[Mlir-commits] [mlir] [mlir][linalg] Vectorize unpack op without masking (PR #89067)

[Andrzej Warzyński]

================
@@ -1560,11 +1574,32 @@ vectorizeAsTensorUnpackOp(RewriterBase &rewriter, tensor::UnPackOp unpackOp,
 
   ArrayRef<int64_t> innerDimPos = unpackOp.getInnerDimsPos();
   ArrayRef<int64_t> innerTiles = unpackOp.getStaticInnerTiles();
+  ArrayRef<int64_t> sourceShape = unpackTensorType.getShape();
+  bool useInBoundsInsteadOfMasking = false;
+  ArrayRef<int64_t> outerDimsPerm = unpackOp.getOuterDimsPerm();
+
+  auto destSize = unpackOp.getDestRank();
+
+  // initVectorShape is the shape of the vector that will be read from the
+  // source tensor. It is set like this: Let's say the sourceShape is 'M' and
+  // the vectorSize (VS) array is size 'N' where N <= M. Thus:
+  // - initVectorShape = sourceShape.take_front(N)
+  // - if outer_dims_perms is present: do that permutation on initVectorShape.
+  // - Multiply all the locations pointed by innerDimPos by the innerTileSize
+  //  attribute value.
+  SmallVector<int64_t> initVectorShape(sourceShape.take_front(destSize));
+  if (inputVectorSizes.empty()) {
+    if (!outerDimsPerm.empty())
+      applyPermutationToVector(initVectorShape, outerDimsPerm);
+    for (auto [i, pos] : llvm::enumerate(innerDimPos))
+      initVectorShape[pos] *= innerTiles[i];
+
+    inputVectorSizes = initVectorShape;
+    useInBoundsInsteadOfMasking = true;
+  }
 
   SmallVector<int64_t> readMaskShape(inputVectorSizes.begin(),
                                      inputVectorSizes.end());
----------------
banach-space wrote:

Good point, I missed that!

`readMaskShape` is used when calling `createReadOrMaskedRead`:
https://github.com/llvm/llvm-project/blob/6d44a1ef55b559e59d725b07ffe1da988b4e5f1c/mlir/lib/Dialect/Linalg/Transforms/Vectorization.cpp#L1613-L1616

And, the signature of `createReadOrMaskedRead` is here: https://github.com/llvm/llvm-project/blob/6d44a1ef55b559e59d725b07ffe1da988b4e5f1c/mlir/include/mlir/Dialect/Vector/Utils/VectorUtils.h#L195-L197

So it's not really `readMaskShape`, it's `readShape` or `readVectorSizes` like you suggested. Am I correct that we only need to calculate this once?

https://github.com/llvm/llvm-project/pull/89067