// RUN โ€” FITS

Can the Intel Gaudi 3 run TRIMELMext (7M)?

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

TRIMELMext (7M) on the Intel Gaudi 3

QuantFits?Max contextSpeedBeyond context
FP16 yes 1774k 153285.7-190285.7 t/s ~1657.1-2057.1 (slow)
Q8 yes 1774k 306571.4-380571.4 t/s ~3314.3-4114.3 (slow)
Q4 yes 1774k 547449-679591.8 t/s ~5918.4-7346.9 (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 Intel Gaudi 3 run TRIMELMext (7M)?โ–ถ

Yes. At Q4 it generates ~547449-679591.8 tok/s and fits up to 1774k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~5918.4-7346.9 tok/s.

How much context fits?โ–ถ

At Q4, up to about 1774k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.