Agent workload simulation and cost model
A discrete-event simulator for LLM agent workloads on a serving cluster: how many concurrent agents the cluster sustains, what fan-out and synchronization cost, and which scheduler …
•
3 min read
Read more
A discrete-event simulator for LLM agent workloads on a serving cluster: how many concurrent agents the cluster sustains, what fan-out and synchronization cost, and which scheduler …