Would it be possible to change the "intel" profile to no longer alias the nvidia.ko and nvidia-uvm.ko modules to "off"? The blacklisting can remain, and would prevent the modules from being loaded automatically in response to the device being present, but wouldn't prevent them from being explicitly loaded if desired.
Since I imagine the reason for preventing the modules from loading is to allow the NVIDIA GPU to be turned off via ACPI, would it be possible to attempt unloading the modules before attempting to shut off the GPU? If they fail to unload, it's presumably because they're in use, so shutting off the GPU doesn't make sense, anyway. Going the other way, it could also be useful to add a udev rule or some other trigger to turn the GPU back on if it's off and somebody tries to load nvidia.ko. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1824282 Title: Add a mode that load nvidia.ko only so that we display on intel and do cuda deep learning in NV GPU To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/nvidia-prime/+bug/1824282/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
