🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the NVIDIA DGX Spark run FvFold?
Yes — here's how fast and up to what context, with real measured numbers.
FvFold on the NVIDIA DGX Spark
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1735k | 10819.5-12894.4 t/s | ~396.3-472.3 (slow) |
| Q8 | yes | 1735k | 21638.9-25788.9 t/s | ~792.6-944.6 (slow) |
| Q4 | yes | 1735k | 38641-46051.5 t/s | ~1415.4-1686.9 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
FAQ
Can the NVIDIA DGX Spark run FvFold?▶
Yes. At Q4 it generates ~38641-46051.5 tok/s and fits up to 1735k context; beyond that it spills to SSD swap and slows to ~1415.4-1686.9 tok/s.
How much context fits?▶
At Q4, up to about 1735k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.