This is an archive of the discontinued LLVM Phabricator instance.

AVX-512: valignq/d intrinsic (Clang)
ClosedPublic

Authored by igorb on Aug 23 2015, 10:32 PM.

Diff Detail

Repository
rL LLVM

Event Timeline

igorb updated this revision to Diff 32932.Aug 23 2015, 10:32 PM
igorb retitled this revision from to AVX-512: valignq/d intrinsic (Clang).
igorb updated this object.
igorb added reviewers: delena, AsafBadouh.
igorb set the repository for this revision to rL LLVM.
igorb added a subscriber: llvm-commits.
vsk added a subscriber: vsk.Aug 24 2015, 5:07 PM

This change looks fine. I'm curious though -- the spec uses "int" for "cnt" in the Compiler Intrinsic Equivalent section, but "imm8" for the instruction encoding. Why are they different?

This revision was automatically updated to reflect the committed changes.
igorb added a comment.Aug 31 2015, 5:26 AM
In D12274#231679, @vsk wrote:

This change looks fine. I'm curious though -- the spec uses "int" for "cnt" in the Compiler Intrinsic Equivalent section, but "imm8" for the instruction encoding. Why are they different?

This is according to Spec.