🔌
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 Intel Gaudi 3 run Llama 3.2 1B?
Yes — here's how fast and up to what context, with real measured numbers.
Llama 3.2 1B on the Intel Gaudi 3
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1736k | 872.4-1082.9 t/s | ~9.4-11.7 (slow) |
| Q8 | yes | 1755k | 1744.7-2165.9 t/s | ~18.9-23.4 (slow) |
| Q4 | yes | 1763k | 3115.6-3867.6 t/s | ~33.7-41.8 (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 Intel Gaudi 3 run Llama 3.2 1B?▶
Yes. At Q4 it generates ~3115.6-3867.6 tok/s and fits up to 1763k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~33.7-41.8 tok/s.
How much context fits?▶
At Q4, up to about 1763k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.