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 FvFold?
Yes — here's how fast and up to what context, with real measured numbers.
FvFold on the Instinct MI430X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 6041k | 796923.9-974018.1 t/s | ~1368.1-1672.1 (slow) |
| Q8 | yes | 6041k | 1593847.8-1948036.2 t/s | ~2736.2-3344.3 (slow) |
| Q4 | yes | 6041k | 2846156.7-3478636 t/s | ~4886.1-5971.9 (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 Instinct MI430X run FvFold?▶
Yes. At Q4 it generates ~2846156.7-3478636 tok/s and fits up to 6041k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~4886.1-5971.9 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.