Skip to content

docs(readme): update readme for scheduling algorithm #43

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 30, 2025
Merged

docs(readme): update readme for scheduling algorithm #43

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 11 additions & 2 deletions src/scheduling/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,17 @@

This directory implements a two-phase scheduler for distributed LLM inference:

- **Phase 1 — Layer allocation**: assign contiguous decoder layer ranges to nodes and rebalance in place.
- **Phase 2 — Request routing**: compute an end-to-end, minimum-latency path across the assigned node ranges.
### Phase 1 — Layer allocation

Assign contiguous decoder layer ranges to nodes and rebalance in place, as illustrated below:

<img width="1874" height="852" alt="parallax_1" src="https://github.com/user-attachments/assets/c57cde77-0cda-48fc-b1ad-6d4aa1b1787b" />

### Phase 2 — Request routing

Compute an end-to-end, minimum-latency path across the assigned node ranges, as illustrated below:

<img width="1828" height="705" alt="parallax_2" src="https://github.com/user-attachments/assets/8a6b4d8f-8d97-402b-ba84-3ce61e4ee313" />

The main entrypoint is `scheduling.scheduler.Scheduler`, which orchestrates allocation, dynamic joins/leaves, health checks, and routing.

Expand Down