Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Feature]支持Nvidia MIG方式支持虚机vGPU功能 #22178

Open
zackzhangkai opened this issue Feb 24, 2025 · 1 comment
Open

[Feature]支持Nvidia MIG方式支持虚机vGPU功能 #22178

zackzhangkai opened this issue Feb 24, 2025 · 1 comment

Comments

@zackzhangkai
Copy link

zackzhangkai commented Feb 24, 2025

What would you like to be added:

  1. vGPU有两种搞法,老的搞法主要是Tesla系列,用的软件实现的方式,要花钱买License;但是不买也可以调试,有20分钟的调试时间,后面还有几天的时间可以用,无非速率降了些;先不着急买;Tesla于2020年后就已经慢慢停产了,后面转向H系列和A系列;

  2. H系列及A系列基本上通过硬件虚拟化的方式支持vGPU,也就是MIG(multi-instance GPU)方式;不需要授权,在平台上先GPU切分好后,挂载到虚拟机上就行;性能好,但是一个GPU最多分8个,基本上也够了,因1:8已经完全够了,对比CPU也就1:10已经超分得不行了; 其实这种硬件方式最开始是AMD的vGPU方式,比Nvidia有优势,Nvidia最新芯片开始转向这种;

  3. Cloudpods支持情况:
    1) 支持Tesla GPU的vGPU,现在看已经过时了;
    2)支持AMD硬件虚拟化;

  4. 需求: GPU MIG硬件虚拟化暂不支持,但是跟AMD的搞法一样,按说只是没有适配;目前看Nvidia发展情况,这是趋势;希望社区能够支持~

https://www.cloudpods.org/docs/guides/onpremise/vminstance/passthrough/vgpu

顺祝Cloudpods越来越好~~

@zexi
Copy link
Member

zexi commented Feb 24, 2025

感谢反馈,我们研究下

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants