Hi all! Red Hat’s CPE team is working on putting together a budget proposal for next year. Is there any “wishlist” hardware for AI enablement, development, or testing which I should ask them to ask for?
Yeah, let them know It would be beneficial to AI in Fedora and the project as a whole (maybe even to world peace) if we could find room in the budget to upgrade my personal workstation.
Specifically the CPU and The HDD.
I feel the i7-13700KF and upgrading the two 8TB hard drives for equivalent m.2 SSD would do the trick.
Faster boot times and no longer being bottlenecked by my CPU would improve productivity and speed up the development of AI in Fedora.
I have a 100-slide PowerPoint ready to go if they want me to present the proposal myself.
( JKJK this is just a joke. At least I think it was funny when it first popped into my head lol 
)
1 Like
Offhand, it would probably make sense to have servers and packager test machines that are equipped with RDNA3 GPUs or CDNA3 accelerators.
2 Likes
You would definitely have a more RESPONSIVE system, of course speed for some task, but you haven’t lived until you experience a Browser or application cut launch times by ~3-5x. . . it feels like running everyting in RAM.
Memes included ?
I would say that for the compatability with ROCm it would have to be CDNA and RDNA has finicky support. . . But since we are on the topic of Accelerators, AMD/XILINX chips have come into play. . .Surely there are cards with those Accelerators that would beneifit AI workloads.
This is gonna sound strange, but Apple hardware also, their SoC’s are pretty decent for AI workstations. A lot of people even develop locally on macOS but deploy to x86_64 Linux in the cloud.
We are even talking about supporting macOS native in some of latest AI offerings.
1 Like