Overview

O2 is a platform for Linux-based high performance computing at Harvard Medical School. The name is derived from being the next generation of the HMS "Orchestra" cluster, hence "O"2.

O2 is managed by the Research Computing Group, part of HMS IT.
O2 is an HPC cluster built on Linux and the Slurm open source job scheduler.
Please submit a request for O2 help or feedback.
Follow us on Twitter for updates and alerts about service outages.

Getting Started with O2 and Slurm

Basic guide to O2 and the Slurm job scheduler ← New users, start here!
Switching workflows from Orchestra to O2 ← Orchestra users, start here!
How to login to O2
Frequently Asked Questions about O2
Security on O2

O2 Jobs and the Slurm scheduler

How to choose a partition (the Slurm equivalent to job queues in LSF)
Troubleshooting your O2 jobs
Parallel Jobs in O2
Using O2 GPU resources
O2 HPC Cluster and Computing Nodes Hardware Information
Job Priority
Examples of O2 commands
Get information about current and past jobs

Software and Programming on O2

Using research applications on O2
Using MATLAB on O2
Using Mathematica on O2
Using Conda on O2
Using Java on O2
Using Jupyter on O2
Maintain your own library of software for different languages
Installing custom software
Using Genome Browsers (IGV, UCSC Genome Browser)
Available Software on O2

General UNIX Information

Data Storage

There are a number of storage options available for research data.
Copying files to and from O2 (including downloading from websites)
If you had an Orchestra account, you will get access to the same home directory and shared network storage from O2.

Cluster Status

DEGRADED PERFORMANCE

Our cluster status page has details of service outages and planned maintenance.

September 18 - Service Degradation

O2 has been experiencing intermittent problems with its authentication system on login, transfer, and compute nodes. This issue can potentially result in: slow or failed logins to O2, Missing group membership, Failed job submissions. We are working with the software vendor to resolve this.

August 9 - Service Degradation

New jobs are intermittently not starting on the cluster (or the sbatch command has errors) due to an issue with cluster-storage communication. We believe that currently running jobs are still executing normally. Disk read/writes may be slower than usual, which can cause other commands to be slow. We will provide details as we get them.

July 8: notes after the July OS/Slurm update

Jupyter Notebooks users should start a new environment and remove any old runtime directories.
"sbatch" no longer uses the "–x11" option in the new version of Slurm. Just remove it from your script and X forwarding should work by default.
- "srun" commands still require "–x11" to enable X forwarding, though.
If you have any custom built software, you may need to recompile or relink it on O2.

Two Factor Authentication:

All O2 cluster logins from outside of the HMS network require two-factor authentication. Please see:

Community Events

User Training

Classes are offered each semester to the HMS community to help you ramp up your research skills!
Please check the User Training page for courses, dates, and registration

Office Hours

RC's Office Hours are held every Wednesday, 1 - 3 PM.
- As part of the HMS response to COVID-19, office hours are being held online, using Zoom.
- During these hours only, you can join the Zoom meeting room for a consultation.
Please contact us first with a support request before joining office hours so we can better help you!

Intermediate Slurm

O2