Qwen 3.8 27B at 67 tok/s on Two $78 Nvidia Tesla P100s
AI 摘要 Running a high-performance 27B parameter LLM at home is possible without expensive consumer hardware by repurposing budget-friendly, older server components. By utilizing two $78 Nvidia Tesla P100 cards, it is possible to achieve 67 tokens per second, significantly exceeding the…
- Qwen 3.8 27B
- VRAM
- Quantization
English摘要 · 转录












