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 ERNIE-Doc (247M)?
Yes — here's how fast and up to what context, with real measured numbers.
ERNIE-Doc (247M) on the Instinct MI455X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 6034k | 24996-30550.6 t/s | ~51-62.3 (slow) |
| Q8 | yes | 6038k | 49991.9-61101.2 t/s | ~102-124.7 (slow) |
| Q4 | yes | 6039k | 89271.3-109109.3 t/s | ~182.2-222.7 (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 ERNIE-Doc (247M)?▶
Yes. At Q4 it generates ~89271.3-109109.3 tok/s and fits up to 6039k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~182.2-222.7 tok/s.
How much context fits?▶
At Q4, up to about 6039k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.