LowerVectorToLLVMPass(const LowerVectorToLLVMOptions &options) {
     this->reassociateFPReductions = options.reassociateFPReductions;
+    this->enableIndexOptimizations = options.enableIndexOptimizations;
   }
   void runOnOperation() override;
 };
 } // namespace
 
 void LowerVectorToLLVMPass::runOnOperation() {
   // Perform progressive lowering of operations on slices and
   // all contraction operations. Also applies folding and DCE.
   {
     OwningRewritePatternList patterns;
     populateVectorToVectorCanonicalizationPatterns(patterns, &getContext());
     populateVectorSlicesLoweringPatterns(patterns, &getContext());
     populateVectorContractLoweringPatterns(patterns, &getContext());
     applyPatternsAndFoldGreedily(getOperation(), patterns);
   }
 
   // Convert to the LLVM IR dialect.
   LLVMTypeConverter converter(&getContext());
   OwningRewritePatternList patterns;
   populateVectorToLLVMMatrixConversionPatterns(converter, patterns);
-  populateVectorToLLVMConversionPatterns(converter, patterns,
-                                         reassociateFPReductions);
+  populateVectorToLLVMConversionPatterns(
+      converter, patterns, reassociateFPReductions, enableIndexOptimizations);
   populateVectorToLLVMMatrixConversionPatterns(converter, patterns);
   populateStdToLLVMConversionPatterns(converter, patterns);
 
   LLVMConversionTarget target(getContext());
-  if (failed(applyPartialConversion(getOperation(), target, patterns))) {
+  if (failed(applyPartialConversion(getOperation(), target, patterns)))
     signalPassFailure();
-  }
-}
 
 std::unique_ptr<OperationPass<ModuleOp>>
 mlir::createConvertVectorToLLVMPass(const LowerVectorToLLVMOptions &options) {
   return std::make_unique<LowerVectorToLLVMPass>(options);
 }

mlir/lib/Dialect/Vector/VectorTransforms.cpp

public:
 
   LogicalResult matchAndRewrite(vector::ConstantMaskOp op,
                                 PatternRewriter &rewriter) const override {
     auto loc = op.getLoc();
     auto dstType = op.getResult().getType().cast<VectorType>();
     auto eltType = dstType.getElementType();
     auto dimSizes = op.mask_dim_sizes();
     int64_t rank = dimSizes.size();
-    int64_t trueDim = dimSizes[0].cast<IntegerAttr>().getInt();
+    int64_t trueDim = std::min(dstType.getDimSize(0),
+                               dimSizes[0].cast<IntegerAttr>().getInt());
 
     if (rank == 1) {
       // Express constant 1-D case in explicit vector form:
       //   [T,..,T,F,..,F].
       SmallVector<bool, 4> values(dstType.getDimSize(0));
       for (int64_t d = 0; d < trueDim; d++)
         values[d] = true;
       rewriter.replaceOpWithNewOp<ConstantOp>(
LogicalResult matchAndRewrite(vector::CreateMaskOp op,
                                 PatternRewriter &rewriter) const override {
     auto loc = op.getLoc();
     auto dstType = op.getResult().getType().cast<VectorType>();
     auto eltType = dstType.getElementType();
     int64_t dim = dstType.getDimSize(0);
     int64_t rank = dstType.getRank();
     Value idx = op.getOperand(0);
 
-    if (rank == 1) {
-      // Express dynamic 1-D case in explicit vector form:
-      //   mask = [0,1,..,n-1] < [a,a,..,a]
-      SmallVector<int64_t, 4> values(dim);
-      for (int64_t d = 0; d < dim; d++)
-        values[d] = d;
-      Value indices =
-          rewriter.create<ConstantOp>(loc, rewriter.getI64VectorAttr(values));
-      Value bound =
-          rewriter.create<IndexCastOp>(loc, rewriter.getI64Type(), idx);
-      Value bounds = rewriter.create<SplatOp>(loc, indices.getType(), bound);
-      rewriter.replaceOpWithNewOp<CmpIOp>(op, CmpIPredicate::slt, indices,
-                                          bounds);
-      return success();
-    }
+    if (rank == 1)
+      return failure(); // leave for lowering
 
