I have a couple systems on which I want to run some software that does GLX operations on images. These seem to rely on the X server having access to a GPU. Without this, they do their operations very very slowly. I have a very decent NVIDIA GPU (GeForce GTX 1660 SUPER) installed in all systems. In my own applications that are specifically compiled to use an NVIDIA GPU (no X server involved), all works great. nvidia-smi correctly shows that my application is using the GPU. The problem I am having with other applications (e.g., nona, a part of hugin) is that it seems that the NVIDIA driver is not loading. In this case, the system does not have a display attached. For my own apps that use the GPU direct, this is no problem. It uses libcuda only. No X. For the problem apps, since the nvdia driver is not loading, the X driver it falls back on does not do GPU. Is it a requirement for the nvidia proprietary drivers that a display be connected? I have googled this, but all discussions are about connected displays not being found or used. This is on Leap 15.1 and NVIDIA 440.82 drivers. If a display is required, could it perhaps be defined in a config file so the nvidia driver loads? We have a cluster of computers like this and we do not want to hook up a display just for this. -- Roger Oberholtzer -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org