StageOnce: Speculative Cross-Job Data Staging for LLM Agent Fleets

Speculative cross-job data staging for fleets of LLM agents.

Abstract only until it is presented at SC ’26.

Scientific workflows increasingly run as fleets of LLM agents analysing the same large immutable datasets. On a shared compute node, independent staging repeats source reads and creates storage contention, while each job discovers its inputs during reasoning. StageOnce coordinates these speculative path predictions through a node-local daemon, coalescing concurrent misses into one immutable copy while preserving dataset versions and private outputs. Tools read staged inputs directly from the filesystem, without per-read coordination. Recorded-session replays in genomics and astronomy on SSD and HDD backends reduce fleet makespan by 31–51% at 12 concurrent jobs against independent staging. Logical staging remains one copy across every tested fleet size. An ablation isolates speculation, which reduces makespan by 8–49% against reactive shared caching.