-
Notifications
You must be signed in to change notification settings - Fork 15
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
Decrease run time #21
Conversation
* Reduce system sizes * Use CPU in MD simulations * Reduce fresnel image resolution and samples * TODO: Still need to update the logging tutorial
due to a bug in hoomd v3.0.0-beta.3
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
@joaander You might find this set of pre-commit hooks helpful as well:
|
If you think it would be useful, feel free to make a PR. We do need to preserve some metadata (nbsphinx hidden cells and a tag that allows the notebook to continue when a cell throws an excption). |
View / edit / reply to this conversation on ReviewNB tommy-waltmann commented on 2021-02-25T13:30:40Z Let's make the variable on the first line be joaander commented on 2021-02-25T13:45:28Z Good catch. |
View / edit / reply to this conversation on ReviewNB tommy-waltmann commented on 2021-02-25T13:30:41Z Same as above comment here |
Does the current runtime of the notebooks ensure they aren't a bottleneck for CI now? |
Good catch. View entire conversation on ReviewNB |
Yes, the runtime reduction from 5 hours to 20 minutes is significant. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The changes LGTM. Also the cut down in time is great!
I’ll pause on #22 until this is merged, to avoid burning tons of CI time. |
Description
Decrease the run time of the examples.
On my linux box, they now run in 10 minutes. On Azure, they run in 20 minutes (compare to 5 hours previously).
Also:
With these changes, the notebook outputs are now as reproducible as they can be. The HOOMD trajectories and the fresnel renderings do not change when rerun on the same system. We may make use of this in the future and have the CI committing the notebook output.
The only changes in the diffs on rerunning (with the same binaries) are now:
There is supposed to be a
record_timing
metadata key and/or nbconvert option to preventiopub
date/time information from changing the cell metadata every time the notebook is rerun. However, I was unable to get this option to have any effect.Checklist: