// RUN โ€” FITS

Can the Apple M1 Ultra 64GB run BIDAF?

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

BIDAF on the Apple M1 Ultra 64GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 856k 104615.4-126153.8 t/s ~1307.7-1576.9 (slow)
Q8 yes 856k 209230.8-252307.7 t/s ~2615.4-3153.8 (slow)
Q4 yes 856k 373626.4-450549.5 t/s ~4670.3-5631.9 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.

FAQ

Can the Apple M1 Ultra 64GB run BIDAF?โ–ถ

Yes. At Q4 it generates ~373626.4-450549.5 tok/s and fits up to 856k context; beyond that it spills to SSD swap and slows to ~4670.3-5631.9 tok/s.

How much context fits?โ–ถ

At Q4, up to about 856k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.