perf(packstream): reuse a per-thread scratch buffer in pack#92
Open
freeeve wants to merge 1 commit into
Open
Conversation
pack() allocated and grew a fresh Vec on every call, so each invocation incurred buffer (re)allocations that scale with the encoded size -- the common record/parameter packing path paid 2 to 10 Rust allocations per call. Hold the encode buffer in a thread-local and reuse it across calls: take it on entry, hand its contents to PyBytes, and clear-and-return it on the way out. After warmup this is 0 Rust allocations per pack operation. Buffers above 64 KiB are dropped instead of pooled so one oversized message cannot pin a large allocation for the life of a thread, and the buffer is parked in a Cell<Option<_>> so a nested pack from a Python dehydration hook safely takes a fresh buffer rather than aliasing.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
pack() allocated and grew a fresh Vec on every call, so each invocation incurred buffer (re)allocations that scale with the encoded size -- the common record/parameter packing path paid 2 to 10 Rust allocations per call.
Hold the encode buffer in a thread-local and reuse it across calls: take it on entry, hand its contents to PyBytes, and clear-and-return it on the way out. After warmup this is 0 Rust allocations per pack operation. Buffers above 64 KiB are dropped instead of pooled so one oversized message cannot pin a large allocation for the life of a thread, and the buffer is parked in a Cell<Option<_>> so a nested pack from a Python dehydration hook safely takes a fresh buffer rather than aliasing.
The throughput my local benchmarks saw from this change was up to ~20% write throughput improvement, with smaller records in the higher end of improvements and bigger records (unwind style bulk) in the lower end.