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

Macrobenchmarks: Django + Skybison #7

Open
2 tasks
bennn opened this issue Aug 7, 2023 · 9 comments
Open
2 tasks

Macrobenchmarks: Django + Skybison #7

bennn opened this issue Aug 7, 2023 · 9 comments

Comments

@bennn
Copy link
Member

bennn commented Aug 7, 2023

Most of the benchmarks so far are tiny! We should try bigger programs.

Ideas:

@tekknolagi
Copy link

https://github.com/tekknolagi/skybison/blob/08f3f441eef002602de86641c443287e0b994711/.github/workflows/cmake.yml#L63-L70

shows how to set up Django minimal from a pinned version. For your experiment you may want to check in the results so you can incrementally add types and make things static.

@vivaan2006
Copy link
Contributor

vivaan2006 commented Aug 28, 2023

@tekknolagi
Copy link

oh the IG Django benchmark I was referring to was https://github.com/facebookarchive/django-workload

@vivaan2006
Copy link
Contributor

Oh ok! Thanks @tekknolagi

@bennn bennn changed the title Macrobenchmarks Macrobenchmarks: Django + Skybison May 21, 2024
@bennn
Copy link
Member Author

bennn commented May 30, 2024

to run skybison we may be able to run these commands (from max's link) one by one:

        mkdir -p /var/tmp/django_minimal
        pushd /var/tmp/django_minimal
        ${{github.workspace}}/benchmarks/benchmarks/django/setup_django_minimal.sh
        popd
        ./benchmarks/quickbench/diffrevs.py --run-django HEAD^ HEAD | tee django-benchmark-results.json

but replacing github.workspace with a path to a clone of the skybison repo

@mrigankpawagi
Copy link
Contributor

Hi, I tried to make skybison work (the QuickBench which involved Django). After quite a lot of configurations I had some limited success in setting up the system. However even then it seems cinder python does not work immediately. Since the underlying benchmark itself is quite simple (much like the one in pyperformance), I think we may be better off by creating our own system (since this one has too many moving parts and is quite difficult to debug...).

@mrigankpawagi
Copy link
Contributor

Moreover I want to remind that since we planned to use something like locust, our system might end up much simpler.

@tekknolagi
Copy link

I am happy to help out a bit in the next week or two

@mrigankpawagi
Copy link
Contributor

Thanks @tekknolagi, that would be great. Looking forward to hear from you.

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

No branches or pull requests

4 participants