Conversation
Relabels the benchmark rows so the progression is legible rather than implied: Hyperfly is the schema-compiled layout alone, Hyperfly + Brotli puts generic compression on top of it, and Hyperfly Profiled adds a dictionary trained on the route's own traffic. The middle row is new — the site previously jumped from bare layout to the full stack, which hid which of the two later stages was doing the work. It turns out to be worth showing. On events Brotli takes 55 bytes off the columnar output and the profile takes 1 231, which is the whole argument for training: a compressor sees one response and has to rediscover the recurring user agents every time. On orders Brotli costs two bytes rather than saving any, and on candles the last two rows are identical to the byte because that route has no string column for a dictionary to key on. All three are left visible. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_hf1
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Relabels the benchmark rows per request:
The middle row is new. The site previously went from bare layout straight to the full stack, which hid which of the two later stages was actually doing the work — and the answer varies enough per route to be the interesting part:
On events Brotli removes 55 bytes and the profile removes 1,231 — that gap is the argument for training, since a compressor sees one response and rediscovers the recurring user agents every time. On orders Brotli costs two bytes. On candles the last two rows are byte-identical because that route's only string sits outside the array, so there is no column to key on. All three left visible.
Bench harness now measures columnar+brotli too, so the site numbers stay reproducible from
bun run bench.🤖 Generated with Claude Code