New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the Instinct MI455X run Pythia-410m?
Yes — here's how fast and up to what context, with real measured numbers.
Pythia-410m on the Instinct MI455X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 6029k | 15058.5-18404.9 t/s | ~30.7-37.6 (slow) |
| Q8 | yes | 6035k | 30117.1-36809.8 t/s | ~61.5-75.1 (slow) |
| Q4 | yes | 6038k | 53780.5-65731.7 t/s | ~109.8-134.1 (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 Instinct MI455X run Pythia-410m?▶
Yes. At Q4 it generates ~53780.5-65731.7 tok/s and fits up to 6038k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~109.8-134.1 tok/s.
How much context fits?▶
At Q4, up to about 6038k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.