Skip to content

refactor: unify engine ownership under Engine and de-templatize schedulers. - #2532

Draft
wai-bao wants to merge 1 commit into
xLLM-AI:mainfrom
wai-bao:refactor/unify-engine-ownership
Draft

wai-bao wants to merge 1 commit into
xLLM-AI:mainfrom
wai-bao:refactor/unify-engine-ownership

Conversation

@wai-bao

@wai-bao wai-bao commented Oct 20, 2026 •

Copy link
Copy Markdown
Collaborator

Unify LLM/VLM target and draft resources under concrete engines and remove the speculative engine wrappers. Schedulers and profiling share one non-template implementation through scheduler callbacks.

  • Centralize worker setup, model loading and cache ownership while keeping cache geometry independent of runtime state.
  • Preserve shared static KV cache, co-located MTP VMM layouts, multimodal memory reservations and model residency paths.
  • Use is_draft_engine consistently and remove unused and test-only production interfaces.
  • Keep speculative verification block tables contiguous across graph capture and replay.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 20, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-20T23:31:58.082657Z ef61631 Draft marked ready
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@wai-bao
wai-bao marked this pull request as draft October 20, 2026 23:13
@wai-bao
wai-bao force-pushed the refactor/unify-engine-ownership branch from e8c8db7 to ef61631 Compare October 20, 2026 23:24
@wai-bao
wai-bao marked this pull request as ready for review October 20, 2026 23:27
@wai-bao
wai-bao marked this pull request as draft October 21, 2026 04:57
@wai-bao
wai-bao force-pushed the refactor/unify-engine-ownership branch 16 times, most recently from b13640e to 8783c79 Compare October 23, 2026 05:46
@wai-bao
wai-bao force-pushed the refactor/unify-engine-ownership branch 17 times, most recently from 1aaf79e to 1ed929e Compare October 24, 2026 09:06
@wai-bao wai-bao closed this Oct 24, 2026
@wai-bao
wai-bao force-pushed the refactor/unify-engine-ownership branch from 1ed929e to 1769d7a Compare October 24, 2026 11:41
@wai-bao wai-bao reopened this Oct 24, 2026
@wai-bao
wai-bao force-pushed the refactor/unify-engine-ownership branch 9 times, most recently from 9cdd4dc to 204244c Compare October 25, 2026 16:37

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant