New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the FuriosaAI RNGD run AWD-LSTM-MoS + dynamic evaluation (PTB, 2018)?
Yes — here's how fast and up to what context, with real measured numbers.
AWD-LSTM-MoS + dynamic evaluation (PTB, 2018) on the FuriosaAI RNGD
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 650k | 20312.5-24687.5 t/s | ~541.7-658.3 (slow) |
| Q8 | yes | 651k | 40625-49375 t/s | ~1083.3-1316.7 (slow) |
| Q4 | yes | 651k | 72544.6-88169.6 t/s | ~1934.5-2351.2 (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 FuriosaAI RNGD run AWD-LSTM-MoS + dynamic evaluation (PTB, 2018)?▶
Yes. At Q4 it generates ~72544.6-88169.6 tok/s and fits up to 651k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1934.5-2351.2 tok/s.
How much context fits?▶
At Q4, up to about 651k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.