[all-commits] [llvm/llvm-project] 1b1010: [mlir] Turn Linalg to LLVM into a partial conversion

[ftynse via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 1b101038dc6603c04fd56c9511a0ef7e8e9474e0
      https://github.com/llvm/llvm-project/commit/1b101038dc6603c04fd56c9511a0ef7e8e9474e0
  Author: Alex Zinenko <zinenko at google.com>
  Date:   2021-02-05 (Fri, 05 Feb 2021)

  Changed paths:
    M mlir/integration_test/Dialect/Linalg/CPU/matmul-vs-matvec.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/rank-reducing-subview.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-1d-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-1d-ncw-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-1d-nwc-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-2d-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-2d-nchw-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-2d-nhwc-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-3d-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-3d-ncdhw-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-conv-3d-ndhwc-call.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-subtensor-insert-multiple-uses.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-subtensor-insert.mlir
    M mlir/integration_test/Dialect/Linalg/CPU/test-tensor-matmul.mlir
    M mlir/integration_test/Sparse/CPU/frostt-example.mlir
    M mlir/integration_test/Sparse/CPU/matrix-market-example.mlir
    M mlir/lib/Conversion/LinalgToLLVM/LinalgToLLVM.cpp
    M mlir/test/Dialect/Linalg/llvm.mlir
    M mlir/test/mlir-cpu-runner/async.mlir
    M mlir/test/mlir-cpu-runner/unranked_memref.mlir
    M mlir/test/mlir-cpu-runner/utils.mlir

  Log Message:
  -----------
  [mlir] Turn Linalg to LLVM into a partial conversion

Historically, Linalg To LLVM conversion subsumed numerous other conversions,
including (affine) loop lowerings to CFG and conversions from the Standard and
Vector dialects to the LLVM dialect. This was due to the insufficient support
for partial conversions in the infrastructure that essentially required
conversions that involve type change (in this case, !linalg.range to
!llvm.struct) to be performed in a single conversion sweep. This is no longer
the case so remove the subsumed conversions and run them as separate passes
when necessary.

Depends On D95317

Reviewed By: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D96008