// RUN โ€” FITS

Can the NVIDIA RTX PRO 6000 Blackwell run CLIP (ResNet-50)?

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

CLIP (ResNet-50) on the NVIDIA RTX PRO 6000 Blackwell

QuantFits?Max contextSpeedBeyond context
FP16 yes 1322k 7382.4-8798.2 t/s ~164.8-196.4 (slow)
Q8 yes 1323k 14764.8-17596.4 t/s ~329.6-392.8 (slow)
Q4 yes 1324k 26365.7-31422.1 t/s ~588.5-701.4 (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 RTX PRO 6000 Blackwell run CLIP (ResNet-50)?โ–ถ

Yes. At Q4 it generates ~26365.7-31422.1 tok/s and fits up to 1324k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~588.5-701.4 tok/s.

How much context fits?โ–ถ

At Q4, up to about 1324k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.