Hi all,
I would like to build and run tests for a GPU-enabled nvidia/cuda program on CircleCI. I have read the documentation parts about GPU-Runners, and it seems that it is only possible on my own server with gpu-instances. Is this correct? Is there a way to get gpu-enabled instances on the standard CircleCI app without having to setup a server?
Thank you for your help!
The closest that CircleCI have to real machines is, I believe, the Machine executor. I would be surprised if that is GPU enabled, since it is probably still a virtual machine (e.g. VM Ware or similar, rather than Docker).
OK thanks! That’s what I thought …
Maybe you can get a GPU-enabled VPS on the web (e.g. something like this), and use CircleCI to orchestrate some remote tests on it? Or maybe there is an emulator framework for GPU?
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.