     VectorType lowType =
         VectorType::get(dstType.getShape().drop_front(), eltType);
     Value trueVal = rewriter.create<vector::CreateMaskOp>(
         loc, lowType, op.getOperands().drop_front());
     Value falseVal = rewriter.create<ConstantOp>(loc, lowType,
                                                  rewriter.getZeroAttr(lowType));
     Value result = rewriter.create<ConstantOp>(loc, dstType,

mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir

+// RUN: mlir-opt %s --convert-vector-to-llvm='enable-index-optimizations=1' | FileCheck %s --check-prefix=CMP32
+// RUN: mlir-opt %s --convert-vector-to-llvm='enable-index-optimizations=0' | FileCheck %s --check-prefix=CMP64
+
+// CMP32-LABEL: llvm.func @genbool_var_1d(
+// CMP32-SAME: %[[A:.*]]: !llvm.i64)
+// CMP32: %[[T0:.*]] = llvm.mlir.constant(dense<[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]> : vector<11xi32>) : !llvm.vec<11 x i32>
+// CMP32: %[[T1:.*]] = llvm.trunc %[[A]] : !llvm.i64 to !llvm.i32
+// CMP32: %[[T2:.*]] = llvm.mlir.undef : !llvm.vec<11 x i32>
+// CMP32: %[[T3:.*]] = llvm.mlir.constant(0 : i32) : !llvm.i32
+// CMP32: %[[T4:.*]] = llvm.insertelement %[[T1]], %[[T2]][%[[T3]] : !llvm.i32] : !llvm.vec<11 x i32>
+// CMP32: %[[T5:.*]] = llvm.shufflevector %[[T4]], %[[T2]] [0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32] : !llvm.vec<11 x i32>, !llvm.vec<11 x i32>
+// CMP32: %[[T6:.*]] = llvm.icmp "slt" %[[T0]], %[[T5]] : !llvm.vec<11 x i32>
+// CMP32: llvm.return %[[T6]] : !llvm.vec<11 x i1>
+
+// CMP64-LABEL: llvm.func @genbool_var_1d(
+// CMP64-SAME: %[[A:.*]]: !llvm.i64)
+// CMP64: %[[T0:.*]] = llvm.mlir.constant(dense<[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]> : vector<11xi64>) : !llvm.vec<11 x i64>
+// CMP64: %[[T1:.*]] = llvm.mlir.undef : !llvm.vec<11 x i64>
+// CMP64: %[[T2:.*]] = llvm.mlir.constant(0 : i32) : !llvm.i32
+// CMP64: %[[T3:.*]] = llvm.insertelement %[[A]], %[[T1]][%[[T2]] : !llvm.i32] : !llvm.vec<11 x i64>
+// CMP64: %[[T4:.*]] = llvm.shufflevector %[[T3]], %[[T1]] [0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32] : !llvm.vec<11 x i64>, !llvm.vec<11 x i64>
+// CMP64: %[[T5:.*]] = llvm.icmp "slt" %[[T0]], %[[T4]] : !llvm.vec<11 x i64>
+// CMP64: llvm.return %[[T5]] : !llvm.vec<11 x i1>
+
+func @genbool_var_1d(%arg0: index) -> vector<11xi1> {
+  %0 = vector.create_mask %arg0 : vector<11xi1>
+  return %0 : vector<11xi1>
+}
+
+// CMP32-LABEL: llvm.func @transfer_read_1d
+// CMP32: %[[C:.*]] = llvm.mlir.constant(dense<[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]> : vector<16xi32>) : !llvm.vec<16 x i32>
+// CMP32: %[[A:.*]] = llvm.add %{{.*}}, %[[C]] : !llvm.vec<16 x i32>
+// CMP32: %[[M:.*]] = llvm.icmp "slt" %[[A]], %{{.*}} : !llvm.vec<16 x i32>
+// CMP32: %[[L:.*]] = llvm.intr.masked.load %{{.*}}, %[[M]], %{{.*}}
+// CMP32: llvm.return %[[L]] : !llvm.vec<16 x float>
+
+// CMP64-LABEL: llvm.func @transfer_read_1d
+// CMP64: %[[C:.*]] = llvm.mlir.constant(dense<[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]> : vector<16xi64>) : !llvm.vec<16 x i64>
+// CMP64: %[[A:.*]] = llvm.add %{{.*}}, %[[C]] : !llvm.vec<16 x i64>
+// CMP64: %[[M:.*]] = llvm.icmp "slt" %[[A]], %{{.*}} : !llvm.vec<16 x i64>
+// CMP64: %[[L:.*]] = llvm.intr.masked.load %{{.*}}, %[[M]], %{{.*}}
+// CMP64: llvm.return %[[L]] : !llvm.vec<16 x float>
+
+func @transfer_read_1d(%A : memref<?xf32>, %i: index) -> vector<16xf32> {
+  %d = constant -1.0: f32
+  %f = vector.transfer_read %A[%i], %d {permutation_map = affine_map<(d0) -> (d0)>} : memref<?xf32>, vector<16xf32>
+  return %f : vector<16xf32>
+}

mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir

 // CHECK-LABEL: func @transfer_read_1d
 //  CHECK-SAME: %[[BASE:[a-zA-Z0-9]*]]: !llvm.i64) -> !llvm.vec<17 x float>
 //
 // 1. Bitcast to vector form.
 //       CHECK: %[[gep:.*]] = llvm.getelementptr {{.*}} :
 //  CHECK-SAME: (!llvm.ptr<float>, !llvm.i64) -> !llvm.ptr<float>
 //       CHECK: %[[vecPtr:.*]] = llvm.bitcast %[[gep]] :
 //  CHECK-SAME: !llvm.ptr<float> to !llvm.ptr<vec<17 x float>>
+//       CHECK: %[[DIM:.*]] = llvm.extractvalue %{{.*}}[3, 0] :
+//  CHECK-SAME: !llvm.struct<(ptr<float>, ptr<float>, i64, array<1 x i64>, array<1 x i64>)>
 //
 // 2. Create a vector with linear indices [ 0 .. vector_length - 1 ].
-//       CHECK: %[[linearIndex:.*]] = llvm.mlir.constant(
-//  CHECK-SAME: dense<[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]> :
+//       CHECK: %[[linearIndex:.*]] = llvm.mlir.constant(dense
+//  CHECK-SAME: <[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]> :
 //  CHECK-SAME: vector<17xi64>) : !llvm.vec<17 x i64>
 //
 // 3. Create offsetVector = [ offset + 0 .. offset + vector_length - 1 ].
 //       CHECK: %[[offsetVec:.*]] = llvm.mlir.undef : !llvm.vec<17 x i64>
 //       CHECK: %[[c0:.*]] = llvm.mlir.constant(0 : i32) : !llvm.i32
 //       CHECK: %[[offsetVec2:.*]] = llvm.insertelement %[[BASE]], %[[offsetVec]][%[[c0]] :
 //  CHECK-SAME: !llvm.i32] : !llvm.vec<17 x i64>
 //       CHECK: %[[offsetVec3:.*]] = llvm.shufflevector %[[offsetVec2]], %{{.*}} [
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32] :
 //  CHECK-SAME: !llvm.vec<17 x i64>, !llvm.vec<17 x i64>
 //       CHECK: %[[offsetVec4:.*]] = llvm.add %[[offsetVec3]], %[[linearIndex]] :
 //  CHECK-SAME: !llvm.vec<17 x i64>
 //
 // 4. Let dim the memref dimension, compute the vector comparison mask:
 //    [ offset + 0 .. offset + vector_length - 1 ] < [ dim .. dim ]
-//       CHECK: %[[DIM:.*]] = llvm.extractvalue %{{.*}}[3, 0] :
-//  CHECK-SAME: !llvm.struct<(ptr<float>, ptr<float>, i64, array<1 x i64>, array<1 x i64>)>
 //       CHECK: %[[dimVec:.*]] = llvm.mlir.undef : !llvm.vec<17 x i64>
 //       CHECK: %[[c01:.*]] = llvm.mlir.constant(0 : i32) : !llvm.i32
 //       CHECK: %[[dimVec2:.*]] = llvm.insertelement %[[DIM]], %[[dimVec]][%[[c01]] :
 //  CHECK-SAME:  !llvm.i32] : !llvm.vec<17 x i64>
 //       CHECK: %[[dimVec3:.*]] = llvm.shufflevector %[[dimVec2]], %{{.*}} [
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32] :

