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

Feature archdetect rpi #223

Closed
wants to merge 999 commits into from
Closed

Conversation

hmeiland
Copy link
Contributor

add raspberry pi support by adding feature to use CPU part for aarch64

@hmeiland
Copy link
Contributor Author

hmeiland commented Jan 19, 2023

checks failed due to HTTP rate limit, not code error...

ocaisa and others added 28 commits March 1, 2024 07:27
…oss/2023a

{2023.06}[foss/2023a] dask v2023.9.2
Make check_missing_installations.sh check against develop branch of EasyBuild
…op back to regular user and do installations
{2023.06}[gompi/2023a,gompi/2023b] OSU Microbenchmarks v 7.1.1, OSU Microbenchmarks v 7.2 w/ CUDA 12.1.1
…t we specify in our RC file, see search order on https://lmod.readthedocs.io/en/latest/145_properties.html?highlight=lmod_rc#the-properties-file-lmodrc-lua . Also, prefix our hooks with eessi_ to avoid unintentional clashes in namespace between site lmod configuration and eessi lmod configuration
…a hook. Lmod only permits the registration of a single function, trying to register a second function will just overwrite the first, as can be seen from https://github.com/TACC/Lmod/blob/a97d68193a6f32ebca4b4bd70dac0f6ac8fddefe/src/Hook.lua#L87
bedroge and others added 20 commits April 4, 2024 17:04
Co-authored-by: Kenneth Hoste <[email protected]>
Co-authored-by: Kenneth Hoste <[email protected]>
Fix Highway hook: change optarch in `pre_prepare_hook` and reset it in `post_prepare_hook`
{2023.06}[2023a] Modflow 6.4.4 and deps
….12-foss-2023a

{2023.06}[foss/2023a] R-bundle-CRAN 2023.12
{2023.06}[system] EasyBuild v4.9.1
@ocaisa
Copy link
Member

ocaisa commented Apr 12, 2024

@hmeiland Can you sync this with upstream, you still have some older CI that's now failing

@ocaisa
Copy link
Member

ocaisa commented Apr 12, 2024

@hmeiland You are targeting the wrong branch with this, it should be 2023.06-software.eessi.io

@ocaisa
Copy link
Member

ocaisa commented Apr 12, 2024

@hmeiland Unfortunately, this is not working as intended, archdetect is no longer recognising neoverse_n1 as an alternative to neoverse_v1 before choosing generic

See https://github.com/EESSI/software-layer/actions/runs/8660373452/job/23748193970?pr=223 , correct output should be
as per https://github.com/EESSI/software-layer/blob/2023.06-software.eessi.io/tests/archdetect/aarch64/neoverse_v1/AWS-awslinux-graviton3.all.output

@ocaisa
Copy link
Member

ocaisa commented Apr 12, 2024

Are we not just better off adding additional required features to eessi_arch_arm.spec to ensure that the Pi 4 doesn't match the condition?

https://marcin.juszkiewicz.com.pl/2022/11/08/from-a-diary-of-aarch64-porter-arm-cpu-features-table/ could help here

@ocaisa
Copy link
Member

ocaisa commented Apr 14, 2024

Alternative to this is merged in #540 where I've ensured that an instruction from the 8.2 instruction set is used to decide on the arch (rather than asimd which already appears in the 8.0 instruction set, and is why the Pi 4 was making a false match). If we wanted an explicit build for the 8.0 instruction set, we'd have to do a bit of work with archdetect (reintroduce asimd as another distinction)

@ocaisa ocaisa closed this Apr 14, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

8 participants