perf(response): pool on-the-fly compression writers#4385
Open
jvoisin wants to merge 1 commit into
Open
Conversation
For every dynamic response above the 1 KiB compression threshold the
builder previously did one of:
brotli.NewWriterV2(b.w, brotli.DefaultCompression)
gzip.NewWriter(b.w)
flate.NewWriter(b.w, -1)
Each constructor allocates the encoder's working set from scratch:
brotli's sliding window + hash tables (~3.5 MiB at quality 6), gzip's
CRC32 + deflate state (~800 KiB), flate's hash chains (~800 KiB).
On a busy server that's per-request churn the GC then has to clean up.
All three writer types expose Reset(dst io.Writer), which rebinds the
destination without touching the internal buffers. So put each behind a
sync.Pool and Get/Reset/Put around the existing Write+Close pair. The
constructor signatures didn't change; only Pool.Get + Reset are new.
Note that brotli.NewWriterV2 (kept from before this change) returns
*matchfinder.Writer, not *brotli.Writer, as V2 is the pure-Go encoder
built on top of github.com/andybalholm/brotli/matchfinder, where the
actual Writer type lives. Hence the matchfinder import.
On a local artificial benchmarks of a 130 KiB HTML-like payload consisting of
250 entry-list items, on a single-core it improves performances by around 10%,
and for multicore under GC pressure, ns/op is reduces by ~80% and B/op by ~99%.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
For every dynamic response above the 1 KiB compression threshold the builder previously did one of:
brotli.NewWriterV2(b.w, brotli.DefaultCompression)
gzip.NewWriter(b.w)
flate.NewWriter(b.w, -1)
Each constructor allocates the encoder's working set from scratch: brotli's sliding window + hash tables (~3.5 MiB at quality 6), gzip's CRC32 + deflate state (~800 KiB), flate's hash chains (~800 KiB). On a busy server that's per-request churn the GC then has to clean up.
All three writer types expose Reset(dst io.Writer), which rebinds the destination without touching the internal buffers. So put each behind a sync.Pool and Get/Reset/Put around the existing Write+Close pair. The constructor signatures didn't change; only Pool.Get + Reset are new.
Note that brotli.NewWriterV2 (kept from before this change) returns *matchfinder.Writer, not *brotli.Writer, as V2 is the pure-Go encoder built on top of github.com/andybalholm/brotli/matchfinder, where the actual Writer type lives. Hence the matchfinder import.
On a local artificial benchmarks of a 130 KiB HTML-like payload consisting of 250 entry-list items, on a single-core it improves performances by around 10%, and for multicore under GC pressure, ns/op is reduces by ~80% and B/op by ~99%.