 //
 // 1. Bitcast to vector form.
 //       CHECK: %[[gep_b:.*]] = llvm.getelementptr {{.*}} :
 //  CHECK-SAME: (!llvm.ptr<float>, !llvm.i64) -> !llvm.ptr<float>
 //       CHECK: %[[vecPtr_b:.*]] = llvm.bitcast %[[gep_b]] :
 //  CHECK-SAME: !llvm.ptr<float> to !llvm.ptr<vec<17 x float>>
 //
 // 2. Create a vector with linear indices [ 0 .. vector_length - 1 ].
-//       CHECK: %[[linearIndex_b:.*]] = llvm.mlir.constant(
-//  CHECK-SAME: dense<[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]> :
-//  CHECK-SAME:  vector<17xi64>) : !llvm.vec<17 x i64>
+//       CHECK: %[[linearIndex_b:.*]] = llvm.mlir.constant(dense
+//  CHECK-SAME: <[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]> :
+//  CHECK-SAME: vector<17xi64>) : !llvm.vec<17 x i64>
 //
 // 3. Create offsetVector = [ offset + 0 .. offset + vector_length - 1 ].
 //       CHECK: llvm.shufflevector {{.*}} [0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32] :
 //  CHECK-SAME: !llvm.vec<17 x i64>, !llvm.vec<17 x i64>
 //       CHECK: llvm.add
 //
func @transfer_read_2d_to_1d(%A : memref<?x?xf32>, %base0: index, %base1: index) -> vector<17xf32> {
   %f7 = constant 7.0: f32
   %f = vector.transfer_read %A[%base0, %base1], %f7
       {permutation_map = affine_map<(d0, d1) -> (d1)>} :
     memref<?x?xf32>, vector<17xf32>
   return %f: vector<17xf32>
 }
 // CHECK-LABEL: func @transfer_read_2d_to_1d
 //  CHECK-SAME: %[[BASE_0:[a-zA-Z0-9]*]]: !llvm.i64, %[[BASE_1:[a-zA-Z0-9]*]]: !llvm.i64) -> !llvm.vec<17 x float>
+//       CHECK: %[[DIM:.*]] = llvm.extractvalue %{{.*}}[3, 1] :
+//  CHECK-SAME: !llvm.struct<(ptr<float>, ptr<float>, i64, array<2 x i64>, array<2 x i64>)>
 //
 // Create offsetVector = [ offset + 0 .. offset + vector_length - 1 ].
 //       CHECK: %[[offsetVec:.*]] = llvm.mlir.undef : !llvm.vec<17 x i64>
 //       CHECK: %[[c0:.*]] = llvm.mlir.constant(0 : i32) : !llvm.i32
 // Here we check we properly use %BASE_1
 //       CHECK: %[[offsetVec2:.*]] = llvm.insertelement %[[BASE_1]], %[[offsetVec]][%[[c0]] :
 //  CHECK-SAME: !llvm.i32] : !llvm.vec<17 x i64>
 //       CHECK: %[[offsetVec3:.*]] = llvm.shufflevector %[[offsetVec2]], %{{.*}} [
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32] :
 //
 // Let dim the memref dimension, compute the vector comparison mask:
 //    [ offset + 0 .. offset + vector_length - 1 ] < [ dim .. dim ]
 // Here we check we properly use %DIM[1]
-//       CHECK: %[[DIM:.*]] = llvm.extractvalue %{{.*}}[3, 1] :
-//  CHECK-SAME: !llvm.struct<(ptr<float>, ptr<float>, i64, array<2 x i64>, array<2 x i64>)>
 //       CHECK: %[[dimVec:.*]] = llvm.mlir.undef : !llvm.vec<17 x i64>
 //       CHECK: %[[c01:.*]] = llvm.mlir.constant(0 : i32) : !llvm.i32
 //       CHECK: %[[dimVec2:.*]] = llvm.insertelement %[[DIM]], %[[dimVec]][%[[c01]] :
 //  CHECK-SAME:  !llvm.i32] : !llvm.vec<17 x i64>
 //       CHECK: %[[dimVec3:.*]] = llvm.shufflevector %[[dimVec2]], %{{.*}} [
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32, 0 : i32,
 //  CHECK-SAME:  0 : i32, 0 : i32, 0 : i32] :

