Professional

Current Position

Senior Principal Engineer, EC2 Networking, Amazon Web Services

Biographical Sketch

I am a Senior Principal Engineer with the Annapurna Labs group in Amazon Web Services.  I am interested in large scale networking challenges, particularly those in High Performance Computing.  Prior to joining Amazon, I was a founding member of the Open MPI project while at Indiana University, the Advanced Computing Laboratory at Los Alamos National Laboratory, and the Scalable System Software group at Sandia National Laboratories.  I have also worked on a scalable communication subsystem for discrete event simulators while at the Information Sciences Institute.

Education

  • Ph.D. Computer Science, Department of Computer Science, Indiana University, March 2009.
    Advisor: Andrew Lumsdaine
  • M.S. Computer Science, Department of Computer Science, Indiana University, August 2003.
    Advisor: Andrew Lumsdaine
  • B.S. Computer Science, University of Notre Dame, May 2001.

Industry Experience

Research Interests

  • High performance network design
  • Communication libraries for parallel programming
  • System software design for next-generation HPC platforms

Honors and Awards

  • Department of Energy High Performance Computer Science fellowship, 2001 – 2003

Thesis

  • Brian W. Barrett. One-Sided Communication for High Performance Computing Applications. Ph.D. Thesis, March 2009. [pdf] [pdf (one page)]

