// RUN โ€” FITS

Can the NVIDIA L40S 48GB run Stable Cascade?

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

Stable Cascade on the NVIDIA L40S 48GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 247k 61.6-73.4 t/s ~2.9-3.4 (slow)
Q8 yes 286k 123.2-146.8 t/s ~5.7-6.8 (slow)
Q4 yes 304k 220-262.2 t/s ~10.2-12.1 (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 Stable Cascade?โ–ถ

Yes. At Q4 it generates ~220-262.2 tok/s and fits up to 304k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~10.2-12.1 tok/s.

How much context fits?โ–ถ

At Q4, up to about 304k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.