This page will help you get started using the Farber cluster, UD's second Community Cluster. For general information and specifications about the Farber cluster, visit the IT Research Computing website. To cite the Farber cluster for grants, proposals and publications, use these HPC templates.

You need a Farber account to access the login node. For example, using SSH to connect to the Farber login node with account traine, type:

    ssh traine@farber.hpc.udel.edu

Your account is configured as a member of an investing-entity group name (workgroup) which determines access to your group's compute nodes, queues and storage resources on Farber. Setting your workgroup environment is required in order to submit jobs to the Farber cluster. For example, the traine account is a member of the it_css workgroup. To start a shell in the it_css workgroup, type:

    workgroup -g it_css

You will need to know Unix/Linux before using Farber. The getting started guide will provide the basics about using IT-supported HPC systems, from accessing the cluster to running applications. The Farber HPC Basics presentation is an overview about Farber focusing on the differences from a standard Linux system.

farber.hpc.udel.edu has live resources: system status, job stats, system alerts.

UD IT HPC has Farber machine information: attributes including a database of node information, milestones, offline nodes and nodes disabled for maintenance.

Cluster monitoring for Farber uses Ganglia to monitor its hardware components.

System alerts: Check here first if you are experiencing problems with the cluster.

Node status notification for Farber is an opt-in service notifying you about the status changes on any of your workgroup's nodes.

Job statistics: Check here for the total number of jobs that ended on each day over a range (week, 2 weeks, month, 6 months, year) with an overlay of the total number of jobs which the job scheduler classified as "failed."

• IT-managed software: A list of installed software that IT builds and maintains for Farber users.
• IT's udbuild process for building and installing software on the Farber cluster.

<box round right width 18em|Farber community cluster overview [available formats]> [PNG] [SVG] </box>

• Using Farber's filesystems

For general information on file transfer see Transferring files

For general information see Computing environment

There are two 64-bit compiler suites on Farber with Fortran, C and C++:

In addition, IT has installed OpenJDK (Open Java Development Kit), which can only be used on the compute nodes.

Farber is based on Intel (code name "Ivy Bridge"): See Intel 64 and IA-32 Architectures Optimization Reference Manual and Intel Math Kernel Library Cookbook.

For general information, see Application development

General information about running applications is applicable to all software. Additional help for specific software may be found on the general software page.

Review the nVidia'a GPU-Accelerated Applications list for applications optimized to work with GPUs. These applications would be able to take advantage of nodes equipped with nVidia Tesla K20X coprocessors.

Opt-in to Farber node status notification service to receive one email if any of your workgroup's nodes transition from being offline, online or accepting-jobs.

If you are experiencing a system related problem, first check Farber cluster monitoring and system alerts. To report a new problem, or you just can't find the help you need on this wiki, then send an e-mail to consult@udel.edu with subject=“Farber Help” and make the first line of the message body be

    Type=Cluster

or submit a Research Computing Help Request specifying High-Performance Computing and Farber cluster for the problem details. .

hpc-ask is a Google group established to stimulate interactions within UD’s broader HPC community and is based on members helping members. This is a great venue to post a question, start a discussion or share an upcoming event with the community. Anyone may request membership. Messages are sent as a daily summary to all group members. This list is archived, public, and searchable by anyone.