🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the NVIDIA GeForce RTX 5090 run OpenCLIP ViT-H-14-378-quickgelu?
Yes — here's how fast and up to what context, with real measured numbers.
OpenCLIP ViT-H-14-378-quickgelu on the NVIDIA GeForce RTX 5090
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 396k | 662.9-790 t/s | ~14.8-17.6 (slow) |
| Q8 | yes | 411k | 1325.8-1580 t/s | ~29.6-35.3 (slow) |
| Q4 | yes | 418k | 2367.5-2821.5 t/s | ~52.8-63 (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.
→ All models on the NVIDIA GeForce RTX 5090
→ OpenCLIP ViT-H-14-378-quickgelu specs
→ Other hardware
FAQ
Can the NVIDIA GeForce RTX 5090 run OpenCLIP ViT-H-14-378-quickgelu?▶
Yes. At Q4 it generates ~2367.5-2821.5 tok/s and fits up to 418k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~52.8-63 tok/s.
How much context fits?▶
At Q4, up to about 418k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.