-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[x86] Generate AVX512 fixed-point instructions #7129
Conversation
Several legit failures here |
Can't quite figure out why the JIT doesn't like |
Ugh, same deal with the |
Just updated the AVX512_Skylake |
Only test failure appears unrelated |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, tests pass on my AVX512 Linux box
* clean-up abs and saturating_pmulhrs, fix AVX512 saturating_ ops * add test coverage for AVX512 fp ops * generate vpabs on AVX512 * faster AVX2 lowering of saturating_pmulhrs
This PR adds support for generating
saturating_(add | sub)
andpmulh(rs)
on Skylake and Cannonlake (i.e. for AVX512BW). It also increases simd_op_check test coverage of fixed-point operations on those archs.I also did a bit of clean-up on the way:
abs
I did not addabs
to codegen because it doesn't appear that LLVM currently exposes non-masked versions of AVX512abs
variants.Fixes #7002