Yen Servers
Introduction to the Yen Servers
At the GSB, we have a collection of Ubuntu Linux servers (the Yen cluster) specifically for doing your research computing work. If you are a faculty member, PhD student, post-doc or research fellow, by default you should have access to these servers. They are administered by the Stanford Research Computing (SRC) and located in Stanford's data centers.
Danger
The Yen servers are NOT designed for teaching or high risk data.
Why Use the Yen Servers?
These servers offer you several advantages over using a laptop or desktop computers.
Better Hardware
Let's use the server yen2.stanford.edu as an example: this machine has 256 processing cores and about 1 TB of RAM. With yen2, you are able to complete memory- or CPU-intensive work that would overwhelm even the best personal laptop!
Long running jobs
Even when your laptop is capable of doing the job, you may still want to offload that work to the external server. The server can free up resources for your laptop to use for other tasks such as browsing websites, reading PDF files, working with spreadsheets, and so forth. If your laptop crashes, it's very convenient for your compute jobs to continue!
Storage
The project files and any large output should live in a dedicated project space (not in your home). The storage capacity is over 1 PB (petabyte).
Licensed software
Tools like Matlab and Stata are installed and licensed to use on the Yen servers.
How to Connect
Tip
New to using a research server? Learn about Getting Started
There are various ways to connect to the Yen servers.
- SSH in to
yen.stanford.edu - A terminal on JupyterLab
- RStudio or Jupyter Notebook on JupyterLab
When you SSH in to yen.stanford.edu, a load-balancer will assign you to one of the interactive Yen server: yen1, yen2, yen3, yen4 or yen5.
The Yen Slurm nodes can only be accessed using the Slurm scheduler.
Tip
Any work running on an interactive server (yen[1-5]) can only be started or stopped from that server.
Overview of the Yen Computing Infrastructure
Important
yen-gpu1 has 4 NVIDIA Ampere A30 GPUs, each with 24 GB of GPU RAM
yen-gpu[2-3] each has 4 NVIDIA Ampere A40 GPUs, each with 48 GB of GPU RAM