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

Enforce cpu explicitly for Mac devicelab test beds #101871

Merged
merged 2 commits into from Apr 15, 2022

Conversation

Copy link
Contributor

@keyonghan keyonghan commented Apr 13, 2022

All existing devicelab testbeds are Macmini8,1 now, but from tests config side, we are not enforcing the mac_model.

On the other hand, we are enabling M1 tests and M1 bots in prod, this will cause existing Mac devicelab tests to run against M1 bots. However, not all tests are supported on M1 bots.

This PR tags mac_model: Macmini8,1 explicitly to avoid interaction between test beds. This will also help avoid potential breakage when model changes in the future.

Context: #101861

@keyonghan keyonghan requested review from godofredoc and jmagman Apr 13, 2022
@jmagman
Copy link
Member

@jmagman jmagman commented Apr 14, 2022

Is there any way we can tag gpu:Intel instead since that's what we really care about here? We don't want to limit to a model if, say, half our ARM fleet are Mac Minis from last year and the other half are Mac Mini models from this year.

@keyonghan
Copy link
Contributor Author

@keyonghan keyonghan commented Apr 14, 2022

Is there any way we can tag gpu:Intel instead since that's what we really care about here? We don't want to limit to a model if, say, half our ARM fleet are Mac Minis from last year and the other half are Mac Mini models from this year.

Gpu tag for arm Mac is blocked on: https://bugs.chromium.org/p/chromium/issues/detail?id=1109628
Does it make sense we use the mac_model for now and update to gpu when that is supported?

@jmagman
Copy link
Member

@jmagman jmagman commented Apr 14, 2022

That issue says we can use cpu:arm-64, does that work?

@keyonghan
Copy link
Contributor Author

@keyonghan keyonghan commented Apr 14, 2022

That issue says we can use cpu:arm-64, does that work?

Arm bots show cpu: arm64, but intel bots show cpu: x86. Is it suffice for intel bots?

@jmagman
Copy link
Member

@jmagman jmagman commented Apr 14, 2022

That issue says we can use cpu:arm-64, does that work?

Arm bots show cpu: arm64, but intel bots show cpu: x86. Is it suffice for intel bots?

Yes, that's the distinction we need!

@keyonghan
Copy link
Contributor Author

@keyonghan keyonghan commented Apr 14, 2022

flutter/cocoon#1731 and https://flutter-review.googlesource.com/c/infra/+/29061 to support cpu dimension defined in .ci.yaml.

@keyonghan keyonghan changed the title Enforce mac_model explicitly for Mac devicelab test beds Enforce cpu explicitly for Mac devicelab test beds Apr 15, 2022
@keyonghan
Copy link
Contributor Author

@keyonghan keyonghan commented Apr 15, 2022

Updated all Mac builders to use cpu: x86. This should not affect existing workflows, as of now mac_model: Macmini8,1 is equal to cpu: x86.

But will land this PR after flutter/cocoon#1731 is pushed to latest cocoon version.

Copy link
Member

@jmagman jmagman left a comment

LGTM, thanks @keyonghan!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
waiting for tree to go green
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants