While we were enabling Nvidia for Bluefin LTS we decided that instead of making an nvidia version for each of these that we would instead focus on just adding Nvidia to the developer experience(dx) image.
This is basically a Bluefin LTS image with nvidia and cuda enabled, so we’ve decided to make an AI/ML workstation out of this image. So we’re going to make bluefin-dx-nvidia:lts be it’s own SKU as Bluefin GDX to signify it’s specific use case.
Is this the same as bluefin-dx-nvidia but based on CentOS instead of Fedora? Or are there more differences.
If it’s the former, could you explain the advantage of each? As I understand with Fedora (gts) nothing can go wrong because the base image is always clean and you can’t end up with a broken or unbootable OS, so what does the CentOS base add compared to Fedora?
Just yesterday, I tried to download and install the iso at Bluefin GDX (Beta) | Bluefin
Got a missing /images/pxeboot/vmlinuz error at boot.
Noticed that the iso is some 2GBs smaller compared to bluefin-dx-nvidia-open-stable.
Hence I think I’ll redownload and retry it once the iso-building issues are resolved.
Cool stuff btw
