Introduction

• The Elder Linux Cluster is a small distributed-memory multi-processor system. Each of the 12 "public" nodes of the cluster contains one dual-core 3-GHz Intel Xeon EMT processor with 4096KB of cache and 32 GB of RAM. The nodes are interconected with Gigabit Ethernet (60-110 Mbytes/sec bandwidth, 50-200 usecs latency). In the default queue, a maximum of 4 cpus may be used by a single user at at time, which can be spread over a maximum of 4 jobs. The maximum time permitted for a single job is 168 hours. The Elder Linux cluster is intended for large-memory jobs that cannot be run on the larger general-purpose Linux clusters. The Elder cluster is also the host for the "condo" program, by which individual research groups can purchase nodes. Each group is assigned a dedicated queue; other users are permitted to use the nodes but their jobs will be pre-empted as soon as a member of the purchaser's group submits a job to their queue. For details please see this document.

Useful Links

• Elder Cluster Status
  Ganglia Toolkit for displaying cluster load and status information.
• Elder Queue Status
  Displays jobs in the queue. Click on the link for a given job ID to get complete information about that job.
• PBS Script Generator
  Generates PBS scripts based on user input to a form.

PBS Documentation

• PBS Pro User's Guide