// RUN โ€” FITS

Can the FuriosaAI RNGD run Ankh_base?

Yes โ€” here's how fast and up to what context, with real measured numbers.

Ankh_base on the FuriosaAI RNGD

QuantFits?Max contextSpeedBeyond context
FP16 yes 628k 658.8-800.7 t/s ~17.6-21.4 (slow)
Q8 yes 640k 1317.6-1601.4 t/s ~35.1-42.7 (slow)
Q4 yes 645k 2352.8-2859.6 t/s ~62.7-76.3 (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 Ankh_base?โ–ถ

Yes. At Q4 it generates ~2352.8-2859.6 tok/s and fits up to 645k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~62.7-76.3 tok/s.

How much context fits?โ–ถ

At Q4, up to about 645k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.