🔌
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 GMKtec EVO-X3 run Megatron-LM (355M)?
Yes — here's how fast and up to what context, with real measured numbers.
Megatron-LM (355M) on the GMKtec EVO-X3
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1724k | 234.4-284.8 t/s | ~9.2-11.1 (slow) |
| Q8 | yes | 1730k | 468.7-569.7 t/s | ~18.3-22.3 (slow) |
| Q4 | yes | 1732k | 837-1017.3 t/s | ~32.7-39.7 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
FAQ
Can the GMKtec EVO-X3 run Megatron-LM (355M)?▶
Yes. At Q4 it generates ~837-1017.3 tok/s and fits up to 1732k context; beyond that it spills to SSD swap and slows to ~32.7-39.7 tok/s.
How much context fits?▶
At Q4, up to about 1732k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.