-
Notifications
You must be signed in to change notification settings - Fork 4.3k
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
[profiling]add show_straggler argument to log_summary() #3579
Merged
tjruwase
merged 17 commits into
deepspeedai:master
from
delock:gma/log_summary_straggler
Jun 30, 2023
Merged
[profiling]add show_straggler argument to log_summary() #3579
tjruwase
merged 17 commits into
deepspeedai:master
from
delock:gma/log_summary_straggler
Jun 30, 2023
Conversation
This file contains 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
tjruwase
reviewed
Jun 29, 2023
tjruwase
reviewed
Jun 29, 2023
tjruwase
approved these changes
Jun 30, 2023
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Outdated suggestions cannot be applied.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.
In tensor parallel inference straggler effect is one of the factor that impacts scaling efficiency. Between any two allreduce of tensor parallel, one worker may run slower than other workers, this worker become the straggler for this duration. Different duration (between two allreduces) might have different straggler.
When do profiling, Straggler effect usually show as part of communication latency. This is because faster worker needs to wait for straggler during communication, this usually cause long communicatio latency. Part of communication latency is waiting for straggler to start, the rest is the time real communication takes.
It would helpful to quantify straggler effect in a workload, this PR extends communciation logging by add an optional argument to
log_summary
:When
show_straggler
is set to True, communication logging would show a new table showing communciation latency and straggler effect. Communication latency is defined as the duration between the time when all ranks started communication and the communication finished. Straggler effect is define as the time between current rank started communication and all ranks started communication.Below is an example:
The column 'Total straggler(ms)' show the straggler time. In this case, for 32KB allreduce, real communciation latency is 966.38ms and straggler effect is 315.26ms. Which shows significant straggler effect in the workload.
In this PR, straggler effect is defined as: