Hello, I am looking at starting up my Proxmox server (its been long time coming on his project) I finally managed to snag a 1U server that supports up to 4 GPU’s on eBay for a reasonable price. I want to pick up a tesla card (Pascal) and instead of using vGPU, pass the whole thing through to a docker VM (yes I know Docker in a VM is a little redundant) specifically for NVENC and other accelerated tasks for various services like Jellyfin (an Emby fork), OBS, and other things that can take advantage of a GPU. I plan on adding 2 more GPUs later and use them with the vGPU unlock script for other VM’s but that is not a concern right not. What I want to know is there anything I should be aware of as far as any issues/concerns I may have with passing through a Pascal based Tesla, to a docker? I’ve read elsewhere I should not have any issues with multiple Dockers sharing GPU resources. I just want to make absolutely sure before I go and blow money on a GPU right now.
This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.