[Merged by Bors] - Optimize process_attestation with active balance cache
#2560
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
michaelsproul
added
A0
bug
|
Testing on a Prater node with
|
michaelsproul
added
ready-for-merge
|
Thank you! bors r+ |
bors bot
pushed a commit
that referenced
this pull request
Sep 3, 2021
## Proposed Changes Cache the total active balance for the current epoch in the `BeaconState`. Computing this value takes around 1ms, and this was negatively impacting block processing times on Prater, particularly when reconstructing states. With a large number of attestations in each block, I saw the `process_attestations` function taking 150ms, which means that reconstructing hot states can take up to 4.65s (31 * 150ms), and reconstructing freezer states can take up to 307s (2047 * 150ms). I opted to add the cache to the beacon state rather than computing the total active balance at the start of state processing and threading it through. Although this would be simpler in a way, it would waste time, particularly during block replay, as the total active balance doesn't change for the duration of an epoch. So we save ~32ms for hot states, and up to 8.1s for freezer states (using `--slots-per-restore-point 8192`).
|
Pull request successfully merged into unstable. Build succeeded: |
bors
bot
changed the title
process_attestation with active balance cacheprocess_attestation with active balance cache
|
thank you! right now https://prater.beaconcha.in is behind because this endpoint is slow (note: it works just fine on pyrmont). also some epochs will be shown with 0 eligible ether and participation-rate, we export epochs with 0 values on timeout right now. will fix these epochs after deployed fix. |
Closed
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.
Proposed Changes
Cache the total active balance for the current epoch in the
BeaconState. Computing this value takes around 1ms, and this was negatively impacting block processing times on Prater, particularly when reconstructing states.With a large number of attestations in each block, I saw the
process_attestationsfunction taking 150ms, which means that reconstructing hot states can take up to 4.65s (31 * 150ms), and reconstructing freezer states can take up to 307s (2047 * 150ms).I opted to add the cache to the beacon state rather than computing the total active balance at the start of state processing and threading it through. Although this would be simpler in a way, it would waste time, particularly during block replay, as the total active balance doesn't change for the duration of an epoch. So we save ~32ms for hot states, and up to 8.1s for freezer states (using
--slots-per-restore-point 8192).