execution: drop BlockStateCache from tx execution workers (P1) - #23146
Draft
sudeepdino008 wants to merge 1 commit into
Draft
execution: drop BlockStateCache from tx execution workers (P1)#23146sudeepdino008 wants to merge 1 commit into
sudeepdino008 wants to merge 1 commit into
Conversation
sudeepdino008
changed the title
This was referenced Aug 11, 2026
Parallel workers read committed pre-block state through a per-block BlockStateCache tier (TxTask.BlockStateCache). That tier is a pure read cache: the base chain (sd.mem -> StateCache -> files) is frozen for a block's whole execution (sd.mem changes only at blockCache.Flush in completeBlock, after which the next block is scheduled), and intra-block isolation comes from the version map layered on top of the reader. So the tier adds no isolation. It does duplicate StateCache and starve it: a slot read 5000 times in a block reached StateCache once, so StateCache's LRU ranked hot keys as cold. Removing the tier lets workers fill and read StateCache directly, restoring its read stream. Values are unchanged because the base is frozen per block. This removes the worker tier only (TxTask.BlockStateCache and its plumbing). The finalize/apply write buffer (be.blockStateCache) is untouched; later steps of #23140 move it to the version map.
sudeepdino008
force-pushed
the
sudeepdino008/bsc-rm-worker-committed-tier
branch
from
August 12, 2026 06:13
3960cee to
d07d799
Compare
sudeepdino008
changed the base branch from
sudeepdino008/sdmem-commitment-lock-split
to
main
August 12, 2026 06:13
sudeepdino008
changed the title
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Part 1 of removing/reducing dependence on
BlockStateCache(#23140).Summary
BlockStateCache is used in two places:
versionedMap(outside of reader) -> BlockCache -> sd.mem -> sd.parent.mem -> StateCache -> temporalDbBlockStateCache has two parts:
readCurrent=falsesoBlockStateCache.currentis anyway skipped.); StateCache provides the "committed" data. So tx execution workers are already covered.StateCacheis already an in-mem cache.BlockStateCache's "committed state" is acting as a duplicate layer on top, interfering withStateCacheretention/LRU policies etc. (so it justifies removal of "commitment state" of BlockCache - will be done in separate PR)