Yale University, its School of Medicine, and the W.M. Keck Foundation Biotechnology Resource have joined together to form a new Center for High Performance Computation in Biology and Biomedicine that will meet the growing need to analyze and interpret the enormous amounts of data that result from several break-through technologies including microarray analysis of gene expression, single nucleotide polymorphism (SNP) genotyping, and high throughput mass spectrometric-based analysis of the human proteome. The instrumentation needed to support this new Center has been funded by a $1.6 million dollar National Institutes of Health Grant which is briefly described here. The Center is open to all Yale investigators and to as many non-Yale investigators as possible. If you would like to join the Center please send your name, department, institution, title and paragraph length description of your project to ![[MAILTO]](/wiki/modern/img/moin-email.png){kind=small}
kenneth.williams@yale.edu.
The Center provides a number of Computational Resources.
The Center's main computational resource is Bulldogi. Follow the link for more details.
There have been a number of recent changes related to the move of Bulldogc and the arrival of the new Bulldogi cluster.
ITS has a website with additional information about Yale clusters, including Bulldogi and Bulldogc ![[WWW]](/wiki/modern/img/moin-www.png){kind=small}
http://research.yale.edu/hpc
NEWS 8/13/09 The cluster is back up following the maintenance downtime. 8/4/09 REMINDER: We have a 2 day downtime on bulldogi and bulldogc scheduled for August 10-11. 7/10/09 Yale ITS Network Services will be performing maintenance on the central ITS PIX firewall on Sunday July 12th from 5:15am - 7:00am. Please expect disruptions to network connectivity to the cluster during this maintenance window. 7/10/09 ITS has requested that we schedule a 2 day downtime during August. The working dates are August 10-11, although that may need to change as the day approaches. The purpose of the downtime is to: - bring all compute nodes to a common, updated kernel - reboot the switch stack - resolve memory accounting issues - perform some other miscellaneous tasks. All of the nodes, including the head node, will be offline during that time, all jobs killed, and filesystems inaccessible. You should be aware that some programs may need to be recompiled in order to run on the new operating system (most do not). We advise you to test programs that you rely on by sshing into one of two compute nodes, c020 and c024, that we have set up as test nodes and trying your program. You do not need to allocate these nodes via PBS. However, please do not consume large amounts of CPU time on them, and run top to check the load before running your program. 6/9/09 1120 Bulldogi is operational at this time. It appears that existing jobs survived the problem, which was due to the PBS server running out of memory. 6/9/09 0920 Bulldogi is unresponsive. ITS is working on the problem. 6/8/09 1015 The PBS scheduler has been restarted. All jobs running at the time the scheduler failed appear to have continued fine. 6/8/09 0915 The PBS scheduler on Bulldogi crashed Sunday afternoon. At the moment, no jobs can be submitted. ITS is investigating the problem. 3/13/09 1640 Bulldogi is unresponsive. ITS has been notified. 2/16/09 1550 Bulldogi became unresponsive, and unfortunately had to be rebooted. We are investigating the cause. 12/1/08 1520 Bulldogi is back in service after replacement of a disk in the Nexan disk array serving /home1 and /home2. 11/29/08 1430 Bulldogi has had a failure of one of the storage units. At this time the cluster is not accessible via ssh. We have reported the problem to the facility and are waiting for further information and an estimate of the repair time.