Access methods

There are a variety of ways you can access the SCF computers.

• Using SSH: The most basic access to SCF computers is via SSH to get to a command line on the machine of your choosing.
  • Example: to SSH to the SCF Linux server named arwen, you'll need to connect to arwen.berkeley.edu. From there you can see your home directory, connect to other SCF machines without using a password, and start jobs on the SCF Linux cluster. 
• Using your web browser: You can use our JupyterHub to access
  • Jupyter notebooks (IPython, R, and Julia notebooks as well as others)
  • Terminal sessions (similar to SSH command-line sessions, but from within your browser)
  • RStudio 
  • VS (Visual Studio) Code sessions
  • Linux desktop sessions
• Using Remote Desktop to get a graphical Linux desktop
• Running VS (Visual Studio) Code, an integrated development environment (IDE) on your personal machine, with VS Code connecting to an SCF machine to do remote development.
  • You'll need to install the 'Remote Development' extension pack and then click on the green button labelled '><' in the lower left corner.
  • Under this mode of using VS Code, you cannot access the cluster nodes (nor the GPUs on some of those nodes). To access the cluster nodes (including GPUs), please use VS Code through JupyterHub.
• You can copy files to and from the SCF filesystem in a variety of ways

What machines can I use?

You can browse the list of the SCF computers on our Grafana dashboards (SCF login required), including a general overview of most machines. Once you are on an SCF machine (e.g., you can initially SSH to arwen.berkeley.edu), run `sitehosts compute` to list machines you can remotely connect to and run jobs on.

Running graphical programs from an SCF machine

You can remotely run GUI-based programs on the SCF machines, displaying the GUI on your local machine. This works for all graphical programs on our Linux machines and for X-windows based programs on our Macs but not for Cocoa-based programs on the Macs. These instructions will tell you how to view such programs on your own Windows or Mac machine.

Note that our JupyterHub provides more responsive access to RStudio and VS Code sessions, and Remote Desktop may be a better option for displaying GUIs.

Other ways to access the SCF filesystem

Copying files using tools such as SFTP, SCP, rsync and Globus can be tedious and requires you to have to deal with the possibility of different file versions on your personal computer and on the SCF filesystem.

• You can mount your SCF home directory as a directory on your personal machine (available for any of Windows, MacOS, or Linux).
• If you have a Mac or Linux desktop that you keep in your office in Evans, the SCF can maintain that computer for you as an SCF-administered machine. In this case, you will have direct access to your home directory, mounted via NFS, as for all SCF machines. Please email us if you are interested in this.