projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 3e6afa7) | patch
[AARCH64][SVE] Do not optimize vector conversions
authorZino Benaissa <zbenaissa@nvidia.com>
Thu, 19 Jan 2023 16:44:23 +0000 (16:44 +0000)
committerSjoerd Meijer <smeijer@nvidia.com>
Thu, 19 Jan 2023 16:50:31 +0000 (16:50 +0000)
commit68f45796edbdc4331a6be0b6e9d58f01d8b1fba0
tree0356873c93e0ce233233abbd43fed4cae6483e3dtree | snapshot
parent3e6afa77b4ec8b5fb5e0b6349f08e4476892c0f8commit | diff
[AARCH64][SVE] Do not optimize vector conversions

shuffle_vector instructions are serialized targeting SVE fixed vectors, see
https://reviews.llvm.org/D139111. This patch disables
optimizeExtendOrTruncateConversion peepholes that generates shuffle_vector.

Differential Revision: https://reviews.llvm.org/D141439
llvm/lib/Target/AArch64/AArch64ISelLowering.cpp diff | blob | history
llvm/test/CodeGen/AArch64/sve-fixed-vector-zext.ll [new file with mode: 0644] blob
Domain: System / Toolchain;