[all-commits] [llvm/llvm-project] 12e41d: [mlir][bufferize] Infer memref types when possible

[Matthias Springer via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 12e41d9264b6f84213be86aab75016fb82ebc1d1
      https://github.com/llvm/llvm-project/commit/12e41d9264b6f84213be86aab75016fb82ebc1d1
  Author: Matthias Springer <springerm at google.com>
  Date:   2022-05-16 (Mon, 16 May 2022)

  Changed paths:
    M mlir/include/mlir/Dialect/Bufferization/IR/BufferizableOpInterface.h
    M mlir/lib/Dialect/Arithmetic/Transforms/BufferizableOpInterfaceImpl.cpp
    M mlir/lib/Dialect/SCF/Transforms/BufferizableOpInterfaceImpl.cpp
    M mlir/lib/Dialect/Shape/Transforms/BufferizableOpInterfaceImpl.cpp

  Log Message:
  -----------
  [mlir][bufferize] Infer memref types when possible

Instead of recomputing memref types from tensor types, try to infer them when possible. This results in more precise layout maps.

Differential Revision: https://reviews.llvm.org/D125614