// RUN โ€” FITS

Can the AMD Instinct MI325X run ADM?

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

ADM on the AMD Instinct MI325X

QuantFits?Max contextSpeedBeyond context
FP16 yes 3554k 3381-4132.4 t/s ~22.5-27.5 (slow)
Q8 yes 3562k 6762.1-8264.8 t/s ~45.1-55.1 (slow)
Q4 yes 3566k 12075.1-14758.5 t/s ~80.5-98.4 (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 MI325X run ADM?โ–ถ

Yes. At Q4 it generates ~12075.1-14758.5 tok/s and fits up to 3566k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~80.5-98.4 tok/s.

How much context fits?โ–ถ

At Q4, up to about 3566k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.