| commit | 2b04291830a2a34b681ae711dabfa1032f6c84f7 | |
| author | Diego Caballero <dieg0ca6aller0@gmail.com> | |
| Fri, 31 Jan 2025 22:21:35 +0000 (31 14:21 -0800) | ||
| committer | GitHub <noreply@github.com> | |
| Fri, 31 Jan 2025 22:21:35 +0000 (31 14:21 -0800) | ||
| tree | 4036897e63019430bb7f09605b140f6ed98d063b | treesnapshot (tar.gz zip) |
| parent | 213a939a792f64e7bfdc684922abdf6cd1d3e388 | commitdiff |
[mlir][Vector] Fix scalable InsertSlice/ExtractSlice lowering (#124861)
It looks like scalable `vector.insertslice/extractslice` ops made their way
through lowering patterns that generate `vector.shuffle` ops. I'm not
sure why this wasn't caught by the verifier, probably because the
shuffle op was folded into something else as part of the same rewrite
and the IR wasn't verified.
This PR fixes the issue by preventing scalable vector.insertslice/extractslice
ops to be lowered to vector shuffles. Instead, they are now lowered to a
sequence of insertslice/extractelement ops using an existing patter.
It looks like scalable `vector.insertslice/extractslice` ops made their way
through lowering patterns that generate `vector.shuffle` ops. I'm not
sure why this wasn't caught by the verifier, probably because the
shuffle op was folded into something else as part of the same rewrite
and the IR wasn't verified.
This PR fixes the issue by preventing scalable vector.insertslice/extractslice
ops to be lowered to vector shuffles. Instead, they are now lowered to a
sequence of insertslice/extractelement ops using an existing patter.
| mlir/lib/Dialect/Vector/Transforms/VectorInsertExtractStridedSliceRewritePatterns.cpp | diffblobblamehistory | |
| mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir | diffblobblamehistory |