[PATCH] D62363: [X86] Enable intrinsics that convert float and bf16 data to each other

[Kan Shengchen via Phabricator via cfe-commits]

skan updated this revision to Diff 203170.
skan added a comment.

When target feature `avx512bf16` is enabled, `avx512vl` is not enabled implicity for some reasons by now. But `_mm_cvtness_sbh`, `_mm256_cvtpbh_ps` , `_mm256_maskz_cvtpbh_ps` , `_mm256_mask_cvtpbh_ps` need `avx512vl` feature. So I move their definition from `avx512bf16intrin` to `avx512vlbf16intrin.h`, and move corresponding test cases from `avx512bf16-builtins.c` to `avx512vlbf16-builtins.c`.

And the test for `_mm_cvtness_sbh` is fixed in the new patch.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D62363/new/

https://reviews.llvm.org/D62363

Files:
  include/clang/Basic/BuiltinsX86.def
  lib/CodeGen/CGBuiltin.cpp
  lib/Headers/avx512bf16intrin.h
  lib/Headers/avx512vlbf16intrin.h
  test/CodeGen/avx512bf16-builtins.c
  test/CodeGen/avx512vlbf16-builtins.c

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D62363.203170.patch
Type: text/x-patch
Size: 14219 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20190605/3f614501/attachment.bin>