// RUN — FITS

Can the Instinct MI430X run AWD-LSTM-DRILL + dynamic evaluation† (WT2)?

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

AWD-LSTM-DRILL + dynamic evaluation† (WT2) on the Instinct MI430X

QuantFits?Max contextSpeedBeyond context
FP16 yes 6041k 215867.6-263838.2 t/s ~370.6-452.9 (slow)
Q8 yes 6041k 431735.3-527676.5 t/s ~741.2-905.9 (slow)
Q4 yes 6041k 770955.9-942279.4 t/s ~1323.5-1617.6 (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 Instinct MI430X run AWD-LSTM-DRILL + dynamic evaluation† (WT2)?

Yes. At Q4 it generates ~770955.9-942279.4 tok/s and fits up to 6041k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1323.5-1617.6 tok/s.

How much context fits?

At Q4, up to about 6041k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.