// RUN โ€” FITS

Can the NVIDIA L40S 48GB run Grover-Mega?

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

Grover-Mega on the NVIDIA L40S 48GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 605k 210.2-250.6 t/s ~9.7-11.6 (slow)
Q8 yes 628k 420.5-501.1 t/s ~19.5-23.2 (slow)
Q4 yes 638k 750.9-894.9 t/s ~34.8-41.4 (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 NVIDIA L40S 48GB run Grover-Mega?โ–ถ

Yes. At Q4 it generates ~750.9-894.9 tok/s and fits up to 638k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~34.8-41.4 tok/s.

How much context fits?โ–ถ

At Q4, up to about 638k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.