projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 2fc3c5c) | patch
[mlir][bufferization][NFC] Cache definitions of read tensors
authorMatthias Springer <springerm@google.com>
Thu, 9 Feb 2023 08:10:48 +0000 (09:10 +0100)
committerMatthias Springer <springerm@google.com>
Thu, 9 Feb 2023 08:27:39 +0000 (09:27 +0100)
commit2b5a020d3e3ca2a887218776bc6e7bd930a656fa
tree0d9e378103c02ee6b92eb4e25c2b35bc54129f2dtree | snapshot
parent2fc3c5c34c4c0ce94a217717a469620e06325fb0commit | diff
[mlir][bufferization][NFC] Cache definitions of read tensors

This is to avoid unnecessary traversals of the IR.

Differential Revision: https://reviews.llvm.org/D143408
mlir/include/mlir/Dialect/Bufferization/Transforms/OneShotAnalysis.h diff | blob | history
mlir/include/mlir/Dialect/Bufferization/Transforms/Transforms.h diff | blob | history
mlir/lib/Dialect/Bufferization/Transforms/EmptyTensorElimination.cpp diff | blob | history
mlir/lib/Dialect/Bufferization/Transforms/OneShotAnalysis.cpp diff | blob | history
Domain: System / Toolchain;