mlir/test/Dialect/Vector/vector-contract-transforms.mlir

 // CHECK: %[[T1:.*]] = vector.insert %[[T0]], %[[C3]] [0] : vector<3x4xi1> into vector<2x3x4xi1>
 // CHECK: return %[[T1]] : vector<2x3x4xi1>
 
 func @genbool_3d() -> vector<2x3x4xi1> {
   %v = vector.constant_mask [1, 1, 3] : vector<2x3x4xi1>
   return %v: vector<2x3x4xi1>
 }
 
-// CHECK-LABEL: func @genbool_var_1d
-// CHECK-SAME: %[[A:.*]]: index
-// CHECK:      %[[C1:.*]] = constant dense<[0, 1, 2]> : vector<3xi64>
-// CHECK:      %[[T0:.*]] = index_cast %[[A]] : index to i64
-// CHECK:      %[[T1:.*]] = splat %[[T0]] : vector<3xi64>
-// CHECK:      %[[T2:.*]] = cmpi "slt", %[[C1]], %[[T1]] : vector<3xi64>
-// CHECK:      return %[[T2]] : vector<3xi1>
+// CHECK-LABEL: func @genbool_var_1d(
+// CHECK-SAME: %[[A:.*]]: index)
+// CHECK:      %[[T0:.*]] = vector.create_mask %[[A]] : vector<3xi1>
+// CHECK:      return %[[T0]] : vector<3xi1>
 
 func @genbool_var_1d(%arg0: index) -> vector<3xi1> {
   %0 = vector.create_mask %arg0 : vector<3xi1>
   return %0 : vector<3xi1>
 }
 
-// CHECK-LABEL: func @genbool_var_2d
-// CHECK-SAME: %[[A:.*0]]: index
-// CHECK-SAME: %[[B:.*1]]: index
-// CHECK:      %[[CI:.*]] = constant dense<[0, 1, 2]> : vector<3xi64>
-// CHECK:      %[[CF:.*]] = constant dense<false> : vector<3xi1>
+// CHECK-LABEL: func @genbool_var_2d(
+// CHECK-SAME: %[[A:.*0]]: index,
+// CHECK-SAME: %[[B:.*1]]: index)
+// CHECK:      %[[C1:.*]] = constant dense<false> : vector<3xi1>
 // CHECK:      %[[C2:.*]] = constant dense<false> : vector<2x3xi1>
 // CHECK:      %[[c0:.*]] = constant 0 : index
 // CHECK:      %[[c1:.*]] = constant 1 : index
-// CHECK:      %[[T0:.*]] = index_cast %[[B]] : index to i64
-// CHECK:      %[[T1:.*]] = splat %[[T0]] : vector<3xi64>
-// CHECK:      %[[T2:.*]] = cmpi "slt", %[[CI]], %[[T1]] : vector<3xi64>
-// CHECK:      %[[T3:.*]] = cmpi "slt", %[[c0]], %[[A]] : index
-// CHECK:      %[[T4:.*]] = select %[[T3]], %[[T2]], %[[CF]] : vector<3xi1>
-// CHECK:      %[[T5:.*]] = vector.insert %[[T4]], %[[C2]] [0] : vector<3xi1> into vector<2x3xi1>
-// CHECK:      %[[T6:.*]] = cmpi "slt", %[[c1]], %[[A]] : index
-// CHECK:      %[[T7:.*]] = select %[[T6]], %[[T2]], %[[CF]] : vector<3xi1>
-// CHECK:      %[[T8:.*]] = vector.insert %[[T7]], %[[T5]] [1] : vector<3xi1> into vector<2x3xi1>
-// CHECK:      return %[[T8]] : vector<2x3xi1>
+// CHECK:      %[[T0:.*]] = vector.create_mask %[[B]] : vector<3xi1>
+// CHECK:      %[[T1:.*]] = cmpi "slt", %[[c0]], %[[A]] : index
+// CHECK:      %[[T2:.*]] = select %[[T1]], %[[T0]], %[[C1]] : vector<3xi1>
+// CHECK:      %[[T3:.*]] = vector.insert %[[T2]], %[[C2]] [0] : vector<3xi1> into vector<2x3xi1>
+// CHECK:      %[[T4:.*]] = cmpi "slt", %[[c1]], %[[A]] : index
+// CHECK:      %[[T5:.*]] = select %[[T4]], %[[T0]], %[[C1]] : vector<3xi1>
+// CHECK:      %[[T6:.*]] = vector.insert %[[T5]], %[[T3]] [1] : vector<3xi1> into vector<2x3xi1>
+// CHECK:      return %[[T6]] : vector<2x3xi1>
 
