nvme

141 gigabytes have to move from somewhere to eight GPUs, and every hop has a bandwidth. Add the CUDA init, the engine warm-up and the graph capture, and the minute is gone unless you design each step.

A paused conversation holds gigabytes of cache that its next turn will need in thirty seconds. Recompute costs prefill; reload costs bandwidth. The comparison is one division per tier, and the answer flips with prompt length.

Reading straight from NVMe into GPU memory skips a copy through host memory, which is worth a lot for one workload and nothing for another. The bandwidth arithmetic for both paths, the two cases where the bounce is the bottleneck, and the far more common case where the CPU has to touch the data anyway.

Six and a half terabytes leaving sixteen thousand GPUs at once is a burst no shared filesystem absorbs, and the fix is to stop trying. What the barrier actually has to wait for, the per-node arithmetic that makes it seconds, and the manifest rule that decides whether a checkpoint is usable at all.

A 70B model's KV cache is 320 KB per token, and a 30k-token conversation is 10 GB that GPU memory cannot keep between turns. The three tiers with their bandwidths, the break-even where reloading a prefix beats recomputing it, the eviction and lookup design, and the traffic shape where the tier is worth its complexity.