[Mlir-commits] [mlir] [mlir][Vector] Support `vector.extract(xfer_read)` folding with dynamic indices (PR #143269)

Diego Caballero llvmlistbot at llvm.org
Mon Jun 16 11:59:49 PDT 2025


https://github.com/dcaballe updated https://github.com/llvm/llvm-project/pull/143269

>From 6bc6e3a3b714485db9d8a39c61300603eb31bf68 Mon Sep 17 00:00:00 2001
From: Diego Caballero <dcaballero at nvidia.com>
Date: Sat, 7 Jun 2025 12:45:44 +0000
Subject: [PATCH 1/2] [mlir][Vector] Support `xfer_read(vector.extract))`
 folding with dynamic indices

This PR is part of the step to remove `vector.extractelement` and
`vector.insertelement` ops. It adds support for folding
`vector.transfer_read(vector.extract) -> memref.load` with dynamic
indices, which is currently supported by `vector.extractelement`.
---
 .../Transforms/VectorTransferOpTransforms.cpp | 24 +++++++++++----
 .../scalar-vector-transfer-to-memref.mlir     | 30 +++++++++++++++++++
 2 files changed, 49 insertions(+), 5 deletions(-)

diff --git a/mlir/lib/Dialect/Vector/Transforms/VectorTransferOpTransforms.cpp b/mlir/lib/Dialect/Vector/Transforms/VectorTransferOpTransforms.cpp
index 7dbb7a334fe62..36197eb1caeb1 100644
--- a/mlir/lib/Dialect/Vector/Transforms/VectorTransferOpTransforms.cpp
+++ b/mlir/lib/Dialect/Vector/Transforms/VectorTransferOpTransforms.cpp
@@ -886,12 +886,26 @@ class RewriteScalarExtractOfTransferRead
     SmallVector<Value> newIndices(xferOp.getIndices().begin(),
                                   xferOp.getIndices().end());
     for (auto [i, pos] : llvm::enumerate(extractOp.getMixedPosition())) {
-      assert(isa<Attribute>(pos) && "Unexpected non-constant index");
-      int64_t offset = cast<IntegerAttr>(cast<Attribute>(pos)).getInt();
       int64_t idx = newIndices.size() - extractOp.getNumIndices() + i;
-      OpFoldResult ofr = affine::makeComposedFoldedAffineApply(
-          rewriter, extractOp.getLoc(),
-          rewriter.getAffineSymbolExpr(0) + offset, {newIndices[idx]});
+
+      // Compute affine expression `newIndices[idx] + pos` where `pos` can be
+      // either a constant or a value.
+      OpFoldResult ofr;
+      if (auto attr = dyn_cast<Attribute>(pos)) {
+        int64_t offset = cast<IntegerAttr>(attr).getInt();
+        ofr = affine::makeComposedFoldedAffineApply(
+            rewriter, extractOp.getLoc(),
+            rewriter.getAffineSymbolExpr(0) + offset, {newIndices[idx]});
+      } else {
+        Value dynamicOffset = cast<Value>(pos);
+        AffineExpr sym0, sym1;
+        bindSymbols(rewriter.getContext(), sym0, sym1);
+        ofr = affine::makeComposedFoldedAffineApply(
+            rewriter, extractOp.getLoc(), sym0 + sym1,
+            {newIndices[idx], dynamicOffset});
+      }
+
+      // Update the corresponding index with the folded result.
       if (auto value = dyn_cast<Value>(ofr)) {
         newIndices[idx] = value;
       } else {
diff --git a/mlir/test/Dialect/Vector/scalar-vector-transfer-to-memref.mlir b/mlir/test/Dialect/Vector/scalar-vector-transfer-to-memref.mlir
index 52b0fdee184f6..9f10063a75092 100644
--- a/mlir/test/Dialect/Vector/scalar-vector-transfer-to-memref.mlir
+++ b/mlir/test/Dialect/Vector/scalar-vector-transfer-to-memref.mlir
@@ -148,3 +148,33 @@ func.func @subvector_extract(%m: memref<?x?xf32>, %idx: index) -> vector<16xf32>
   return %1 : vector<16xf32>
 }
 
