#Rusticl + #radeonsi works with #hashcat! The dictionary mode at least.

Trying to use brute force fails on the self test.

Still better than clover, which makes the GPU hang up so badly I can't even turn off the computer.

This is awesome cause this #APU (#Ryzen 2400G) is not supported by #rocm. Rusticl is the best #opencl I can get right now.

The #rocm examples at https://github.com/amd/rocm-examples/blob/develop/HIP-Basic/runtime_compilation/main.hip are nice, but it is annoying that they are in C++ when the underlying API is in C.

Here, have a few #ponies #AI generated today from my local computer setup.

Had issues starting off, because after a crash and not dumping memory. Long story short I generate batches of 4 images at 768x768 at a time, then when I get one I like I move to img2img regenerate for hopefully better image and higher resolution at 1536x1536 - can only do one at a time though

The issue lies in #ROCm not supporting FP16 operations, so uses 2x more memory and takes 2x longer than it should.

Need to run large MPI jobs across multiple #AMD GPUs?

Don't miss our latest #AMDlabnotes blog post covering #GPU-aware MPI with #ROCm support 👇

https://gpuopen.com/learn/amd-lab-notes/amd-lab-notes-gpu-aware-mpi-readme/?utm_source=mastodon&utm_medium=social&utm_campaign=amd-lab-notes

#ROCm 5.5.1 Released For #AMD's #OpenSource Compute Stack
The officially supported GPUs currently with ROCm are the Instinct MI50, Instinct MI60, Radeon VII, Radeon Pro VII, Radeon Pro W6800, Radeon Pro V620, Instinct MI100, and Instinct MI300. While there has been #RDNA3 movement with ROCM, so far the recently announced Radeon Pro W7800/W7900 series are yet to be officially listed.
https://www.phoronix.com/news/AMD-ROCm-5.5.1

@sjstulga I upgraded to a 6800xt and I'm quite happy with the performance. Went 6800+ for the out of box #rocm support without having to wait for rdna3 support to land.

Software just worked out of the box. Drop in upgrade from the vega56 for #pytorch .

#AMD #ROCm 5.5 Released With RDNA3 Improvements, Many Changes

https://www.phoronix.com/news/ROCm-5.5-Released

Original tweet : https://twitter.com/phoronix/status/1653349014429999105

#AMD #ROCm 5.5 In The Process Of Being Released

-- Much better RDNA3 support!

https://www.phoronix.com/news/ROCm-5.5-Releasing

Original tweet : https://twitter.com/phoronix/status/1653143397593501700

@framaka C'est un peu pour ça que j'ai mis en pause ma contribution à l'empaquetage #debian de #ROCm. Le débat est ouvert.
Vu passer cet article de Cédric #Sauviat aujourd'hui en lien, https://basta.media/vendredi-debat-faut-il-resister-a-l-intelligence-artificielle #IA

혹시 AMD 그래픽 카드 쓰시는 분 중에 SD 돌리실 분 있으면... 쓰세요.

https://github.com/perillamint/rocm-docker-sd-webui

#AIArt #ROCm #AMD

#Rusticl es uno de los componentes más prometedores de Mesa, ya que puede brindar un buen soporte de #OpenCL para #AMD #Radeon sin tener que tirar de componentes raros que son soportados por pocos sistemas.

Parece que en rendimiento puede dar alcance a #ROCm, pero habrá que ver si soluciones como DaVinci Resolve dan soporte a través de él, ya que el software profesional es muy tendente a seguir un modelo más bien propio de principios de los años 90 del siglo pasado. https://www.phoronix.com/news/Rusticl-RadeonSI-Near

The end might be neigh for my trusty Vega 56 #GPU. :blobcatsad:

* #ROCm wonder
* #Linux friendly drivers
* ML thesis workhorse
* CIV6/Modded Skyrim time-waster-enabler
* Pandemic survivor

...but playing with #stablediffusion models now causes screen artifacts followed by machine reboots. Every time. Oof.

You will be missed old friend.

Any recommendations on a logical ROCm-supported replacement attainable by mere mortals? RX 6800?

The @gnutools #GDB Debugger Adds Support For Debugging @amdradeon #ROCm / HIP Code

https://www.phoronix.com/news/GDB-ROCm-HIP-Debugging

Original tweet : https://twitter.com/phoronix/status/1621171679643987969

Want to build #ROCm on your #Linux workstation? 🐧

#AMDlabnotes presents a new blog post covering three possible ways to install ROCm.

Take a look now over on on #gpuopen

https://gpuopen.com/learn/amd-lab-notes/amd-lab-notes-rocm-installation-readme/?utm_source=mastodon&utm_medium=social&utm_campaign=amdlabnotes

#AMD #ROCm 5.4.2 Released As Another Small Update To The Compute Stack

https://www.phoronix.com/news/AMD-ROCm-5.4.2

Original tweet : https://twitter.com/phoronix/status/1614227281635061762

Receiving an #AMD RX 7900 #GPU on Wednesday. Rather curious whether #ROCm will work at all on it. I like AMD hardware, but their software often makes me want to rip my hair out...

ROCm in [community]

https://lists.archlinux.org/archives/list/arch-dev-public@lists.archlinux.org/thread/BQ7HYEXRPW2W5PMRCDRYKEFQS7NIDUIW/

#ArchLinux #ROCm #linux

#Radeon #ROCm 5.4.1 Released

https://www.phoronix.com/news/Radeon-ROCm-5.4.1

Original tweet : https://twitter.com/phoronix/status/1603536839347937280

#AMD #ROCm 5.4 Released With HIP Improvements

https://www.phoronix.com/news/AMD-ROCm-5.4-Released#rocm

Original tweet : https://twitter.com/phoronix/status/1598048591586918400

Okay, tried my first hand at #StableDiffusion and the results were pretty interesting. However also pretty slow (10 minutes for a batch of 6 images on my 7950x).

I don't want to mess with #ROCm too much, so I wonder if there is another way to make it run on the GPU. #rusticl / #clvk? Also I started wondering how hard it would be to add enough #vulkan support to #pytorch ? (I always wanted a usecase for VK_NV_cooperate_matrix 🙃 )

Theoretically my GPU should be enough for a 10x speedup ...