Repository navigation
Conversation
Owner
|
/benchmark -f viltrum --save |
Contributor
|
👋 Benchmark request received. A collaborator will review and approve the run. |
Contributor
|
|
Contributor
Benchmark ResultsFramework: No results captured Full log |
Owner
|
/benchmark -f viltrum --save |
Contributor
|
👋 Benchmark request received. A collaborator will review and approve the run. |
Contributor
Benchmark ResultsFramework:
Full log |
This branch has not been deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Updates the viltrum entry to 1.0.1, a rewrite onto a zero-allocation core.
The last run showed viltrum using only 11–23 of the 64 cores. 1.0 removes per-request GC allocation from the framework (enforced by a test in CI) and drops spawn-per-connection in favour of one
SO_REUSEPORTepoll worker per CPU.c.json) and framework gzip per request. No responses are cached or precomputed.GET /baseline11endpoint; the workers block inepoll_waitwith no busy-poll or fixed-cadence timer, and size to the cpuset). echo-ws is removed because WebSocket is not in viltrum 1.0 yet; it returns on the new core later.VILTRUM_REF=v1.0.1. 1.0.1 fixes a JSON encoder slowdown under many threads (a V-prodGC lock on the hot path).validate.shpasses locally.Could you run
/benchmark -f viltrumwhen convenient? Thanks!