University of Rochester
Calendar :: A-Z :: Directory
Dandelion
Schools and Colleges About Us Admissions Alumni and Parents Athletics Graduate Studies Information Technology Libraries and Museums Medicine Research Student Life Working at the University

Center for Research Computing (CRC)

From CRC

(Redirected from Main Page)

Contents

[edit]

Recent System Updates

2009-11-11 6:00 pm - The BlueHive and Nova clusters have been merged and upgraded. See New_Cluster_Environment for details.

[edit]

Background

The University of Rochester, founded in 1850, is one of the nation's leading private universities. With just over 4,500 undergraduates, Rochester is one of the smallest and most collegiate in character among the top research universities.

In 2005, the University recognized the need for a shared research computing resource to support both domain-specific and interdisciplinary research activities. A faculty team, consisting of 17 researchers from across the University documented the need for a central research computing resource and the opportunities presented by this resource. This team released their report, Regaining the Frontier: Research Excellence and the University Center for Computational Arts, Science & Engineering, in April 2006. In addition to recommending the purchase of a large High-Performance Computing (HPC) cluster, this team’s report recommended the formation of a University center to provide research computing support, expertise, training, and collaboration activities for the entire university community.

In late 2007, the College of Arts, Sciences, and Engineering and the Medical Center jointly funded the Center for Research Computing (CRC). In mid 2008, Gregory R. Warnes, Ph.D. was hired as the first CRC Program Director, the initial computing hardware (dubbed the ‘BlueHive’ cluster) was purchased, and the CRC was officially launched.

Since its launch in mid 2008, the Center for Research Computing has expanded to include 5 computing clusters providing more than 24 TeraFlops of computational power and 200 TB of storage space. As of July 2009, the CRC supports more than 150 users, and has provided more than 5 Million hours of compute time to more than 400,000 compute jobs.

[edit]

Vision

The Center for Research Computing enables University of Rochester researchers to effectively utilize computation in their research activities by providing computational resources, expertise, and support.

The CRC will evolve to provide expertise and training to assist researchers in applying High Performance Computing (HPC) tools to their own research. CRC consultants can provide support for a wide range of activities, including: conceptual assistance in determining how computing can be applied to a specific research problem; support in determining what software or algorithms are applicable for a problem; assistance in porting and tuning applications; software design and implementation support; and assistance in managing software lifecycle. Further, in partnership with academic departments across the university, the CRC provides training courses enabling faculty, students, and staff to develop key research computing skills. Course topics include: introduction to HPC technologies; use of specific HPC applications; HPC software development, porting, and tuning; software lifecycle management; and data visualization techniques and technologies.

In addition to providing computational resources and expertise, the CRC also serves as a focal point for interdisciplinary collaboration among researchers. The efforts related to the CRC have already yielded a number of cross-discipline interactions that are developing into novel research efforts.

[edit]

Value

The current capabilities of the CRC target the needs of groups within the Medical Center and the College of Arts, Sciences, and Engineering that currently conduct computationally intensive research, and thus reflect the need for large integrated computational platforms. CRC systems are considerably larger and more powerful than those available to any individual department or research group.

[edit]

Users

Despite the recent establishment of the Center for Research Computing, the CRC has already demonstrated its value for researchers spanning a variety of fields. While these initial pilot users represent traditional users of computational resources, the CRC will evolve to support all areas of scholarship including those that are not traditionally associated with computational research, such as Art History, Linguistics, and Music Composition. To this end, CRC staff will engage in a variety outreach activities to enable research computing to benefit these areas of scholarship.

Current CRC users include:

  • Biochemistry (URMC)
  • Biochemistry and Biophysics (URMC)
  • Biology
  • Biomedical Engineering (URMC)
  • Biostatistics and Computational Biology (URMC)
  • Brain and Cognitive Sciences
  • Cardiology (URMC)
  • Chemistry
  • Chemical Engineering
  • Center for Visual Science and Earth and Planetary Science
  • Community and Preventive Medicine (URMC)
  • Computer Science
  • Economics
  • Electrical and Computer Engineering
  • Functional Genomics Center (URMC)
  • Institute of Optics
  • Mechanical Engineering
  • Medicine M&D (URMC)
  • Microbiology and Immunology (URMC)
  • Neural Development and Disease (URMC)
  • Neurobiology and Anatomy (URMC)
  • Physics and Astronomy
  • Political Science

The center hosts a monthly CRC User Day to highlight research projects that use CRC resources and to introduce scientific and technical topics to the University community. Workshops and consulting sessions are provided to users immediately following CRC User Days.

[edit]

Technology

The CRC offers University researchers state of the art computing technology and software including support for the use and implementation of these resources. CRC currently maintains two high-performance computing (HPC) clusters, BlueHive (currently in pre-production "pilot" phase) and Nova. Compute clusters:

  • BlueHive Cluster – 672 CPU Cores – 24 TB Storage – 1.3 TB RAM – approx 7 TFLOPS
    • Head node: two 3.0 GHz Dual-core Xeon CPUs (4 cores), 8GB RAM, 73 GB local storage
    • 84 compute nodes: two 3.0 GHz Quad-core Xeon CPUs (8 cores), 16 GB RAM, and 73 GB local storage
    • 2 Storage nodes: two 3.0 GHz Quad-core Xeon CPUs (8 cores), 8 GB RAM, and 73 GB local storage
    • 20 TB storage array
    • Private Gigabit Ethernet Network
    • Torque/Maui batch queuing system
  • Nova Cluster - 144 CPU cores - 144GB RAM – 2.1 TB storage - approx 1.5 TeraFLOPS
    • Head node: two 3.0 GHz Dual-core Xeon CPUs (4 cores), 8GB RAM, 80 GB local storage
    • 32 compute nodes:
    • 24 nodes: two 3.2 GHz Dual-core Xeon CPUs (4 cores), 4 GB RAM, and 40 GB local storage
    • 8 nodes: two 2.33GHz Dual-core Xeon (8 cores), 8 GB RAM, and 80 GB local storage
    • 2.1 TB storage array
    • Private Gigabit Ethernet Network
  • Shell Cluster - (formerly Nebula Cluster) 60 CPU cores - 120GB RAM - 100 GB storage
    • Head node: one 2,2 GHz Dual-core Xeon CPUs (2 cores), 4 GB RAM
    • 28 compute nodes
    • 100 GB storage array
    • Private Gigabit Ethernet Network
[edit]

More Information

[edit]

General Information 

 * Web Site: http://www.crc.rochester.edu
 * Email: ResearchComputing@rochester.edu
[edit]

Account Requests

Accounts must be requested by a faculty member using the online account request form at http://www.crc.rochester.edu/AccountRequest.

[edit]

Support Requests 

 * Email: ResearchComputing@rochester.edu 
 * Phone: University IT Help Desk  585-275-2000
[edit]

CRC Program Director: Gregory R. Warnes, Ph.D. 

 * Morey Hall, Suite 220
 * Tel: 585-273-2794
 * Fax: 585-276-2097
 * Email: gregory.warnes@rochester.edu
Retrieved from "http://www.rochester.edu/its/web/wiki/crc/index.php/Center_for_Research_Computing_%28CRC%29"