The Data Vortex Network

The Data Vortex Network

Presenter

  • Coke Reed, Data Vortex Technologies, Oak Ridge
March 20, 2018 - 1:30pm to 2:30pm

Abstract 

Dr. Reed will describe the basic architectural features, scalability, and performance of the Data Vortex network. The high radix crossbar switches used in traditional crossbar networks tend to transmit only large packets in order to “hide” the time to compute the setting of the switch. Since there is no setting time for the self-routing Data Vortex switch, short packets can be used effectively. The Data Vortex may be used as a network connecting cores on a single chip or as a network connecting multiple computing elements in several cabinets. The Data Vortex excels in fine-grained, congestion-free data movement. For certain examples, including sparse linear algebra and Fourier analysis for solving differential equations, Dr. Reed will highlight how to calculate bandwidth and latency through a Data Vortex network of a given radix. The high bandwidth, low latency Data Vortex network excels at tackling large scientific problems as well as particularly interesting sets of graphs. The conversation will end with various uses and implementations of the network in present-day and future architectures.

Additional Information 

About the Speaker:

Dr. Coke Reed is the founder, inventor, and chairman of Data Vortex Technologies. Coke has authored over thirty patents on the Data Vortex and related technologies. He received all his mathematics degrees from The University of Texas at Austin and has previously worked for the Institute for Defense Analysis in Princeton, Bowie and La Jolla, Los Alamos National Laboratories, the Microelectronics and Computer Corporation and NASA. On November 15, 1990, Dr. Reed was awarded the Exceptionally Meritorious Civilian Service Award Medal for his contribution to national security.

Sponsoring Organization 

Computing and Computational Sciences Directorate,Computational Research and Development Programs

Location

  • Research Office Building
  • Building: 5700
  • Room: L-204

Contact Information

Share