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 Routing Transformer (WT-103)?
Yes — here's how fast and up to what context, with real measured numbers.
Routing Transformer (WT-103) on the Intel Gaudi 3
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1772k | 13496.9-16754.7 t/s | ~145.9-181.1 (slow) |
| Q8 | yes | 1773k | 26993.7-33509.4 t/s | ~291.8-362.3 (slow) |
| Q4 | yes | 1773k | 48203.1-59838.3 t/s | ~521.1-646.9 (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 Routing Transformer (WT-103)?▶
Yes. At Q4 it generates ~48203.1-59838.3 tok/s and fits up to 1773k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~521.1-646.9 tok/s.
How much context fits?▶
At Q4, up to about 1773k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.