+// -----
+
+//       CHECK: #[[$MAP:.*]] = affine_map<()[s0, s1] -> (s0 + s1)>
+// CHECK-LABEL: func @transfer_read_1d_extract_dynamic(
+//  CHECK-SAME:     %[[MEMREF:.*]]: memref<?xf32>, %[[M_IDX:.*]]: index, %[[E_IDX:.*]]: index
+//       CHECK:   %[[APPLY:.*]] = affine.apply #[[$MAP]]()[%[[M_IDX]], %[[E_IDX]]]
+//       CHECK:   %[[RES:.*]] = memref.load %[[MEMREF]][%[[APPLY]]]
+func.func @transfer_read_1d_extract_dynamic(%m: memref<?xf32>, %idx: index,
+                                            %offset: index) -> f32 {
+  %cst = arith.constant 0.0 : f32
+  %vec = vector.transfer_read %m[%idx], %cst {in_bounds = [true]} : memref<?xf32>, vector<5xf32>
+  %elem = vector.extract %vec[%offset] : f32 from vector<5xf32>
+  return %elem : f32
+}
+
+// -----
+
+//       CHECK: #[[$MAP:.*]] = affine_map<()[s0, s1] -> (s0 + s1)>
+// CHECK-LABEL: func @transfer_read_2d_extract_dynamic(
+//  CHECK-SAME:     %[[MEMREF:.*]]: memref<?x?xf32>, %[[M_IDX:.*]]: index, %[[ROW:.*]]: index, %[[COL:.*]]: index
+//       CHECK:   %[[ROW_APPLY:.*]] = affine.apply #[[$MAP]]()[%[[M_IDX]], %[[ROW]]]
+//       CHECK:   %[[COL_APPLY:.*]] = affine.apply #[[$MAP]]()[%[[M_IDX]], %[[COL]]]
+//       CHECK:   %[[RES:.*]] = memref.load %[[MEMREF]][%[[ROW_APPLY]], %[[COL_APPLY]]]
+func.func @transfer_read_2d_extract_dynamic(%m: memref<?x?xf32>, %idx: index,
+                                            %row_offset: index, %col_offset: index) -> f32 {
+  %cst = arith.constant 0.0 : f32
+  %vec = vector.transfer_read %m[%idx, %idx], %cst {in_bounds = [true, true]} : memref<?x?xf32>, vector<10x5xf32>
+  %elem = vector.extract %vec[%row_offset, %col_offset] : f32 from vector<10x5xf32>
+  return %elem : f32
+}

>From f98e3480519924dd8f72b4155b3bc5b2807ecc02 Mon Sep 17 00:00:00 2001
From: Diego Caballero <dcaballero at nvidia.com>
Date: Mon, 16 Jun 2025 18:58:02 +0000
Subject: [PATCH 2/2] Feedback

---
 .../Vector/Transforms/VectorTransferOpTransforms.cpp   | 10 +++++-----
 .../Vector/scalar-vector-transfer-to-memref.mlir       | 10 +++++-----
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/mlir/lib/Dialect/Vector/Transforms/VectorTransferOpTransforms.cpp b/mlir/lib/Dialect/Vector/Transforms/VectorTransferOpTransforms.cpp
index 36197eb1caeb1..384717aeca665 100644
--- a/mlir/lib/Dialect/Vector/Transforms/VectorTransferOpTransforms.cpp
+++ b/mlir/lib/Dialect/Vector/Transforms/VectorTransferOpTransforms.cpp
@@ -890,27 +890,27 @@ class RewriteScalarExtractOfTransferRead
 
