[PATCH] D48712: [X86] Lowering integer truncation intrinsics to native IR

[Mikhail Dvoretckii via Phabricator via cfe-commits]

mike.dvoretsky created this revision.
mike.dvoretsky added a reviewer: craig.topper.
Herald added a subscriber: cfe-commits.

This patch lowers the _mm[256|512]_cvtepi{64|32|16}_epi{32|16|8} intrinsics to native IR in cases where the result's length is less than 128 bits.

The resulting IR is folded into VPMOV instructions in https://reviews.llvm.org/D46957, with the exception of _mm_cvtepi64_epi8, where a PSHUFB instruction is currently produced instead.


Repository:
  rC Clang

https://reviews.llvm.org/D48712

Files:
  clang/lib/Headers/avx512vlbwintrin.h
  clang/lib/Headers/avx512vlintrin.h
  clang/test/CodeGen/avx512vl-builtins.c
  clang/test/CodeGen/avx512vlbw-builtins.c

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D48712.153272.patch
Type: text/x-patch
Size: 9777 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20180628/b3603dc9/attachment-0001.bin>