Research Computing

The Pronghorn Supercomputer

The Research Computing Team at the University of Nevada, Reno is excited to announce the arrival of the Pronghorn Supercomputer. This new, $1.3 million, 310 TFLOPS high-performance cluster will support the University's initiative to reach R1 Carnegie Research Classification.

Named after the American antelope, the fastest mammal on the North American continent, Pronghorn is designed to handle a wide range of campus research computing tasks, including big data analysis, machine learning, artificial intelligence, bioinformatics, and robotics. 

Dell EMC and DDN Storage were selected by a cross-disciplinary technical committee to build Pronghorn. Their high-quality proposal best addressed the engineering requirements gathered from 49 University researchers and the Cyberinfrastructure Committee.

Initial project funding was provided through the State of Nevada Knowledge Fund (facilitated by the Governor's Office for Economic Development), a donation from University supporter Mick Hitchcock, the Research & Innovation division, the Office of Information Technology and faculty investors. Pronghorn's computing and storage capacities will be expanded to meet the demands of the University's research computing community, as funding allows.

Pronghorn will be located at Switch TAHOE RENO, The Citadel Campus, a world-class, Tier-4 Gold data center. Switch is assisting in the development of Pronghorn by contributing critical infrastructure support such as space, power and security for the initial installation and future expansions at no cost to the University.

Initial hardware installation is scheduled for completion in September 2017 with testing starting shortly thereafter. Faculty-investor and campus-wide availability is anticipated in November 2017 and January 2018 respectively.

The University's Nevada Center for Applied Research will coordinate industry access to Pronghorn, once the system is fully operational, to those who have a connection to the University.

Pronghorn Specifications

CPU Nodes
Processor Count 32
Processor Specs E5-2683 v4, 2.1GHz Intel Xeon
Memory 256GB
Networking 100Gb/s Intel Omni-Path
GPU Nodes
Processor Count 32
Processor Specs E5-2683 v4, 2.1GHz Intel Xeon
Memory 256GB
Networking 100Gb/s Intel Omni-Path
GPU Count 4
GPU Specs NVIDIA Tesla P100, 16GB Memory, NVLink
Storage
File System DDN GridScaler (IBM Spectrum Scale/GPFS)
Capacity 1PB (Petabyte, or 1000TB)
Throughput 9GB/s
Scalability Factor Very High
System
CPU Node Count 64
GPU Node Count 11
Processor Count 2400
GPU Count 44
Memory 19TB
Networking 100Gb/s Intel Omni-Path
Storage Capacity 1PB
Theoretical Performance 310TFLOPS

Support

The Office of Information Technology supports Pronghorn and research computing resources.  Support is provided by:

  • John Anderson, Senior HPC Systems Administrator
  • Artin Matousian, Senior HPC Applications Specialist
  • Sebastian Smith, Senior HPC Engineer

For more information about Pronghorn email hpc@unr.edu.