Increase number of cores used to calculate upsample for UNET #16351
+9
−10
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
nkpatel-tt
changed the title
nkpatel-tt
force-pushed
the
nkpatel/upsample_perf_improve
branch
from
52cdc41 to
893a523
Compare
nkpatel-tt
changed the title
nkpatel-tt
force-pushed
the
nkpatel/upsample_perf_improve
branch
from
893a523 to
ff77bb7
Compare
overall performace is increased by more than 5%. Signed-off-by: Nilaykumar Patel <[email protected]>
nkpatel-tt
force-pushed
the
nkpatel/upsample_perf_improve
branch
from
ff77bb7 to
9e97e87
Compare
nkpatel-tt
changed the title
esmalTT
reviewed
Jan 27, 2025
| @@ -17,7 +17,7 @@ | |||
| @pytest.mark.models_device_performance_bare_metal | |||
| @pytest.mark.parametrize( | |||
| "batch, groups, expected_device_perf_fps", | |||
| ((1, 2, 1035.0),), | |||
| ((1, 2, 978.0),), | |||
There was a problem hiding this comment.
This is a pretty big perf regression for UNet upsample, why is this happening?
There was a problem hiding this comment.
The main reason is next convolutions internally has to perform reshard. Let me just put it back in draft and see if I can manage to bypass those reshard by changing something in model(mainly while concating).
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Outdated suggestions cannot be applied.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.
Ticket
Link to Github Issue
Problem description
Currently, Upsample processes one row per core, which can impact the performance.
What's changed
Instead single row per core granularity, make it work for minimum possible shard size for each core. More than 5% of perf improvement is seen.
Checklist