Repository navigation
[perf] Make LPT parallel scheduler the default, drop --lpt opt-in flag - #8785
Closed
TomasVotruba wants to merge 2 commits into
Closed
TomasVotruba wants to merge 2 commits into
TomasVotruba wants to merge 2 commits into
Conversation
The LPT (Longest Processing Time first) bucket scheduler is now the only parallel scheduler. It balances files across workers by size up front and keeps each worker alive for its whole bucket, preserving warm caches. - promote LptScheduleFactory to ScheduleFactory - promote ExperimentalParallelFileProcessor to ParallelFileProcessor - move BucketSchedule out of the Experimental namespace - remove the old shared-pool ScheduleFactory, ParallelFileProcessor and Schedule - drop the --lpt CLI option and its CI usages
TomasVotruba
changed the title
- regenerate expected output of the parallel e2e tests for the LPT file order (single worker, bucket ordered by file size); the change set is identical, only the reporting order differs - update test comments that referenced the removed mid-run worker respawn - drop the obsolete ProcessPool nullable ignore; the promoted processor uses a non-nullable ProcessPool
Member
|
I already created PR: for it |
This file contains hidden or 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The LPT (Longest Processing Time first) bucket scheduler, previously opt-in behind
--lpt, is now the only parallel scheduler.It splits files into one fixed bucket per worker balanced by total byte size, and keeps each worker alive for its whole bucket, so warm in-memory caches (reflection, parsed AST, scope resolver) survive the run instead of being thrown away on worker recycling.
Changes
LptScheduleFactory->Rector\\Parallel\\ScheduleFactoryExperimentalParallelFileProcessor->Rector\\Parallel\\Application\\ParallelFileProcessorBucketScheduleout of theExperimentalnamespaceScheduleFactory,ParallelFileProcessorandSchedulevalue object--lptCLI option and remove it from therectorandbuild_scoped_rectorworkflowsNet -381 lines. The parallel path was smoke-tested with a dry-run over a multi-file directory.
Note:
PARALLEL_LPT_POC_NOTES.mdat the repo root still documents the old--lptflag; left untouched here, can be removed in a follow-up.