llvm-project

History

Nicolas Vasilache 0b17336f79 [mlir][Vector] Make vector.shape_cast based size-1 foldings opt-in and separate.

This is in prevision of dropping them altogether and using insert/extract based patterns.

Reviewed By: antiagainst

Differential Revision: https://reviews.llvm.org/D113928

2021-11-15 21:17:57 +00:00

CMakeLists.txt

[mlir][Linalg] NFC - Add additional control to lower vector.shape_cast ops

2021-10-27 08:12:57 +00:00

VectorInsertExtractStridedSliceRewritePatterns.cpp

[mlir][Vector] Add InsertStridedSliceOp -> ShuffleOp for the rank-1 cases.

2021-10-27 07:57:17 +00:00

VectorMultiDimReductionTransforms.cpp

[mlir] Move min/max ops from Std to Arith.

2021-11-15 13:19:17 +01:00

VectorOps.cpp

[mlir][Vector] Make vector.shape_cast based size-1 foldings opt-in and separate.

2021-11-15 21:17:57 +00:00

VectorTransferOpTransforms.cpp

[mlir][vector] Move transferOp on tensor opt to folder/canonicalization

2021-04-16 08:13:10 -07:00

VectorTransferPermutationMapRewritePatterns.cpp

[mlir][Linalg] NFC - Add additional control to lower vector.shape_cast ops

2021-10-27 08:12:57 +00:00

VectorTransforms.cpp

[mlir][Vector] Make vector.shape_cast based size-1 foldings opt-in and separate.

2021-11-15 21:17:57 +00:00

VectorUtils.cpp

[mlir] Switch arith, llvm, std & shape dialects to accessors prefixed both form.

2021-10-24 18:36:33 -07:00