-
-
Notifications
You must be signed in to change notification settings - Fork 51
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
Performance degraded by 8x #253
Comments
Looks like with 6.5.2 we were using the clang version, not sure why we'd be switching to the python one but it's coming from here 437113f#diff-004ace7ca04755ff5d00d7eeb3045b910a59d9a6af99c7ecd7af0acd53297d65R143-R148 |
Thanks for bringing this up. Unfortunately I don't have the capacity to look into this. Can anyone help out? The tests (CI/CD) should be expanded to test both the C implementation and without (similar to the Numba alternative). As a temporary workaround you could try enabling the Numba implementation or downgrading for a Speedup. |
Thanks for the prompt reply, we'll downgrade for the time being.
Is there documentation for this? |
@jvacek could you please test if you can locally import and use the c lang helpers (see the respective python module and tests). |
For documentation, please check: https://timezonefinder.readthedocs.io/en/latest/7_performance.html#numba |
Could you please run the tests of this package (Pytest) locally? These includes running the clang version of the point in polygon implementation. |
|
Could you please navigate to the directory where What is the value of |
I noticed that the tests are failing because the C extension is not being loaded when I run the tests locally on MacOS: #255 |
Will run the pytest later, but FYI this also happened in a cronjob on our k8s cluster, which is running python 3.10.3 on I think a buster-slim image |
Could you please try again with the latest version |
6.5.4 works well, and it's using clang again. thanks for the super fast fix! Cheers |
Between 6.5.2 and 6.5.3 we noticed an ~800% increase in time taken to complete the same loop.
After profiling, it seems most of the time goes to
pt_in_poly_python
.Let me know if there's any other info you need to help troubleshoot this.
The text was updated successfully, but these errors were encountered: