// RUN โ€” FITS

Can the NVIDIA GeForce RTX 4090 run BART-IT?

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

BART-IT on the NVIDIA GeForce RTX 4090

QuantFits?Max contextSpeedBeyond context
FP16 yes 310k 2628-3132 t/s ~104.3-124.3 (slow)
Q8 yes 312k 5256-6264 t/s ~208.6-248.6 (slow)
Q4 yes 313k 9385.7-11185.7 t/s ~372.4-443.9 (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 NVIDIA GeForce RTX 4090 run BART-IT?โ–ถ

Yes. At Q4 it generates ~9385.7-11185.7 tok/s and fits up to 313k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~372.4-443.9 tok/s.

How much context fits?โ–ถ

At Q4, up to about 313k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.