Skip to content
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

Use sharded shuffle in Hybrid #1517

Merged

Conversation

akoshelev
Copy link
Collaborator

@akoshelev akoshelev commented Dec 27, 2024

It turns out, we've never used sharded shuffle in Hybrid, running local shuffles (per shard) instead. This change fixes that.
Also, drop Shuffle implementation for sharded contexts to prevent this from happening again.

Using sharded shuffle costs about 20% in terms of latency, according to the latest benchmarks.
with sharded shuffle - 64 min for 1B rows
without sharded shuffle - 53 min for 1B rows

This change also contains a small fix for PRF - to use a resharding step that we forgot to use as well.

It turns out, we've never used sharded shuffle in Hybrid, running local shuffles (per shard) instead. This change fixes that.

Also, drop `Shuffle` implementation for sharded contexts to prevent this from happening again.
Copy link

codecov bot commented Dec 27, 2024

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.85%. Comparing base (3413c3c) to head (251d1f3).
Report is 3 commits behind head on main.

Additional details and impacted files
@@            Coverage Diff             @@
##             main    #1517      +/-   ##
==========================================
- Coverage   93.07%   92.85%   -0.22%     
==========================================
  Files         242      242              
  Lines       44174    44173       -1     
==========================================
- Hits        41113    41019      -94     
- Misses       3061     3154      +93     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@benjaminsavage benjaminsavage merged commit 8a7d3c2 into private-attribution:main Dec 28, 2024
12 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants