-
Notifications
You must be signed in to change notification settings - Fork 1.7k
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
add fbgemm/cci.20210316 #4927
add fbgemm/cci.20210316 #4927
Conversation
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
I think I understand, fbgemm relies on a new C++17 behavior for |
This comment has been minimized.
This comment has been minimized.
@jgsogo It fails on Macos agents because |
According to this http://csharpmulticore.blogspot.com/2014/12/how-to-check-intel-avx2-support-on-mac-os-x-haswell.html, we can check for AVX2 support using
I'll run this line on our MacOS machines... |
Checked our infra. Indeed, MacOS machines doesn't support AVX2 instruction set. I'm opening a ticket to the infra team to check if this is something we can request, if we can migrate to other machines. |
macOS issues have been fixed, we are now in control of the Mac workers and now they all use the same hardware 🎉 . All of them report AVX2 from CPU features (and they are able to build the library). Now I'm facing issues with Linux workers, it looks like it takes more than 5Gb to build these libraries and the docker containers run out of memory (not all of them, I retrying the PR again and again to see if it fails consistently for some configurations). Also, the failure is not reported because the k8s POD dies, and the CI keeps waiting for a response that will never arrive. Fixes:
|
fbgemm is not the only recipe suffering of memory issues. There are also PCL #1891 (comment), and maybe OpenCASCADE #4094.
Would it be possible to allocate more memory to containers running build of specific recipes? |
There are three proposals right now on the table:
|
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
Is this just about memory? Do we know what is the amount it needs? We have modified a bunch of things and now it is easy for us to redirect this build to a cluster with more memory. |
Failure in build 16 ( An unexpected error happened and has been reported. Help is on its way! 🏇 |
This is failing always the same way: processes are killed. @SSE4 , can we confirm if it is about memory and maybe use larger cluster for this one as well? Thanks! |
Closing this PR as stale due to inactivity. PCL was eventually merged after reaching out to the PCL maintainers - turns out was a specific bottleneck in the compilation that required some additional build-system help to prevent a massive amount of memory. This really was an issue outside of the Conan CI realm, and had been reported by other users - if this is really an issue my advice would be to reach out and report it to the authors of fbgemm |
Specify library name and version: fbgemm/cci.20210316
conan-center hook activated.
closes #4428
There are still several issues with: