Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
As we have recently discovered there was a very subtle issue with the optimizer that wasn't fully restored on resume. And it's possible that some spikes we have encountered were related to that.
So I think it'd help to log resume events.
I'm thinking perhaps simply logging
world_sizeonce per run, which will also have a second use which will tell us if we switched to a higher or lower number of nodes and perhaps see some additional correlations there.Since TB will end up creating a flat line and we won't see the restart events I came up with the following hack, where on restart it logs world_size, and then logs another step with 0, so we end up having nice spikes for each restart with the height of the spike also showing the world size for that run.
Here is an example of 3 restarts with a world_size=2 on my machine:
Perhaps TB has another way to flag sparse points and not making them disappear by drawing a line between them. Like it'd be nice to have some way of making those points extra fat or something... but I didn't find a way.
Feedback and suggestions for improvements are very welcome.