Reproduce “PrefixPlace: Provable Prefix Key-Value Placement for Large Language Model Serving under Heterogeneous Compute and Transfer Costs” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhiyu Wang, Rajkumar Buyya
PrefixPlace plans placements of reusable language model attention states across workers. It compares profiled recomputation and transfer costs while enforcing memory budgets and prefix-complete targets. In 432 three-worker instances, it achieved a mean of 99.84% of exact optimum. In a RAG replay, modeled savings were 40.29% higher than vLLM-APC; these are not end-to-end latency results, and routing remains an input.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.