[clang] [llvm] [clang-tools-extra] [mlir] [MLIR][Linalg] Support dynamic sizes in `lower_unpack` (PR #75494)

lorenzo chelini via cfe-commits cfe-commits at lists.llvm.org
Mon Dec 18 02:12:29 PST 2023


================
@@ -434,8 +429,21 @@ FailureOr<LowerUnPackOpResult> linalg::lowerUnPack(RewriterBase &rewriter,
       RankedTensorType::Builder(packedTensorType).setShape(stripMinedShape);
   RankedTensorType collapsedType = tensor::CollapseShapeOp::inferCollapsedType(
       stripMinedTensorType, packingMetadata.reassociations);
+
+  // Get dynamic dims from input tensor in order of stripMinedTensor
----------------
chelini wrote:

Not sure if I understand the comment. I would rephrase as: `Get dynamic dims from input tensor based on lastDimsToInsertPositionsPerm permutation`.

https://github.com/llvm/llvm-project/pull/75494


More information about the cfe-commits mailing list