Closed jack-morrison closed 6 months ago
Can I test this patch?
Got some tests to fix up here, it seems.
@jack-morrison since you are mentioning the model name of the CPU, do you think it could be added too?
@jack-morrison since you are mentioning the model name of the CPU, do you think it could be added too?
I think that's already available in the topology files as arch
(relevant archspec docs), on the line immediately above this change.
Edit: I guess it's the microarchitecture name that's already presented as arch
. This what I was intending to refer to. I'm happy to add any (or all) other attributes from archspec.cpu
though if that's desirable.
@jack-morrison The CI fails because you need to update the configuration schema as well:
@jack-morrison since you are mentioning the model name of the CPU, do you think it could be added too?
I think that's already available in the topology files as
arch
(relevant archspec docs), on the line immediately above this change.Edit: I guess it's the microarchitecture name that's already presented as
arch
. This what I was intending to refer to. I'm happy to add any (or all) other attributes fromarchspec.cpu
though if that's desirable.
Yeah, it's indeed the microarchitecture, the model name I can retrieve using: archspec.cpu.detect.raw_info_dictionary()['model name']
.
All modified and coverable lines are covered by tests :white_check_mark:
Comparison is base (
ead9f84
) 86.64% compared to head (9d122a0
) 86.64%.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Hello @jack-morrison, Thank you for updating!
Cheers! There are no PEP8 issues in this Pull Request!Do see the ReFrame Coding Style Guide
By having the CPU vendor identifier available, test authors can use "GeniuneIntel", "AuthenticAMD", etc. as a logic condition where high-level characterization is acceptable. There are cases where this may actually be more useful than the particular CPU model name.
Fixes #https://github.com/reframe-hpc/reframe/issues/2995