       // Compute affine expression `newIndices[idx] + pos` where `pos` can be
       // either a constant or a value.
-      OpFoldResult ofr;
+      OpFoldResult composedIdx;
       if (auto attr = dyn_cast<Attribute>(pos)) {
         int64_t offset = cast<IntegerAttr>(attr).getInt();
-        ofr = affine::makeComposedFoldedAffineApply(
+        composedIdx = affine::makeComposedFoldedAffineApply(
             rewriter, extractOp.getLoc(),
             rewriter.getAffineSymbolExpr(0) + offset, {newIndices[idx]});
       } else {
         Value dynamicOffset = cast<Value>(pos);
         AffineExpr sym0, sym1;
         bindSymbols(rewriter.getContext(), sym0, sym1);
-        ofr = affine::makeComposedFoldedAffineApply(
+        composedIdx = affine::makeComposedFoldedAffineApply(
             rewriter, extractOp.getLoc(), sym0 + sym1,
             {newIndices[idx], dynamicOffset});
       }
 
       // Update the corresponding index with the folded result.
-      if (auto value = dyn_cast<Value>(ofr)) {
+      if (auto value = dyn_cast<Value>(composedIdx)) {
         newIndices[idx] = value;
       } else {
         newIndices[idx] = rewriter.create<arith::ConstantIndexOp>(
-            extractOp.getLoc(), *getConstantIntValue(ofr));
+            extractOp.getLoc(), *getConstantIntValue(composedIdx));
       }
     }
     if (isa<MemRefType>(xferOp.getBase().getType())) {
diff --git a/mlir/test/Dialect/Vector/scalar-vector-transfer-to-memref.mlir b/mlir/test/Dialect/Vector/scalar-vector-transfer-to-memref.mlir
index 9f10063a75092..7a1d6b3a8344a 100644
--- a/mlir/test/Dialect/Vector/scalar-vector-transfer-to-memref.mlir
+++ b/mlir/test/Dialect/Vector/scalar-vector-transfer-to-memref.mlir
@@ -167,14 +167,14 @@ func.func @transfer_read_1d_extract_dynamic(%m: memref<?xf32>, %idx: index,
 
 //       CHECK: #[[$MAP:.*]] = affine_map<()[s0, s1] -> (s0 + s1)>
 // CHECK-LABEL: func @transfer_read_2d_extract_dynamic(
-//  CHECK-SAME:     %[[MEMREF:.*]]: memref<?x?xf32>, %[[M_IDX:.*]]: index, %[[ROW:.*]]: index, %[[COL:.*]]: index
-//       CHECK:   %[[ROW_APPLY:.*]] = affine.apply #[[$MAP]]()[%[[M_IDX]], %[[ROW]]]
-//       CHECK:   %[[COL_APPLY:.*]] = affine.apply #[[$MAP]]()[%[[M_IDX]], %[[COL]]]
+//  CHECK-SAME:     %[[MEMREF:.*]]: memref<?x?xf32>, %[[ROW_IDX:.*]]: index, %[[COL_IDX:.*]]: index, %[[ROW_OFFSET:.*]]: index, %[[COL_OFFSET:.*]]: index
+//       CHECK:   %[[ROW_APPLY:.*]] = affine.apply #[[$MAP]]()[%[[ROW_IDX]], %[[ROW_OFFSET]]]
+//       CHECK:   %[[COL_APPLY:.*]] = affine.apply #[[$MAP]]()[%[[COL_IDX]], %[[COL_OFFSET]]]
 //       CHECK:   %[[RES:.*]] = memref.load %[[MEMREF]][%[[ROW_APPLY]], %[[COL_APPLY]]]
-func.func @transfer_read_2d_extract_dynamic(%m: memref<?x?xf32>, %idx: index,
+func.func @transfer_read_2d_extract_dynamic(%m: memref<?x?xf32>, %row_idx: index, %col_idx: index,
                                             %row_offset: index, %col_offset: index) -> f32 {
   %cst = arith.constant 0.0 : f32
-  %vec = vector.transfer_read %m[%idx, %idx], %cst {in_bounds = [true, true]} : memref<?x?xf32>, vector<10x5xf32>
+  %vec = vector.transfer_read %m[%row_idx, %col_idx], %cst {in_bounds = [true, true]} : memref<?x?xf32>, vector<10x5xf32>
   %elem = vector.extract %vec[%row_offset, %col_offset] : f32 from vector<10x5xf32>
   return %elem : f32
 }



More information about the Mlir-commits mailing list