🔌
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 RTX 6000 Ada 48GB run ELECTRA?
Yes — here's how fast and up to what context, with real measured numbers.
ELECTRA on the NVIDIA RTX 6000 Ada 48GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 641k | 1046-1246.6 t/s | ~43.6-51.9 (slow) |
| Q8 | yes | 646k | 2091.9-2493.1 t/s | ~87.2-103.9 (slow) |
| Q4 | yes | 648k | 3735.6-4452 t/s | ~155.7-185.5 (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 6000 Ada 48GB run ELECTRA?▶
Yes. At Q4 it generates ~3735.6-4452 tok/s and fits up to 648k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~155.7-185.5 tok/s.
How much context fits?▶
At Q4, up to about 648k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.