perf: use memoryview in write() to avoid copies#954
Merged
Conversation
Merging this PR will improve performance by ×410
Performance Changes
Tip Curious why this is faster? Comment Comparing 
|
2 tasks
mbeijen
force-pushed
the
perf/memoryview-write
branch
from
c0bff4b to
f42a743
Compare
This was referenced May 17, 2026
Member
|
I'm experimenting some benchmark strategies. Sorry the noise. |
Use memoryview slicing in TLSinTLSStream.write() and SyncStream.write() so that each iteration of the send loop does a zero-copy view advance instead of copying the remaining buffer bytes. Most visible on large POST uploads where partial sends cause repeated re-copies of the tail. Ported from https://codeberg.org/httpxyz/httpcorexyz/pulls/4 Closes pydantic#912 Closes encode/httpcore#1029 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Co-Authored-By: naarob <laforge@forge-sync.local>
mbeijen
force-pushed
the
perf/memoryview-write
branch
from
519a8d6 to
9b70201
Compare
The test only verifies correctness on a normal socket and does not regression-test the perf fix (it passes equally well against the old bytes-slicing code).
Member
|
I've run some tests locally, and I see this dramatically increases the performance. The benchmark setup needs improvements tho. 😞 I'm dropping the test. I prefer to have regression tests focused on high level API. |
Kludex
approved these changes
May 20, 2026
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Use memoryview slicing in TLSinTLSStream.write() and SyncStream.write() so that each iteration of the send loop does a zero-copy view advance instead of copying the remaining buffer bytes. Most visible on large POST uploads where partial sends cause repeated re-copies of the tail.
Ported from https://codeberg.org/httpxyz/httpcorexyz/pulls/4
Closes #912
Closes encode/httpcore#1029