If you had a bunch of GPUs lying around, what would you actually build with them? (Running LLMs is off the table) [D]
热度趋势
趋势数据积累中
百分比基于当前可用热度信号，而非评论数或独立用户人数。
一个开发者社区的讨论提出了一个问题：如果拥有大量高端GPU，在排除运行大型语言模型（LLMs）的情况下，会用它们来构建什么？参与者被鼓励分享具体且“略显疯狂”的想法，尤其欢迎那些涉及研究和新技术的建议。该讨论旨在收集对这一假设情景的诚实和创新性回应。
Be honest if someone dropped a stack of high-end GPUs on your desk tomorrow, what would you actually do with them?
And before the usual answers roll in: running local LLMs is banned for this thread. It’s been done to death and feels pretty pointless at this point.
So… what else?
- Some niche scientific/simulation workload?
- Weird generative stuff that isn’t text?
- Distributed something-or-other?
- Rendering / media pipeline?
- Homelab experiments that actually need the horsepower?
- Completely unhinged personal projects?
Drop your ideas. The more specific (and slightly unhinged), the better.
Great Ideas but are there some with more of research and new tech.