Is there a way to select the gpu for channels dvr to use in a multiple GPU system during hardware encoding. I currently have an internal intel mobo vid card and a 10 series nvidia which are both enabled. I havent found a way to allocate a gpu for hardware acceleration in a VM or I would address it that way.
If you use the containerized version, you can selectively choose which device is added into the container image. The same for using a systemd-nspawn instance.
If you are using qemu, you can do a PCI passthrough, but in that instance the GPU will be unavailable on the host system, and only available to the VM.