 func @genbool_var_2d(%arg0: index, %arg1: index) -> vector<2x3xi1> {
   %0 = vector.create_mask %arg0, %arg1 : vector<2x3xi1>
   return %0 : vector<2x3xi1>
 }
 
+// CHECK-LABEL: func @genbool_var_3d(
+// CHECK-SAME: %[[A:.*0]]: index,
+// CHECK-SAME: %[[B:.*1]]: index,
+// CHECK-SAME: %[[C:.*2]]: index)
+// CHECK:      %[[C1:.*]] = constant dense<false> : vector<7xi1>
+// CHECK:      %[[C2:.*]] = constant dense<false> : vector<1x7xi1>
+// CHECK:      %[[C3:.*]] = constant dense<false> : vector<2x1x7xi1>
+// CHECK:      %[[c0:.*]] = constant 0 : index
+// CHECK:      %[[c1:.*]] = constant 1 : index
+// CHECK:      %[[T0:.*]] = vector.create_mask %[[C]] : vector<7xi1>
+// CHECK:      %[[T1:.*]] = cmpi "slt", %[[c0]], %[[B]] : index
+// CHECK:      %[[T2:.*]] = select %[[T1]], %[[T0]], %[[C1]] : vector<7xi1>
+// CHECK:      %[[T3:.*]] = vector.insert %[[T2]], %[[C2]] [0] : vector<7xi1> into vector<1x7xi1>
+// CHECK:      %[[T4:.*]] = cmpi "slt", %[[c0]], %[[A]] : index
+// CHECK:      %[[T5:.*]] = select %[[T4]], %[[T3]], %[[C2]] : vector<1x7xi1>
+// CHECK:      %[[T6:.*]] = vector.insert %[[T5]], %[[C3]] [0] : vector<1x7xi1> into vector<2x1x7xi1>
+// CHECK:      %[[T7:.*]] = cmpi "slt", %[[c1]], %[[A]] : index
+// CHECK:      %[[T8:.*]] = select %[[T7]], %[[T3]], %[[C2]] : vector<1x7xi1>
+// CHECK:      %[[T9:.*]] = vector.insert %[[T8]], %[[T6]] [1] : vector<1x7xi1> into vector<2x1x7xi1>
+// CHECK:      return %[[T9]] : vector<2x1x7xi1>
+
+func @genbool_var_3d(%arg0: index, %arg1: index, %arg2: index) -> vector<2x1x7xi1> {
+  %0 = vector.create_mask %arg0, %arg1, %arg2 : vector<2x1x7xi1>
+  return %0 : vector<2x1x7xi1>
+}
+
 #matmat_accesses_0 = [
   affine_map<(m, n, k) -> (m, k)>,
   affine_map<(m, n, k) -> (k, n)>,
   affine_map<(m, n, k) -> (m, n)>
 ]
 #matmat_trait_0 = {
   indexing_maps = #matmat_accesses_0,
   iterator_types = ["parallel", "parallel", "reduction"]