Repository navigation
Conversation
ale18V
force-pushed
the
bench/worker-peer-transport
branch
from
November 18, 2026 15:09
d76e6cc to
0f455df
Compare
ale18V
changed the base branch from
feat/training-worker-page-client-alternative
to
fix/worker-model-recovery
November 18, 2026 15:09
ale18V
marked this pull request as ready for review
November 18, 2026 15:10
ale18V
marked this pull request as draft
November 18, 2026 15:36
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What & why
Follow-up to #1287, stacked on #1295 to inherit its lifecycle fixes and browser-test startup timeout. A real-worker harness separates MessagePort dispatch, copy/transfer, cloning and the production peer bridge, then compares direct and bridged WebRTC loopback exchange. In one Chrome/Linux run, one-peer median WebRTC RTT increased by roughly 3–6% for 1–64 MiB payloads; the small CPU training workload reduced maximum page-timer lag from about 36 ms to 1 ms. The recorded report includes raw JSON, hardware metadata, the measured source revision and limitations. Validation covers 195 recorded measurements across three repetitions, payload integrity/byte counts, lint, formatting, type checks, dependency checks and the production build.
Technical plan
Warm up each path and rotate/reverse order across repetitions, measuring realistic payloads with one or four peers using the unmodified production adapter and WebRTC peers. A separate Titanic CPU training-and-exchange workload starts both paths from identical serialized weights and reports training, encoding, exchange, applying and page responsiveness. Keep the benchmark out of ordinary Cypress runs and the production app bundle, and provide commands to run it and summarize its JSON report.
Deviations from plan
None.
Todo
Repeat on other hardware and browsers, measure peak memory separately, and validate remote-network/full Disco sessions and larger GPU models before generalizing the results.