New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the NVIDIA H200 run DDPM-IP (CelebA)?
Yes — here's how fast and up to what context, with real measured numbers.
DDPM-IP (CelebA) on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1947k | 5939-7078 t/s | ~49.5-59 (slow) |
| Q8 | yes | 1952k | 11878-14155.9 t/s | ~99-118 (slow) |
| Q4 | yes | 1954k | 21210.7-25278.5 t/s | ~176.8-210.7 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to system RAM offload (depends on your system RAM) and speed drops to the "beyond" figure.
FAQ
Can the NVIDIA H200 run DDPM-IP (CelebA)?▶
Yes. At Q4 it generates ~21210.7-25278.5 tok/s and fits up to 1954k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~176.8-210.7 tok/s.
How much context fits?▶
At Q4, up to about 1954k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.