百元级废矿卡跑大模型实测
推荐指数 65.0 NO. 012 · 2026.07.14
发布2026/07/13Score76Comments24
为什么值得看
对15张退役NVIDIA Tesla企业卡（K80/P100/V100）进行现代AI负载基准测试，验证60-200美元二手卡在LLM推理和训练中的实际可用性。为预算有限的AI工程师和创业者提供高密度、低成本的GPU集群搭建方案，VRAM单价远低于消费级显卡。
编辑判断
消费级显卡被禁运和涨价后，二手企业卡成了国内小团队绕开算力封锁的暗线。K80虽然架构老旧，但24GB显存能跑7B模型量化版，关键是PCIe被动散热设计可以塞进4U机箱密集部署，这是RTX 4090做不到的。
真正的问题是功耗和电费账：K80双芯250W，跑一年电费可能超过卡本身。建议只考虑P100或V100，后者有Tensor Core且支持FP16，性价比曲线在二手市场目前是最优解。作者提到的定制散热器值得关注，企业卡改风冷是这类方案能否规模化的瓶颈。
社区反馈
正面 24 条评论
核心争论：废矿卡性价比 vs 新卡能效比，以及极限堆VRAM的可行性边界
相关内容
Great read. I'd love to know more about how power consumption changes as cards get newer too!
Thanks! Yeah this is a major consideration. I have looked at power consumption throughout runs in the past (https://esologic.com/gpu-server-benchmark/#gpu-box-benchmark) and found that for many of these enterprise class cards, they're happy to slam right into the max TDP. So, for
Darn, I was hoping to see bc-250's (aka PS5 chips) in there. They've recently become popular for inference and they are only about $200 on ebay. They hold a special place in my heart because I deployed 20k of them and I'm glad to see they are finding a purpose now and not just e-waste.