You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm working on updating the mypy benchmark to be compatible with Python 3.11 (primarily by just updating the dependencies). However, I've run into a small snag. The mypyc benchmark, which is a "subbenchmark" of mypy, doesn't actually seem to run mypyc -- it just runs the same thing as the mypy benchmark with 50 loops rather than 20. Additionally, there is a bug when running in the pyperformance harness since the results of both of these use the same name (mypy).
I think we should either:
Add a new benchmark that legitimately runs mypyc
Just remove the mypyc benchmark
Thoughts?
The text was updated successfully, but these errors were encountered:
Ah thanks for pointing this out, looks like this needs a little bit more work so I just removed the mypyc benchmark for now.
Just a word of caution that the mypy wheels on pypi are built using mypyc whereas their sdist is configured to not use mypyc, so this means that if you're comparing a Python version that has a mypy wheel vs one that doesn't you will see a large performance difference due to this. I think this should be addressed by python/pyperformance#195 in newer versions of pyperformance
I'm working on updating the mypy benchmark to be compatible with Python 3.11 (primarily by just updating the dependencies). However, I've run into a small snag. The
mypyc
benchmark, which is a "subbenchmark" ofmypy
, doesn't actually seem to runmypyc
-- it just runs the same thing as themypy
benchmark with 50 loops rather than 20. Additionally, there is a bug when running in thepyperformance
harness since the results of both of these use the same name (mypy
).I think we should either:
mypyc
mypyc
benchmarkThoughts?
The text was updated successfully, but these errors were encountered: