New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the NVIDIA B200 run DARTS (second order) (PTB)?
Yes — here's how fast and up to what context, with real measured numbers.
DARTS (second order) (PTB) on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2672k | 126956.5-151304.3 t/s | ~634.8-756.5 (slow) |
| Q8 | yes | 2672k | 253913-302608.7 t/s | ~1269.6-1513 (slow) |
| Q4 | yes | 2672k | 453416.1-540372.7 t/s | ~2267.1-2701.9 (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 B200 run DARTS (second order) (PTB)?▶
Yes. At Q4 it generates ~453416.1-540372.7 tok/s and fits up to 2672k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~2267.1-2701.9 tok/s.
How much context fits?▶
At Q4, up to about 2672k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.