New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the Intel Gaudi 3 run DDPM-IP (CelebA)?
Yes — here's how fast and up to what context, with real measured numbers.
DDPM-IP (CelebA) on the Intel Gaudi 3
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1765k | 3637.3-4515.3 t/s | ~39.3-48.8 (slow) |
| Q8 | yes | 1769k | 7274.6-9030.5 t/s | ~78.6-97.6 (slow) |
| Q4 | yes | 1771k | 12990.3-16125.9 t/s | ~140.4-174.3 (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 DDPM-IP (CelebA)?▶
Yes. At Q4 it generates ~12990.3-16125.9 tok/s and fits up to 1771k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~140.4-174.3 tok/s.
How much context fits?▶
At Q4, up to about 1771k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.