[Mlir-commits] [mlir] [mlir][vector] Handle rank mismatch when unrolling broadcasts (PR #227351)

[purnima shrivastava]

================

----------------
purnima-nlp wrote:

Thanks! I switched the regression to the smaller `vector<2x4xi16> -> vector<1x2x4xi16>` case and simplified the CHECKs accordingly.

https://github.com/llvm/llvm-project/pull/227351