GPUs & Macs › CIX CD8180
CIX CD8180: local LLM speed
Llama 2 7B (Q4_0) generates 5.5 tokens/s on the CIX CD8180 (Vulkan), and processes prompts at 3 tokens/s.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 2 7B | 5.5 | 3 | Q4_0 | Vulkan FA off | tg128 | llama.cpp #10879 details |
What fits on the CIX CD8180
Estimates for this chip are scaled from a measured run: Llama 2 7B Q4_0 at 5.5 tok/s (Vulkan, source). Fit assumes the whole model in GPU memory with ~8K context.
Memory size unknown for this device.