AMD ROCm 5.5 Now Available on GitHub

[erek]

It's released

"Surprisingly, the release notes do not officially mention RDNA 3 improvements in its release notes, but those have been already tested and confirmed. The GPU support list is pretty short including AMD GFX9, RDNA, and CDNA GPUs, ranging from Radeon VII, Pro VII, W6800, V620, and Instinct lineup. The release notes do mention new HIP enhancements, enhanced stack size limit, raising it from 16k to 128k, new APIs, OpenMP enhancements, and more. You can check out the full release notes, downloads, and more details over at GitHub."

Source: https://www.techpowerup.com/308091/amd-rocm-5-5-now-available-on-github

 

[sc5mu93]

they announced they are bringing ROCm to windows in a separate announcement. has anyone seen anything about that yet?

 

[Ready4Dis]

Can't find any benchmarks for it yet, any idea how much improved it may be over older versions? Was thinking about stuffing an AMD card into my homelab server to run stable diffusion and possibly some other ML workloads (like image recognition and training).

 

[Lakados]

Can't find any benchmarks for it yet, any idea how much improved it may be over older versions? Was thinking about stuffing an AMD card into my homelab server to run stable diffusion and possibly some other ML workloads (like image recognition and training).

Pharonix is working on something it seems, but their forums are not impressed with this update at all and are complaining it lacks features and support for things that were promised years ago.

 

[LukeTbk]

Can't find any benchmarks for it yet, any idea how much improved it may be over older versions? Was thinking about stuffing an AMD card into my homelab server to run stable diffusion and possibly some other ML workloads (like image recognition and training).

Maybe you already seen it, but apparently some models have really food stable diffusion performance per dollars: