Reproduce “LaFlash: Capacity-Driven Acceleration for Coding Agent Workflows via QLC-HBF” 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
Yuseon Choi, Eunjong Lee, Sanghyeon Park, Seongyon Hong, Sangjin Kim, Thierry Tambe
LaFlash accelerates coding-agent workloads by reserving memory for dormant states, including context and KV caches, rather than using extra capacity only for requests. Capacity-Driven Acceleration leaves room for reusable states, while its QLC-HBF layout balances flash reads and block use. Trace-driven simulations report average 7.04× throughput and 10.96× energy efficiency over SLC-HBF with batch expansion; results are modeled, not prototype measurements.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.