Improve performance of hybrid encrypt CLI #1483
Merged
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
There were quite a few bottlenecks here: * Writes were done serially, writing one file at a time. * Shares were encrypted on a single CPU core I almost used `rayon` to parallelize encryption, but the problem is that we need to get the output sorted to maintain total order across files. Rayon can do that, but requires collecting `ParallelIterator` which would be bad for generating 100M+ reports. Our goal is to be able to share and encrypt 1B, so streaming and manual fiddling with thread pools is justified imo. The way this CLI works right now: it keeps a compute pool for encryption (thread-per-core) and a separate pool of 3 threads to write data for each helper in parallel I also made a few tweaks to improve code re-usability in this module. ## Benchmarks Done locally on M1 Mac Pro (10 cores) Before this change: Encryption process is completed. 442.15834075s After this change Encryption process is completed. 55.63269625s
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #1483 +/- ##
==========================================
- Coverage 93.37% 93.27% -0.10%
==========================================
Files 239 239
Lines 43476 43675 +199
==========================================
+ Hits 40594 40739 +145
- Misses 2882 2936 +54 ☔ View full report in Codecov by Sentry. |
| ) | ||
| }) | ||
| .collect(), | ||
| next_worker: 0, |
There was a problem hiding this comment.
why is next_worker always 0? what's the point of it if it's constant?
There was a problem hiding this comment.
mutation occurs inside encrypt_share
eriktaubeneck
approved these changes
Dec 7, 2024
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.
There were quite a few bottlenecks here:
I almost used
rayonto parallelize encryption, but the problem is that we need to get the output sorted to maintain total order across files. Rayon can do that, but requires collectingParallelIteratorwhich would be bad for generating 100M+ reports.Our goal is to be able to share and encrypt 1B, so streaming and manual fiddling with thread pools is justified imo.
The way this CLI works right now: it keeps a compute pool for encryption (thread-per-core) and a separate pool of 3 threads to write data for each helper in parallel
I also made a few tweaks to improve code re-usability in this module.
Benchmarks
Time it takes to encrypt 1M reports done locally on M1 Mac Pro (10 cores)
Before this change:
Encryption process is completed. 442.15834075s
After this change
Encryption process is completed. 55.63269625s