🔌
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 Falcon Perception?
Yes — here's how fast and up to what context, with real measured numbers.
Falcon Perception on the NVIDIA DGX Spark
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1717k | 166.1-197.9 t/s | ~6.1-7.3 (slow) |
| Q8 | yes | 1726k | 332.2-395.9 t/s | ~12.2-14.5 (slow) |
| Q4 | yes | 1730k | 593.1-706.9 t/s | ~21.7-25.9 (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 Falcon Perception?▶
Yes. At Q4 it generates ~593.1-706.9 tok/s and fits up to 1730k context; beyond that it spills to SSD swap and slows to ~21.7-25.9 tok/s.
How much context fits?▶
At Q4, up to about 1730k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.