[all-commits] [llvm/llvm-project] a36348: [mlir][bufferize] Fix bug in AllocTensorElimination

[Matthias Springer via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: a36348c5868def196062fe74714b9a8ab754cdd0
      https://github.com/llvm/llvm-project/commit/a36348c5868def196062fe74714b9a8ab754cdd0
  Author: Matthias Springer <springerm at google.com>
  Date:   2022-08-15 (Mon, 15 Aug 2022)

  Changed paths:
    M mlir/lib/Dialect/Bufferization/Transforms/AllocTensorElimination.cpp
    M mlir/test/Dialect/Bufferization/Transforms/one-shot-bufferize-alloc-tensor-elimination.mlir

  Log Message:
  -----------
  [mlir][bufferize] Fix bug in AllocTensorElimination

AllocTensorElimination does currently not support chains where the type is
changing. AllocTensorElimination used to generate invalid IR for such
inputs. With this commit, AllocTensorElimination does no longer apply to
such inputs. (It can be extended to support such IR if needed.)

Differential Revision: https://reviews.llvm.org/D131880