Lab Information

El Capitan is our on-campus Linux server with four NVidia V100 GPUs (32 GB RAM each) dedicated to research use.

Half Dome is our Windows workstation with two Nvidia RTX 3090 GPUs and a Meta Quest 2 headset. This workstation is ideal for running ArcGIS Pro and virtual reality experiments.

Frequently Asked Questions

Here are answers to frequently asked questions about working on the El Capitan server.

How do I log in?

You can log in with your Cal Poly credentials. You need to be on the Cal Poly VPN if you are off campus.

To avoid having to type in your password, you can set up SSH keys:

• On your laptop, see if you already have a public key at ~/.ssh/id_rsa.pub. If not, create it with ssh-keygen -t rsa and accept all the defaults.
• Copy the contents of ~/.ssh/id_rsa.pub to ~/.ssh/authorized_keys on the server.

You should now be able to SSH to the server without entering your password.

Where should I store my files?

You won’t have access to your network home directory on El Capitan, as you would on other Cal Poly servers. Instead, your home directory /home/<username> is on a local 1 TB drive.

El Capitan also has two internal 8TB drives: /code (for code) and /files (for datasets and large files). Since /home has limited storage, you should store the majority of your code under /code/<username> and data on /files.

How do I run a Juptyer notebook?

It is not easy to run a Jupyter server on El Capitan and I don’t recommend it. Instead, you should use VS Code to run notebooks (see next question.)

How do I use VS Code on the server?

You can connect to the server using VS Code through the “Remote - SSH” extension. Please disable the “Typescript and Javascript Language Services” extension before connecting:

• Navigate to “Extensions” (click the building blocks button in the left toolbar)
• Search for @builtin TypeScript
• Disable the “Typescript and Javascript Language Services” extension

How do I install Python packages?

You can’t install packages to the system Python, since that requires sudo access. Instead, you will need to use some kind of virtual environment.

VS Code can help you manage either Python virtual environments or conda environments. Python virtual environments will work without needing to install anything extra. I recommend first trying Python virtual environments.

Python virtual environments

I think VS Code is the easiest way to handle virtual environments. You can create a separate virtual environment for each project (workspace).

To create a virtual environment through the command line:

cd <project directory>
python3 -m venv .venv

To activate the virtual environemt:

source .venv/bin/activate

To deactivate the virtual environment:

deactivate

Conda

Conda is a bit more powerful than venv since it can install things other than Python modules, such as specific CUDA library versions, etc. If you need to use conda, do the following:

curl -LO http://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh
rm Miniconda3-latest-Linux-x86_64.sh

Then, after exiting out and reconnecting, run:

conda update -y conda

VS Code should automatically detect your conda environments so that you can select the appropriate Python kernel when running code.

My program stops running when I disconnect from SSH – how can I run a long process?

You should run your sessions inside of tmux to avoid SSH timeout and be able to return to sessions after disconnecting. Start a session with tmux, use ctl-b ctl-d to leave tmux and tmux attach to get back into it. You can finds lots other keyboard commands for tmux online.

How do I upload and download files to and from the server?

You can use scp or sftp from the command line. For a nicer experience, you need an SFTP client. Some free options are FileZilla and WinSCP.

I am getting CUDA out-of-memory errors, what do I do?

There are four GPUs on the machine but typically your code will only use the first one unless you tell it otherwise.

You can use

nvidia-smi

to check GPU availability. Then, to select a particular GPU to use:

export CUDA_VISIBLE_DEVICES=<0,1,2, or 3>

What if I can’t correctly set up my environment in conda? How do I compile C++ programs? Can you please install a specific program or CUDA version?

In cases where you need specific versions of drivers, libraries, etc. installed or you need to compile C++, you might need to use Docker so that you can manage the system environment yourself. You need to be added to the docker group in order to use docker – ask me about it and I will help you determine if Docker is actually necessary.

To use Docker, you need to create a Dockerfile and store it in a sub-directory. For example, create a text file at “docker/Dockerfile.” Here is an example Dockerfile for Tensorflow:

FROM tensorflow/tensorflow:latest-gpu-py3
RUN apt-get update ; apt-get install vim git wget -y
RUN pip install --upgrade scikit-image tqdm opencv-python

To build the Dockerfile:

docker build docker -t <tag>

You can choose whatever you want for the tag.

Docker normally runs as root which is dangerous. However, with this command you can run Docker as your user rather than root:

docker run -u $(id -u):$(id -g) --rm --gpus device=<gpu number> -it -v /code:/code -v /data:/data -v /data2:/data2 -e USER=$USER -e HOME=/code/$USER -w $PWD <tag> bash