Data Release Production

Princeton HPC Systems

In addition to the regular LSST-provided compute systems (lsst-dev, the Verification Cluster, etc), DRP team members have access to two clusters hosted by the Research Computing Group in Princeton. Please refer to the Research Computing Group’s pages for general information on getting started, usage policies, FAQs, etc, and be aware that you must comply with all their rules when using these systems

Obtaining Accounts

Accounts are issued on demand at the request of an appropriate PI. For our group, that means you should speak to either Robert or John, and they will arrange one for you. When your account has been created, you should check that you are a member of the groups astro, hsc, and lsst (use the groups command).

Available Systems

Tiger

Tiger is a Dell/SGI cluster, providing around 10,000 CPU cores, with at least 4 GB of RAM per core, spread over 644 compute nodes. When connecting to Tiger, you should connect to the head node tiger-sumire.princeton.edu (also known as tiger3.princeton.edu) which has been reserved for HSC/LSST use.

Tiger runs version 6.8 of Springdale Linux (a derivative of RHEL). As such, the default toolchain is too old to work with the LSST stack. You should therefore enable devtoolset-3 before proceeding:

scl enable devtoolset-3 bash

A regularly-updated LSST “shared stack” is available in /tigress/HSC/LSST/stack_tiger/ (note that the name of the cluster is embedded in the path). To get started, try:

. /tigress/HSC/LSST/stack_tiger/loadLSST.bash
setup lsst_apps

Perseus

Perseus is a Dell Beowulf cluster, providing 8,960 CPU cores, with 4.5 GB of RAM per core, spread over 320 compute nodes. It provides broadly equivalent capabilities to Tiger, but is often less heavily loaded. Unlike Tiger, there is no head node reserved for HSC/LSST use. Connect to perseus.princeton.edu, and be especially considerate of other users before starting long-running jobs on the head node.

Tiger runs version 7.3 of Springdale Linux (a derivative of RHEL). This provides a default toolchain which is appropriate for working with the LSST stack. You should not enable devtoolset-3 (or any other devtoolset) when working with LSST code on this system.

A regularly-updated LSST “shared stack” is available in /tigress/HSC/LSST/stack_perseus/ (note that the name of the cluster is embedded in the path). To get started, try:

. /tigress/HSC/LSST/stack_perseus/loadLSST.bash
setup lsst_apps

Storage

HSC data (both public data releases and private data, which may not be shared outside the collaboration) is available in /tigress/HSC on both clusters. This filesystem is available from both clusters, and you may use it to store your results. However, note that space is at a premium, especially during our periodic HSC data release processing: please clean up any data you are not actively using.

Space is also available in your home directory, but note that it is not shared across clusters.

Cluster Usage

Jobs are managed on both systems using SLURM; refer to its documentation for details.

It is occasionally useful to be able to bring up an interactive shell on a compute node. The following should work:

salloc --nodes 1 --ntasks 16 --time=1:00:00  # hh:mm:ss

Connecting from Outside Princeton

Access to all of the Princeton clusters is only available from within the Princeton network. If you are connecting from the outside, you will need to bounce through another host on campus first. Options include:

If you choose the first option, you may find the ProxyCommand option to SSH helpful. For example, adding the follwing to ~/.ssh/config will automatically route your connection to the right place when you run ssh tiger:

Host tiger
    Hostname tiger3.princeton.edu
    ProxyCommand ssh hammer5.astro.princeton.edu nc %h %p

See also the Peyton Hall tips on using SSH.

Help & Support

Contact cses@princeton.edu for technical support when using these systems. Note that neither the regular Peyton Hall sysadmins (help@astro) nor the LSST Project can provide help.