lm-sys / lm-sys/FastChat

Will the cache kv become invalid?

Open
#3,245 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
39.5k
Forks
4.8k
PR merge metrics
No merged PRs in 30d

Description

In a multi-threaded situation, if the GPU server resources are insufficient, will cache kv preemption occur? For example, there are two conversations at the same time, both of which are long. If the two conversations are halfway through and conversation a cuts into conversation b, the cache kv in conversation b should be lost, that is, the cache kv of conversation a is used. Due to the involvement of gpu computing and insufficient resources, verification cannot be carried out

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No files, tests, or entry points are identified in the issue. Start by locating the multi-threaded GPU serving and KV-cache management paths, then investigate two concurrent long conversations under insufficient GPU resources; done would require verified behavior or a documented answer about cache preemption and invalidation.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
ai
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.