[all-commits] [llvm/llvm-project] 12b495: [mlir][vector] Add missing support for scalable ve...

[Andrzej Warzyński via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 12b4951866622cfc38342629808a1a6542624689
      https://github.com/llvm/llvm-project/commit/12b4951866622cfc38342629808a1a6542624689
  Author: Andrzej Warzynski <andrzej.warzynski at arm.com>
  Date:   2023-08-10 (Thu, 10 Aug 2023)

  Changed paths:
    M mlir/lib/Dialect/Vector/IR/VectorOps.cpp
    M mlir/lib/Dialect/Vector/Transforms/LowerVectorTransfer.cpp
    M mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir

  Log Message:
  -----------
  [mlir][vector] Add missing support for scalable vectors

This patch adds the missing logic so that the
`TransferReadPermutationLowering` can be used for scalable vectors. To
this end:
  * TransferOp custom C++ builder is updated to support scalable
    vectors,
  * `TransferOpReduceRank` is also updated to support scalable vectors.

This pattern is relevant when lowering `linalg.matmul` via
`vector_multi_reduction` for scalable vectors.

I've also updated relevant code in `TransferOpReduceRank` not to use
`llvm::to_vector` for constructing `SmallVector` from `ArrayRef`. That
hook doesn't work for `ArraryRef<bool>` (*), so for consistency I
switched to an explicit constructor (so that both `newShape` and
`newScalableDim` are constructed in a similar fashion).

(*) IIUC, that's due how implicit narrowing conversions between `bool`
and `*bool` work. Note that these narrowing conversions change when
using initializer lists, see
  * https://en.cppreference.com/w/cpp/language/list_initialization.

Depends on D157092

Differential Revision: https://reviews.llvm.org/D157268