// RUN โ€” FITS

Can the NVIDIA GeForce RTX 5090 run CLIP (ViT L/14@336px)?

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

CLIP (ViT L/14@336px) on the NVIDIA GeForce RTX 5090

QuantFits?Max contextSpeedBeyond context
FP16 yes 415k 1767.8-2106.8 t/s ~39.5-47 (slow)
Q8 yes 421k 3535.6-4213.6 t/s ~78.9-94.1 (slow)
Q4 yes 423k 6313.5-7524.3 t/s ~140.9-168 (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 5090 run CLIP (ViT L/14@336px)?โ–ถ

Yes. At Q4 it generates ~6313.5-7524.3 tok/s and fits up to 423k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~140.9-168 tok/s.

How much context fits?โ–ถ

At Q4, up to about 423k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.