diff options
| author | Andrzej Warzyński <andrzej.warzynski@arm.com> | 2024-03-28 14:53:21 +0000 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-03-28 14:53:21 +0000 |
| commit | d3aa92ed142409266ebcc9cbc20e5f2c2d0209c0 (patch) | |
| tree | 9e8388e2c471e47a432cc86ee5ad5ffc4c2a6ae8 /mlir/test/lib/Dialect/Vector/TestVectorTransforms.cpp | |
| parent | ffed554f2d6590acd5cc8d66af916ec1938326b9 (diff) | |
[mlir][vector] Add support for scalable vectors to VectorLinearize (#86786)
Adds support for scalable vectors to patterns defined in
VectorLineralize.cpp.
Linearization is disable in 2 notable cases:
* vectors with more than 1 scalable dimension (we cannot represent
vscale^2),
* vectors initialised with arith.constant that's not a vector splat
(such arith.constant Ops cannot be flattened).
Diffstat (limited to 'mlir/test/lib/Dialect/Vector/TestVectorTransforms.cpp')
| -rw-r--r-- | mlir/test/lib/Dialect/Vector/TestVectorTransforms.cpp | 4 |
1 files changed, 3 insertions, 1 deletions
diff --git a/mlir/test/lib/Dialect/Vector/TestVectorTransforms.cpp b/mlir/test/lib/Dialect/Vector/TestVectorTransforms.cpp index f14fb18706d1..006225999105 100644 --- a/mlir/test/lib/Dialect/Vector/TestVectorTransforms.cpp +++ b/mlir/test/lib/Dialect/Vector/TestVectorTransforms.cpp @@ -489,7 +489,9 @@ struct TestFlattenVectorTransferPatterns Option<unsigned> targetVectorBitwidth{ *this, "target-vector-bitwidth", llvm::cl::desc( - "Minimum vector bitwidth to enable the flattening transformation"), + "Minimum vector bitwidth to enable the flattening transformation. " + "For scalable vectors this is the base size, i.e. the size " + "corresponding to vscale=1."), llvm::cl::init(std::numeric_limits<unsigned>::max())}; void runOnOperation() override { |