// RUN โ€” FITS

Can the AMD Instinct MI455X run ByT5-XXL?

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

ByT5-XXL on the AMD Instinct MI455X

QuantFits?Max contextSpeedBeyond context
FP16 yes 2259k 478.6-585 t/s ~1-1.2 (slow)
Q8 yes 2338k 957.2-1169.9 t/s ~2-2.4 (slow)
Q4 yes 2373k 1709.3-2089.1 t/s ~3.5-4.3 (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 ByT5-XXL?โ–ถ

Yes. At Q4 it generates ~1709.3-2089.1 tok/s and fits up to 2373k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~3.5-4.3 tok/s.

How much context fits?โ–ถ

At Q4, up to about 2373k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.