Date: 07.08.2025

AMD Instinct Mi50 GPU

Hot links

If not exist check archive.org

• Download driver
• Install installer
• Install driver option 1
• Install driver option 2

Limitations

• Linux only, there is no driver for windows
• This GPU is considered outdated; future versions of ROCm may drop support for it
• Required external fun

Test environment

• Workstation 40 GB RAM, 200GB SSD, 750W Power supply
• Ubuntu 24.04 LTS HWE Kernel
• Install python 3.12

My test environment: HP Z440

Ubuntu preparation

sudo apt-get install --install-recommends linux-generic-hwe-24.04
hwe-support-status --verbose
sudo apt dist-upgrade
sudo reboot

Driver setup

• Install drivers

  gpg --keyserver hkp://keyserver.ubuntu.com:80 --recv 9386B48A1A693C5C
  gpg --export --armor 9386B48A1A693C5C | sudo apt-key add -
  cd /tmp
  wget https://repo.radeon.com/amdgpu-install/6.4.2.1/ubuntu/noble/amdgpu-install_6.4.60402-1_all.deb
  sudo apt install ./amdgpu-install_6.4.60402-1_all.deb
  sudo amdgpu-install --usecase=workstation --vulkan=pro --opencl=rocr
  sudo apt install rocm-smi clinfo rocminfo
  

• Add your user to the required groups to enable access to ROCm drivers and GPU hardware

  sudo groupadd kfd
  sudo usermod -aG video $USER
  sudo usermod -aG render $USER
  sudo usermod -aG kfd $USER
  

• Check installation

  clinfo
  sudo /opt/rocm/bin/rocminfo
  sudo rocm-smi
  

Check ROCm in python

• Priparing PyTorch

  mkdir -p ~/llm && cd ~/llm
  python3 -m venv .venv_llm
  source ./.venv_llm/bin/activate
  python -m pip install --upgrade pip
  pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.0
  python3 -c "import torch; print(torch.__version__); print(torch.cuda.is_available()); print(torch.version.hip);print(torch.cuda.get_device_name(0));"
  

• Expected responce

  2.4.1+rocm6.0
  True
  6.0.32830-d62f6a171
  AMD Instinct MI50/MI60
  

Done!

Your workstation ready to test most common models from huggingface.co

For example: Mistral 7b ROCm PyTorch Test