Hello
I am a Linux user and used to my desktop environment settings(workspaces, shortcuts, editor, etc).
I need to develop a plugin/extension for a 3D modeling software(Rhino3D) that runs on windows. The plug-in will interact with a server that uses deep learning, I will also develop that. The server will run on Linux.
All the development is supposed to be done in one machine. Except the execution of the code that train the DL models, that will be done in the cloud. Locally we code, check if few iterations of the training are ok, do debugging and inference.
So I have gpu demand from Rhino on windows and from my DL development environment on Linux. I might alternate between them with a considerable frequency.
I am trying to figure what machine setup fits this situation. I was naive, thinking that with virtual box would be possible to just create a VM with some amount of memory of my gpu. Maybe would is possible with a high end quadro and VMware, but …
The budget probably will allow only a single RTX2060 with 8Gb or something similar on this local machine. Btw I still haven’t bought any piece of hardware yet.
So now I am now into this rabbit hole of GPU passthrough
What do you guys suggest?
I am wandering of how hard is to dynamically switch what GPU is sent to the host.
I also have a doubt , if its possible to pass the ryzen 5 3400g integrated graphics to the guest OS instead of the RTX.
Thanks in advance!