Repository navigation
docs: name what was measured - #33
Merged
Merged
Conversation
The benchmark tables used generic labels — "barrier-ring", "shared-ring broadcast", "point-to-point queue" — so a reader could not tell what had been measured or reproduce it. The repo's own owner asked what "barrier" meant, which settled the question. Every row now names the crate and version: disruptor 4.0, tokio::sync::broadcast 1.53, crossbeam-channel 0.5, flume 0.11. Benchmark ids match, so the published tables and `cargo bench` output line up. The site now also carries the cross-thread table, which is where the loss at one consumer lives, and explains the mechanism rather than gesturing at it: consumers coordinate through a shared sequence barrier by publishing their positions, and the producer folds them into a minimum before it may overwrite a slot. That is free with one consumer and grows with the audience, which is both why this crate loses at N=1 and why it can offer non-gating observers and dead-consumer recovery — there is no shared minimum for a stalled consumer to poison. Naming a library in a reproducible measurement is accuracy. What this repo avoids is organising its story around someone else's, which these tables do not do.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The benchmark tables used generic labels — "barrier-ring", "shared-ring broadcast", "point-to-point queue" — so a reader could not tell what had been measured or reproduce it. The repo's own owner asked what "barrier" meant, which settled the question.
Every row now names the crate and version:
disruptor4.0,tokio::sync::broadcast1.53,crossbeam-channel0.5,flume0.11. Benchmark ids match, so published tables andcargo benchoutput line up.The site now also carries the cross-thread table — which is where this crate's loss at one consumer lives — and explains the mechanism instead of gesturing at it: consumers coordinate through a shared sequence barrier by publishing their positions, and the producer folds them into a minimum before it may overwrite a slot. Free with one consumer, growing with the audience. That is both why this crate loses at N=1 and why it can offer non-gating observers and dead-consumer recovery: there is no shared minimum for a stalled consumer to poison.
Naming a library in a reproducible measurement is accuracy. What this repo avoids is organising its story around someone else's, which these tables do not do.
Verified: 219 tests, clippy clean, fmt clean, both benches build, HTML tag balance checked.