Rollout
In reinforcement learning, generating a trajectory of model actions and outcomes used to compute training rewards.
In other contexts, Rollout also means A staged deployment of a change — firmware, driver, configuration, or a new site — gated by canaries, health checks, and a tested way to stop or roll back. (Parts 6.4, 14.8, 14.14).
Current numbers
~80%of wall-clock spent on rollout generation in agentic/reasoning RL post-training
10K–100K+tokens per RL trajectory for reasoning/agentic tasks — the rollout that dominates cost
~8:1 → ~2:1 and belowGPU:CPU norm rebalancing toward more CPU per node as agentic RL adds rollout/tool/env load