Gordon: Data-Intensive Supercomputing

The SDSC Gordon Compute Cluster, a unique data-intensive supercomputer sponsored by the NSF XSEDE program, went into production January 1, 2012 Large graph problems, data mining, de novo genome assembly, database applications, and quantum chemistry are some of the fields of research benefitting from Gordon's unique architecture. The system is characterized by:

  • 1,024 dual-socket Intel Sandy Bridge nodes, each with 64 GB DDR3–1333 memory
  • Over 300 TB of high performance Intel flash memory SSDs via 64 dual-socket Intel Westmere I/O nodes
  • Large memory supernodes capable of presenting over 2 TB of cache coherent memory
  • Dual rail QDR InfiniBand network.
  • Data Oasis high performance parallel file system with over 4 PB capacity and sustained rates of 100 GB/s

New XSEDE Allocation Request Cycle Dates

The next Gordon request cycle runs from September 15, 2013 through October 15, 2013, for 12-month allocations beginning January 1, 2014. Please see the XSEDE User Portal for details on how to prepare and submit an allocation request.

Providing a new allocation model for data-intensive computing

In addition to the standard Research Allocation requests done through the XSEDE Allocation Request System, Gordon provides an additional new and unique allocation opportunity. Users can request dedicated, long-term use of Gordon's I/O nodes for data-intensive applications like databases and data mining environments. With 4.8 TB of high performance flash memory in each I/O node, Gordon I/O nodes provide exceptionally high performance and capacity for solving the most challenging data problems. For example, this may be of value where new instruments create data volumes exceeding the project environment capacity.

This special access to Gordon I/O nodes is done only through the startup allocation process, giving XSEDE users and staff a chance to evaluate the novel allocation model. Although startup awards require only a one-page proposal, applicants must provide strong justification of their need for dedicated access.

Gordon I/O nodes are considered a storage resource within the XSEDE project. Therefore, a node must be requested as 4.8 TB of flash storage.

Review the Gordon resource catalog entries for more information on these distinct offerings.

Getting help with your allocation request

Here are some suggestions to help you prepare a successful allocation request:

  1. Review talks from the Grand Challenges in Data-Intensive Discovery conference and the 2011 Gordon Summer Institute to understand the motivation for using Gordon.
  2. Review the process for submitting an allocation request.
  3. Review successful allocation requests.

Recent Gordon News & Events

Gordon Ranks No. 88

May 29, 2015
Still One of World's Fastest HPC Systems Gordon, the unique data-intensive flash-based HPC resource at the San Diego Supercomputer Center (SDSC) at t... more news...

Past Events

Quick Links