🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the Instinct MI430X run KoGPT?
Yes — here's how fast and up to what context, with real measured numbers.
KoGPT on the Instinct MI430X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2927k | 1190.2-1454.7 t/s | ~2-2.5 (slow) |
| Q8 | yes | 2974k | 2380.4-2909.4 t/s | ~4.1-5 (slow) |
| Q4 | yes | 2994k | 4250.8-5195.4 t/s | ~7.3-8.9 (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 KoGPT?▶
Yes. At Q4 it generates ~4250.8-5195.4 tok/s and fits up to 2994k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~7.3-8.9 tok/s.
How much context fits?▶
At Q4, up to about 2994k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.