// RUN โ€” FITS

Can the AMD Instinct MI325X run ProteinBERT?

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

ProteinBERT on the AMD Instinct MI325X

QuantFits?Max contextSpeedBeyond context
FP16 yes 3570k 118125-144375 t/s ~787.5-962.5 (slow)
Q8 yes 3571k 236250-288750 t/s ~1575-1925 (slow)
Q4 yes 3571k 421875-515625 t/s ~2812.5-3437.5 (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 AMD Instinct MI325X run ProteinBERT?โ–ถ

Yes. At Q4 it generates ~421875-515625 tok/s and fits up to 3571k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~2812.5-3437.5 tok/s.

How much context fits?โ–ถ

At Q4, up to about 3571k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.