// RUN โ€” FITS

Can the AMD Radeon Pro W7900 48GB run MegaMolBART?

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

MegaMolBART on the AMD Radeon Pro W7900 48GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 650k 6048-7392 t/s ~280-342.2 (slow)
Q8 yes 650k 12096-14784 t/s ~560-684.4 (slow)
Q4 yes 651k 21600-26400 t/s ~1000-1222.2 (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 Radeon Pro W7900 48GB run MegaMolBART?โ–ถ

Yes. At Q4 it generates ~21600-26400 tok/s and fits up to 651k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1000-1222.2 tok/s.

How much context fits?โ–ถ

At Q4, up to about 651k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.