Skip to content

Victor-UVA/Servers

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 

Repository files navigation

Victor Lab Servers

(this document is a work in progress)

Server List

Name IP GPU OS Applications
hal9000 172.28.235.138 RTX 3090 Ubuntu 20
wall-e 10.0.0.7 RTX 3080 Ubuntu 20, Windows 10
r2d2 172.28.235.240 RTX 3080 Ubuntu 20
ava 10.0.0.x RTX 3080
chappie 10.0.0.x RTX 3080
t800 10.0.0.x RTX 3080

Hardware Assembly

  1. Flash motherboard (if using same board) https://www.gigabyte.com/Motherboard/X570S-AORUS-ELITE-AX-rev-11/support#support-dl-bios
  2. Put parts together

Software Installation

  1. Install Ubuntu (20.04.4 LTS)

    • Password should not be password
  2. Install SSH server

    sudo apt update
    sudo apt install -y net-tools openssh-server
    sudo systemctl enable ssh
    

    Validate by opening a shell from another computer on the subnet.

    ssh username@ipaddress, "yes", enter password

also ips static etc

  1. Install GPU drivers

    There doesn't seem to be a definitive online resource for every command required to correctly install the necessary software for enabling the GPU for computation. You should follow the instructions from Nvidia for the most part, however a few additional steps are listed below.

    1. Find the latest version for this machine.

      ubuntu-drivers devices

      The above command should output something similar to ...

      == /sys/devices/pci0000:00/0000:00:03.1/0000:0a:00.0 ==
      modalias : pci:v000010DEd00002216sv00003842sd00004897bc03sc00i00
      vendor   : NVIDIA Corporation
      driver   : nvidia-driver-510-server - distro non-free
      driver   : nvidia-driver-510 - third-party non-free
      driver   : nvidia-driver-470 - third-party non-free
      driver   : nvidia-driver-495 - third-party non-free
      driver   : nvidia-driver-515 - third-party non-free recommended
      driver   : nvidia-driver-515-server - distro non-free
      driver   : nvidia-driver-460 - third-party non-free
      driver   : nvidia-driver-470-server - distro non-free
      driver   : xserver-xorg-video-nouveau - distro free builtin

      Since 515 is the highest number, this will be the version we install.

    2. Pre install several packages that seem to not be auto installed later.

      export version=515
      sudo apt update
      sudo apt install -y \
          libnvidia-extra-$version \
          libnvidia-compute-$version \
          nvidia-driver-$version
    3. Read through and follow along with Nvidia's official instructions.

      You can find the name of the dpkg file by navigating the distro website.

    sudo apt update
    sudo apt install cuda
    sudo reboot
    

    Validate by installing and running tensorflow

    sudo apt install python3-pip
    pip install tensorflow-gpu
    echo "alias python=python3" >> ~/.bashrc && source ~/.bashrc
    python
    import tensorflow as tf
    print(tf.config.list_physical_devices('GPU'))
  2. Install ROS

About

Documentation for servers and networking at Victor Lab

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors