[all-commits] [llvm/llvm-project] 5122a2: [mlir][sparse] allow for direct-out passing of spa...

[Aart Bik via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 5122a2c2320c7b14f6585e63b7fc43ac82a550c2
      https://github.com/llvm/llvm-project/commit/5122a2c2320c7b14f6585e63b7fc43ac82a550c2
  Author: Aart Bik <ajcbik at google.com>
  Date:   2024-04-11 (Thu, 11 Apr 2024)

  Changed paths:
    M mlir/include/mlir/Dialect/SparseTensor/Transforms/Passes.h
    M mlir/include/mlir/Dialect/SparseTensor/Transforms/Passes.td
    M mlir/lib/Dialect/SparseTensor/Transforms/SparseAssembler.cpp
    M mlir/lib/Dialect/SparseTensor/Transforms/SparseTensorConversion.cpp
    M mlir/lib/Dialect/SparseTensor/Transforms/SparseTensorPasses.cpp
    A mlir/test/Dialect/SparseTensor/external_direct.mlir

  Log Message:
  -----------
  [mlir][sparse] allow for direct-out passing of sparse tensor buffers (#88327)

In order to support various external frameworks (JAX vs PyTorch) we need
a bit more flexibility in [dis]assembling external buffers to and from
sparse tensors in MLIR land. This PR adds a direct-out option that
avoids the rigid pre-allocated for copy-out semantics.

Note that over time, we expect the [dis]assemble operations to converge
into something that supports all sorts of external frameworks. Until
then, this option helps in experimenting with different options.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications