[PATCH] D35996: [X86][AVX] Ensure vector non-temporal load/store intrinsics force pointer alignment (PR33830)

Simon Pilgrim via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Fri Jul 28 07:08:59 PDT 2017


RKSimon created this revision.

Clang specifies a max type alignment of 16 bytes on darwin targets (annoyingly in the driver not via cc1), meaning that the builtin nontemporal stores don't correctly align the loads/stores to 32 or 64 bytes when required, resulting in lowering to temporal unaligned loads/stores.

This patch casts the vectors to explicitly aligned types prior to the load/store to ensure that the require alignment is respected.


Repository:
  rL LLVM

https://reviews.llvm.org/D35996

Files:
  lib/Headers/avx2intrin.h
  lib/Headers/avx512fintrin.h
  lib/Headers/avxintrin.h
  test/CodeGen/x86-nontemporal.c

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D35996.108646.patch
Type: text/x-patch
Size: 8163 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20170728/e7cc3108/attachment-0001.bin>


More information about the cfe-commits mailing list