// RUN โ€” FITS

Can the Intel Arc B580 12GB run Falcon-OCR?

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

Falcon-OCR on the Intel Arc B580 12GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 136k 440.8-547.2 t/s ~38.7-48 (slow)
Q8 yes 141k 881.6-1094.4 t/s ~77.3-96 (slow)
Q4 yes 143k 1574.3-1954.3 t/s ~138.1-171.4 (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 Intel Arc B580 12GB run Falcon-OCR?โ–ถ

Yes. At Q4 it generates ~1574.3-1954.3 tok/s and fits up to 143k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~138.1-171.4 tok/s.

How much context fits?โ–ถ

At Q4, up to about 143k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.