Hiya folks,

I have always installed the Nvidia blob when installing Nvidia drivers...it's just easier, IMHO.

Now, I would like to install cuda, and the slackbuild mentions that I must have the nvidia driver/kernel driver installed. Does this mean that I have to have those installed via the slackbuilds way, or that the cuda build will find my already installed nvidia blob?

Every time I've tried to install the nvidia drivers the slackbuilds way, it's never worked right and I always ended up having to reinstall the Slackware install.

So if anyone happens to know if the blob will work, I'd sure appreciate hearing it. Reason being, I got the Krita appimage of 5.2.2 on my system and have the AI extensions installed, but if I try to use anything 'AI', it throws out a CUDA error, presumably because I don't have CUDA installed.

In my experience the SlackBuild scripts for nVidia driver and kernel module has worked fine and has also been possible to uninstall. That is IMHO the major advantage of using the SlackBuild scripts instead of using the .run file provided by nVidia directly which might break OpenGL libraries when the driver is uninstalled.

That said, there is no clear definition of the "slackbuilds way". Basically, slackbuilds.org will provide you with SlackBuild scripts. Those scripts will most likely work for you if you manually have resolved all dependencies even though not by installing packages created by SlackBuilds from slackbuilds.org.

Slackbuilds.org will also provide .info files for each SlackBuild script. Those files will list dependencies and tools like slpkg will by default use the information in those files to resolve the dependencies with other SlackBuild scripts. You might be able to tell your favorite tool to not care about dependencies.

regards Henrik