News for hackers. Stuff that matters.
Re:Ollama is now available on Windows in preview by freedomben on Saturday February 17, 2024 @02:49PM (#39412803)
I've had no end of difficulty installing the Pro drivers and/or ROCm. The "solution" that was recommended was to install a different distro (I use Fedora and installing CentOS or Ubuntu was recommended). When I finally could get it installed, I got kernel panics and my system frequently became unbootable. Then once it was installed, getting user space programs to recognize it was the next major pain point.
Re:Ollama is now available on Windows in preview by michaelmrose on Saturday February 17, 2024 @03:53PM (#39413451)
I've been using Nvidia and it stopped being challenging in about 2006. I hear perpetually that Nvidia is horrible and I should try AMD. The 2 times I did admitted a long time ago it was... not great.
Re:Ollama is now available on Windows in preview by freedomben on Saturday February 17, 2024 @04:02PM (#39413524)
Do you use Ubuntu LTS? If so, then indeed Nvidia is not a problem.

But if you run a distro that has anywhere near new kernels such as Fedora and Arch, you'll be constantly in fear of receiving new kernel updates. And every so often the packages will be broken and you'll have to use Nvidia's horrible installer. Oh and every once in a while they'll subtly drop support for older cards and you'll need to move to the legacy package, but the way you'll find out is that your system suddenly doesn't boot and you just happen to think about it being the old Nvidia card so you Kagi that and discover the change.

Re:Ollama is now available on Windows in preview by 65a on Saturday February 17, 2024 @04:52PM (#39413893)
I found it much easier to make ROCm/AMD work for AI (including on an laptop) than getting nvidia work with Xorg on an optimus laptop with an intel iGPU/nvidia dGPU. I swore off nvidia at that point.
Re:Ollama is now available on Windows in preview by michaelmrose on Saturday February 17, 2024 @06:46PM (#39414629)
Changing kernels automatically as new releases came out was never an optimal strategy even if its what you get by default in Arch. Notably arch has linux-lts presently at 6.6 whereas mainline is 6.7.

Instead of treating it like a dice roll and living in existential dread at the entirely predictable peril of Linus cutting releases that necessarily occasionally front run NVIDIA which releases less frequently I simply don't install kernels first released yesterday, pull in major kernel version updates daily, don't remove the old kernel automatically when the new one is installed, and automatically make snapshots on update against any sort of issue that might obtain.

If that seems like too much work one could simply at least keep the prior kernel version around and reboot and your only out 45 seconds of your life. This actually seems like a good idea no matter what.

I don't think I have used nvidia's installer since 2003 on Fedora "Core"–as the nomenclature used to be—One. One simply doesn't need to. Also generally speaking one doesn't need to use a legacy package until a card is over 10 years old. For instance the oldest consumer card unsupported right now is a 600 series from 2012.

If you still own a 2012 GPU you should probably put it where it belongs in the trash but when you get to the sort of computers that require legacy support which is 2009-2012 you are apt to need to worry about other matters like distros that still support 32 bit, simple environments like xfce, software that works well in ram constrained environments. Needing to install a slightly different driver seems tractable.

Re:Ollama is now available on Windows in preview by spookie on Saturday February 17, 2024 @08:20PM (#39415212)
Try to use the runfile provided by Nvidia and use DKMS. The biggest issue is just that flatpaks aren't really updated for CUDA drivers, but you can just not use them if your distro isn't old or niche.
Re:Ollama is now available on Windows in preview by slavik81 on Saturday February 17, 2024 @10:43PM (#39415972)
On Fedora 40, I believe you can install llama.cpp's ROCm dependencies with:

    dnf install hipcc rocm-hip-devel rocblas-devel hipblas-devel
Re:Ollama is now available on Windows in preview by slavik81 on Sunday February 18, 2024 @01:23AM (#39416699)
So, after a bit of experimentation, it seems that Fedora is built primarily for RDNA 3 while Debian is built for RDNA 2 and earlier. These are llama-cpp build instructions for Fedora: https://gist.github.com/cgmb/bb661fccaf041d3649f9a90560826eb.... These are llama-cpp build instructions for Debian: https://gist.github.com/cgmb/be113c04cd740425f637aa33c3e4ea3....
Re:Ollama is now available on Windows in preview by freedomben on Monday February 19, 2024 @12:07AM (#39426416)
Great, thanks I will give this a try once I upgrade!