🔌
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 MI440X run TurboVLA?
Yes — here's how fast and up to what context, with real measured numbers.
TurboVLA on the Instinct MI440X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 6035k | 30870-37730 t/s | ~63-77 (slow) |
| Q8 | yes | 6039k | 61740-75460 t/s | ~126-154 (slow) |
| Q4 | yes | 6040k | 110250-134750 t/s | ~225-275 (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 MI440X run TurboVLA?▶
Yes. At Q4 it generates ~110250-134750 tok/s and fits up to 6040k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~225-275 tok/s.
How much context fits?▶
At Q4, up to about 6040k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.