change base compiler for intel/2023.03 to GCCcore/12.3.0

[boegel]

Since there's no newer Intel compilers release than 2023 update 1, which is used in intel/2023.03, we need to change the base compiler to GCCcore/12.3.0, since that's the one used for foss/2023.05 (which will become foss/2023a).

[boegel]

Submitting a test report for this is a bit tricky, only using --from-pr doesn't suffice to ensure that the updated intel-compilers-2023.1.0.eb is used for dependency resolution, which leads to:

ERROR: Missing dependencies: UCX/1.14.1-intel-compilers-2023.1.0 (no easyconfig file or existing module found)

while UCX-1.14.1-GCCcore-12.3.0.eb should be considered here...

[Micket]

This should include imkl-FFTW-2023.1.0-iimpi-2023.03.eb and iimpi-2023.03.eb.

Maybe this is causing the problem?

[Micket]

Test report by @Micket
SUCCESS
Build succeeded for 5 out of 5 (3 easyconfigs in total)
vera-c1 - Linux Rocky Linux 8.7, x86_64, Intel Xeon Processor (Skylake), Python 3.6.8
See https://gist.github.com/Micket/4f172bcb1fba02f1942700ccfb41f44c for a full test report.

[Micket]

Ignore previous comment (as you already know).

I poked a bit.
The issue looks to be caused by the fact that --from-pr only append the robot paths

== 2023-07-05 19:23:20,189 robot.py:75 INFO Extended list of robot search paths with ['/dev/shm/eb-snph1s7m/files_pr18263']: ['/apps/Common/software/EasyBuild/4.7.2/easybuild/easyconfigs', '/dev/shm/eb-snph1s7m/files_pr18263']

Seems natural that this would prepended, but i guess it's to preserve some behaviour of what easyconfigs to use for dependencies if you specify only specific easyconfigs with eb --from-pr XXX Foo-1.2.3.eb, but that is surely some pretty exotic situation where you'd like to not use one of the dependencies of Foo-1.2.3.eb from your local robot paths instead.

And combined with the fact that (which can be triggered outside of --from-pr):

eb /dev/shm/eb-snph1s7m/files_pr18263/i/impi/impi-2021.9.0-intel-compilers-2023.1.0.eb /dev/shm/eb-snph1s7m/files_pr18263/i/intel-compilers/intel-compilers-2023.1.0.eb -Dr

it doesn't consider command line given easyconfigs first when resolving toolchain, which perhaps it also should.

Removing the conflicting easyconfigs from the robot paths beforehand seems like the only workaround right to get this fixed in time for release.

[Micket]

lgtm

[Micket]

So, a bit backwards here, but merging this first, and then triggering a rebuild on with boegelbot should suffice i believe.

[boegel]

Test report by @boegel
SUCCESS
Build succeeded for 6 out of 6 (3 easyconfigs in total)
node3129.skitty.os - Linux RHEL 8.6, x86_64, Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz (skylake_avx512), Python 3.6.8
See https://gist.github.com/boegel/ce3d3c53875dd8d944d49bceec6b6eb1 for a full test report.

edit: test was run after checking out the intel-2023.03_GCCcore-12.3.0 branch that was used to create this PR, to side-step the missing dependency problem...

[Micket]

Going in, thanks @boegel!

[Micket]

@boegelbot please test @ generoso

[boegelbot]

@Micket: Request for testing this PR well received on login1

PR test command 'EB_PR=18263 EB_ARGS= EB_CONTAINER= /opt/software/slurm/bin/sbatch --job-name test_PR_18263 --ntasks=4 ~/boegelbot/eb_from_pr_upload_generoso.sh' executed!

• exit code: 0
• output:

Submitted batch job 11216

Test results coming soon (I hope)...

- notification for comment with ID 1623241478 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

[Micket]

@boegelbot please test @ jsc-zen2

[boegelbot]

Test report by @boegelbot
SUCCESS
Build succeeded for 3 out of 3 (3 easyconfigs in total)
cns1 - Linux Rocky Linux 8.5, x86_64, Intel(R) Xeon(R) CPU E5-2667 v3 @ 3.20GHz (haswell), Python 3.6.8
See https://gist.github.com/boegelbot/11b58c1baa9b5e787ea36bdf4de721ca for a full test report.

[boegel]

@boegelbot please test @ jsc-zen2

[boegelbot]

@boegel: Request for testing this PR well received on jsczen2l1.int.jsc-zen2.easybuild-test.cluster

PR test command 'EB_PR=18263 EB_ARGS= /opt/software/slurm/bin/sbatch --mem-per-cpu=4000M --job-name test_PR_18263 --ntasks=8 ~/boegelbot/eb_from_pr_upload_jsc-zen2.sh' executed!

• exit code: 0
• output:

Submitted batch job 2960

Test results coming soon (I hope)...

- notification for comment with ID 1624022460 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

[boegelbot]

Test report by @boegelbot
SUCCESS
Build succeeded for 3 out of 3 (3 easyconfigs in total)
jsczen2c1.int.jsc-zen2.easybuild-test.cluster - Linux Rocky Linux 8.5, x86_64, AMD EPYC 7742 64-Core Processor (zen2), Python 3.6.8
See https://gist.github.com/boegelbot/f91749d7bbe82b6b281e137d36d8c0e6 for a full test report.