Skip to content
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

use -march=x86-64 -mtune=generic instead of -xSSE2 when using Intel oneAPI compilers #4147

Merged
merged 5 commits into from
Dec 27, 2022

Conversation

boegel
Copy link
Member

@boegel boegel commented Dec 21, 2022

Required since xSSE2 is no longer supported by oneAPI compilers:

error: unknown target CPU 'SSE2'
note: valid target CPU values are: nocona, core2, penryn, bonnell, atom, silvermont, slm, goldmont, goldmont-plus, tremont, gracemont, nehalem, corei7, westmere, sandybridge, corei7-avx, ivybridge, core-avx-i, haswell, core-avx2, broadwell, common-avx512, skylake, skylake-avx512, skx, cascadelake, cooperlake, cannonlake, icelake-client, rocketlake, icelake-server, tigerlake, sapphirerapids, alderlake, knl, knm, k8, athlon64, athlon-fx, opteron, k8-sse3, athlon64-sse3, opteron-sse3, amdfam10, barcelona, btver1, btver2, bdver1, bdver2, bdver3, bdver4, znver1, znver2, znver3, x86-64, x86-64-v2, x86-64-v3, x86-64-v4

tested with HPL in easybuilders/easybuild-easyconfigs#16962 using eb --from-pr 16962 HPL-2.3-intel-2022b.eb --optarch=GENERIC

@boegel boegel added the bug fix label Dec 21, 2022
@boegel boegel added this to the next release (4.7.0) milestone Dec 21, 2022
Copy link
Member

@migueldiascosta migueldiascosta left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

lgtm

bartoldeman
bartoldeman previously approved these changes Dec 22, 2022
Copy link
Contributor

@bartoldeman bartoldeman left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

'intel-compilers': ('2021.4.0', "-xSSE2 -fp-speculation=safe -fp-model precise"),
'intel-compilers': ('2021.4.0', "-march=x86-64 -mtune=generic -fp-speculation=safe -fp-model precise"),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is a bit too lax. It should test intel-compilers with and without oneapi enabled.
So perhaps skip the enforcing of oneapi=true below and have two intel-compilers, one pre-2022.2.0 (like the 2021.4.0 above) with -xSSE2 and one intel-compilers using version > 2022.2.0 to get oneapi auto enabled.
Not sure how to do that easy since you can't have two intel-compilers in this dict...

perhaps an array of single compiler dicts?

Copy link
Member Author

@boegel boegel Dec 24, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@akesandgren Very good point for raising this: I've extended this test extensively in 848293d, and doing so actually uncovered a bug, which was fixed in 63099bf

@boegel boegel force-pushed the oneapi_generic branch 10 times, most recently from df2a695 to e4d3f1e Compare December 24, 2022 15:37
… oneapi toolchain option in test_optarch_generic
…if oneapi toolchain option is not set to True or False
Copy link
Member

@migueldiascosta migueldiascosta left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

lgtm, but it looked good to me before, so waiting for an extra pair of eyes before merging

@bartoldeman bartoldeman merged commit e061d91 into easybuilders:develop Dec 27, 2022
@migueldiascosta
Copy link
Member

nevermind :)

@migueldiascosta
Copy link
Member

ah, it was on auto-merge, missed that... even then, I would have expected that a more recent commit would clear the auto-merge, same as dismissing an outdated review...

@boegel boegel deleted the oneapi_generic branch December 27, 2022 10:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants