GPU Setup @ Ubuntu22+i7+3090

  • Disable nouveau!!!1

    <!-- sudo gedit /etc/modprobe.d/blacklist-nouveau.conf -->
    blacklist nouveau
    options nouveau modeset=0
    
  • Install driver sudo apt install nvidia-driver-*** to support dual monitors
  • Install CUDA Toolkit 11.82
  • Install cuDNN 8.63

    sudo dpkg -i cudnn-local-repo-ubuntu2204-8.6.0.163_1.0-1_amd64.deb
    sudo cp /var/cudnn-local-repo-ubuntu2204-8.6.0.163/cudnn-local-FAED14DD-keyring.gpg /usr/share/keyrings/
    sudo apt-get update
    sudo apt-get install libcudnn8=8.6.0.163-1+cuda11.8
    sudo apt-get install libcudnn8-dev=8.6.0.163-1+cuda11.8
    sudo apt-get install libcudnn8-samples=8.6.0.163-1+cuda11.8
    
  • (optional) sudo apt install libstdc++-12-dev4
  • a tool: pip3 install --upgrade nvitop5

Toolkit:

  • Python
  • C++
  • CUDA = Compute Unified Device Architecture
  • PyTorch (A replacement for NumPy to use the power of GPUs)
  • pybind11
  • CasADi

A kernel is executed in parallel by an array of threads:

  • All threads run the same code
  • Each thread has an ID that it uses to compute memory addresses and make control decisions

Materials:

Simulator

  • MJX = Brax(Google Research) ❤️ MuJoCo6(DeepMind)
    • JAX7 & CUDA using pip wheels! pip install --upgrade "jax[cuda11_pip]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
  • Tiny Differentiable Simulator
    • C++ & CUDA
  • Isaac Gym
    • Python 3.8!
    • PyTorch & CUDA