Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -8,3 +8,4 @@ easyconfigs:
options:
# see https://github.com/easybuilders/easybuild-easyconfigs/pull/22371
from-commit: 29f82e43229663c22f0c76cb3fc7b6dd5c407cd7
- RAxML-8.2.13-gompi-2023b-avx2.eb
Copy link
Member

Choose a reason for hiding this comment

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

This is not going to work on systems that don't have AVX2, you need to create a variant of https://github.com/easybuilders/easybuild-easyconfigs/blob/develop/easybuild/easyconfigs/r/RAxML/RAxML-8.2.13-gompi-2023a-standard.eb with a newer toolchain

Copy link
Contributor

Choose a reason for hiding this comment

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

Actually, this is a case where I think we should diverge from the "same modules everywhere" goal.

The AVX2 variant of RAxML is definitely worth installing for x86_64 CPUs, but clearly doesn't make any sense for aarch64.

So we should install "fake" modules for the latter, and get them to produce a meaningful error to explain why that installation is not available on aarch64 targets?

Copy link
Member

Choose a reason for hiding this comment

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

I don't think we need to make this exception, we need a better easyconfig. For example, https://github.com/stamatak/standard-RAxML/blob/master/Makefile.HYBRID.gcc has the mechanism to support both x86 and Arm, we should really be using that.

Copy link
Member

Choose a reason for hiding this comment

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

PR for this particular case in easybuilders/easybuild-easyconfigs#23975

RAxML does seem to have some intrinsics available, but the specific code targeted Haswell and hasn't been touched in 10 years, I think a native CPU compilation is a reasonable option

Loading