Why it matters
CPU offload extends effective memory. Understanding shapes design.
Advertisement
The architecture
Identify what to offload (optim + activations + inference KV).
PCIe transfer overhead.
Prefetch overlap.
Advertisement
How it works end to end
CPU RAM 5-10x GPU HBM.
PCIe 5.0 ~64 GB/s.
NVLink C2C (GB200) 900 GB/s.
Prefetch critical.