[all-commits] [llvm/llvm-project] 3a41ff: [mlir][SCF] Peel scf.for loops for even step divison

[Matthias Springer via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 3a41ff4883fe8b9e34a4f30aa9eecaf2ecb2ef44
      https://github.com/llvm/llvm-project/commit/3a41ff4883fe8b9e34a4f30aa9eecaf2ecb2ef44
  Author: Matthias Springer <springerm at google.com>
  Date:   2021-08-03 (Tue, 03 Aug 2021)

  Changed paths:
    M mlir/include/mlir/Dialect/SCF/Passes.h
    M mlir/include/mlir/Dialect/SCF/Passes.td
    M mlir/include/mlir/Dialect/SCF/Transforms.h
    M mlir/lib/Dialect/SCF/Transforms/LoopSpecialization.cpp
    A mlir/test/Dialect/SCF/for-loop-peeling.mlir
    M utils/bazel/llvm-project-overlay/mlir/BUILD.bazel

  Log Message:
  -----------
  [mlir][SCF] Peel scf.for loops for even step divison

Add ForLoopBoundSpecialization pass, which specializes scf.for loops into a "main loop" where `step` divides the iteration space evenly and into an scf.if that handles the last iteration.

This transformation is useful for vectorization and loop tiling. E.g., when vectorizing loads/stores, programs will spend most of their time in the main loop, in which only unmasked loads/stores are used. Only the in the last iteration (scf.if), slower masked loads/stores are used.

Subsequent commits will apply this transformation in the SparseDialect and in Linalg's loop tiling.

Differential Revision: https://reviews.llvm.org/D105804