New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the Instinct MI430X run GPT-2 (124M)?
Yes — here's how fast and up to what context, with real measured numbers.
GPT-2 (124M) on the Instinct MI430X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 6038k | 59189.5-72342.7 t/s | ~101.6-124.2 (slow) |
| Q8 | yes | 6040k | 118379-144685.5 t/s | ~203.2-248.4 (slow) |
| Q4 | yes | 6041k | 211391.1-258366.9 t/s | ~362.9-443.5 (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 Instinct MI430X run GPT-2 (124M)?▶
Yes. At Q4 it generates ~211391.1-258366.9 tok/s and fits up to 6041k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~362.9-443.5 tok/s.
How much context fits?▶
At Q4, up to about 6041k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.