// RUN โ€” FITS

Can the AMD Instinct MI455X run Falcon-40B?

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

Falcon-40B on the AMD Instinct MI455X

QuantFits?Max contextSpeedBeyond context
FP16 yes 964k 154.4-188.7 t/s ~0.3-0.4 (slow)
Q8 yes 1086k 308.7-377.3 t/s ~0.6-0.8 (slow)
Q4 yes 1140k 551.3-673.8 t/s ~1.1-1.4 (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 AMD Instinct MI455X run Falcon-40B?โ–ถ

Yes. At Q4 it generates ~551.3-673.8 tok/s and fits up to 1140k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1.1-1.4 tok/s.

How much context fits?โ–ถ

At Q4, up to about 1140k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.