Refereed Papers

  • Alvin, Ken, Brian Barrett, Ron Brightwell, Sudip Dosanjh, Al Geist, Scott Hemmert, Mike Heroux, Doug Kothe, Richard Murphy, Jeff Nichols, Ron Oldfield, Arun Rodrigues, and Jeffrey Vetter. On the Path to Exascale. to appear in the International Journal of Distributed Systems and Technologies.
  • Torsten Hoefler, Greg Bronevetsky, Brian Barrett, Bronis de Supinski, and Andrew Lumsdaine. Efficient MPI Support for Advanced Hybrid Programming Models. In Proceedings, 17th European MPI Users’ Group Meeting, Stuttgart, Germany, September 2010. [pdf]
  • K. Scott Hemmert, Brian W. Barrett, and Keith Underwood. Using Triggered Operations to Offload Collective Communication Operations. In Proceedings, 17th European MPI Users’ Group Meeting, Stuttgart, Germany, September 2010. [pdf]
  • Brian W. Barrett and K. Scott Hemmert. An Application Based MPI Message Throughput Benchmark. In Proceedings of IEEE Cluster 2009, August 2009. [pdf]
  • Brian W. Barrett, Jonathan W. Berry, Richard C. Murphy, and Kyle B. Wheeler. Implementing a Portable Multi-threaded Graph Library: the MTGL on Qthreads. In Proceedings of the 23rd IEEE International Parallel and Distributed Processing Symposium (IPDPS 2009), Workshop on Multithreaded Architectures and Applications, 2009.
  • Ralph Castain, Tim Woodall, David Daniel, Jeff Squyres, and Brian W. Barrett. The Open Run-Time Environment (OpenRTE): A Transparent Multi-Cluster Environment for High-Performance Computing. In Future Generation Computer Systems, Vol. 24, No. 2, February 2008.
  • Brian W. Barrett, Galen M. Shipman, and Andrew Lumsdaine. Analysis of Implementation Options for MPI-2 One-sided. In Proceedings, 14th European PVM/MPI Users’ Group Meeting, Paris, France, September 2007. [pdf]
  • Richard L. Graham, Ron Brightwell, Brian W. Barrett, George Bosilca, and Jelena Pjesivac-Grbovic. An Evaluation of Open MPI’s Matching Transport Layer on the Cray XT. In  Proceedings, 14th European PVM/MPI Users’ Group Meeting, Paris, France, September 2007. [pdf]
  • Galen M. Shipman, Ron Brightwell, Brian W. Barrett, Jeffrey M. Squyres, and Gil Bloch. Investigations on InfiniBand: Efficient Network Buffer Utilization at Scale. In  Proceedings, 14th European PVM/MPI Users’ Group Meeting, Paris, France, September 2007. [pdf]
  • Richard L. Graham, Brian W. Barrett, Galen M. Shipman, Timothy S. Woodall and George Bosilca. Open MPI: A High Performance, Flexible Implementation of MPI Point-to-Point Communications. In Parallel Processing Letters, Vol. 17, No. 1, March 2007. [preprint pdf]
  • Christopher Gottbrath, Brian Barrett, Bill Gropp, Ewing Rusty Lusk, and Jeff Squyres. An Interface to Support the Identification of Dynamic MPI 2 Processes for Scalable Parallel Debugging. In Proceedings, 13th European PVM/MPI Users’ Group Meeting, Bonn, Germany, September 2006. [pdf]
  • Richard L. Graham, Galen M. Shipman, Brian W. Barrett, Ralph H. Castain, and George Bosilca. Open MPI: A High Performance, Heterogeneous MPI. In Proceedings, Fifth International Workshop on Algorithms, Models and Tools for Parallel Computing on Heterogeneous Networks, Barcelona, Spain, September 2006. [pdf]
  • Brian W. Barrett, Ron Brightwell, Jeffrey M. Squyres, and Andrew Lumsdaine. Implementation of Open MPI on the XT3. Cray Users Group 2006, Lagano, Switzerland, May 2006. [pdf]
  • Sriram Sankaran, Jeffrey M. Squyres, Brian Barrett, Andrew Lumsdaine, Jason Duell, Paul Hargrove, and Eric Roman. The LAM/MPI Checkpoint/Restart Framework: System-Initiated Checkpointing. International Journal of High Performance Computing Applications, 19(4):479–493, Winter 2005. [preprint pdf]
  • B. Barrett, J. M. Squyres, A. Lumsdaine, R. L. Graham, and G. Bosilca. Analysis of the Component Architecture Overhead in Open MPI. In Proceedings, 12th European PVM/MPI Users’ Group Meeting, Sorrento, Italy, September 2005. [pdf]
  • R. H. Castain, T. S. Woodall, D. J. Daniel, J. M. Squyres, B. Barrett, and G. E. Fagg. The Open Run-Time Environment (OpenRTE): A Transparent Multi-Cluster Environment for High-Performance Computing. In Proceedings, 12th European PVM/MPI Users’ Group Meeting, Sorrento, Italy, September 2005. [pdf]
  • Brian Barrett and Thomas Gottschalk. Advanced Message Routing for Scalable Distributed Simulations. In Proceedings, Interservice/Industry Training, Simulation, and Education Conference (I/ITSEC), Orlando, FL 2004. [preprint pdf]
  • Edgar Gabriel, Graham E. Fagg, George Bosilca, Thara Angskun, Jack J. Dongarra, Jeffrey M. Squyres, Vishal Sahay, Prabhanjan Kambadur, Brian Barrett, Andrew Lumsdaine, Ralph H. Castain, David J. Daniel, Richard L. Graham, and Timothy S. Woodall. Open MPI: Goals, Concept, and Design of a Next Generation MPI Implementation. In Proceedings, 11th European PVM/MPI Users’ Group Meeting, Budapest, Hungary, September 2004. [pdf]
  • T.S. Woodall, R.L. Graham, R.H. Castain, D.J. Daniel, M.W. Sukalski, G.E. Fagg, E. Gabriel, G. Bosilca, T. Angskun, J.J. Dongarra, J.M. Squyres, V. Sahay, P. Kambadur, B. Barrett, and A. Lumsdaine. Open MPI’s TEG Point-to-Point Communications Methodology: Comparison to Existing Implementations. In Proceedings, 11th European PVM/MPI Users’ Group Meeting, Budapest, Hungary, September 2004. [pdf]
  • T.S. Woodall, R.L. Graham, R.H. Castain, D.J. Daniel, M.W. Sukalski, G.E. Fagg, E. Gabriel, G. Bosilca, T. Angskun, J.J. Dongarra, J.M. Squyres, V. Sahay, P. Kambadur, B. Barrett, and A. Lumsdaine. TEG: A High-Performance, Scalable, Multi-Network Point-to-Point Communications Methodology. In Proceedings, 11th European PVM/MPI Users’ Group Meeting, Budapest, Hungary, September 2004. [pdf]
  • Brian Barrett, Jeff Squyres, and Andrew Lumsdaine. Integration of the LAM/MPI environment and the PBS scheduling system. In Proceedings, 17th Annual International Symposium on High Performance Computing Systems and Applications, Quebec, Canada, May 2003. [preprint pdf]
  • John Mugler, Thomas Naughton, Stephen L. Scott, Brian Barrett, Andrew Lumsdaine, Jeffrey M. Squyres, Benoit des Ligneris, Francis Giraldeau, and Chokchai Leangsuksun. OSCAR Clusters. In Proceedings of the Ottawa Linux Symposium (OLS’03), Ottawa, Canada, July 23-26, 2003. [pdf]
  • Sriram Sankaran, Jeffrey M. Squyres, Brian Barrett, Andrew Lumsdaine, Jason Duell, Paul Hargrove, and Eric Roman. The LAM/MPI Checkpoint/Restart Framework: System-Initiated Checkpointing. In LACSI Symposium, October 2003. [preprint pdf]

Invited Talks, Presentations, Technical Reports, and Unrefereed Articles

  • Jeffrey M. Squyres, Brian W. Barrett, George Bosilca, Richard L. Graham, Galen M. Shipman, and Timothy S. Woodall. Open MPI Developer’s Workshop, San Jose, CA, April 2006. [link]
  • Richard L. Graham, Brian W. Barrett, Galen M. Shipman, and Timothy S. Woodall. Open MPI: A High Performance, Flexible Implementation of MPI Point-To-Point Communications. In Proceedings, Clusters and Computational Grids for cientific Computing, Flat Rock, North Carolina, September 2006.
  • Brian W. Barrett, Jeffrey M. Squyres, and Andrew Lumsdaine. Implementation of Open MPI on Red Storm. Technical report LA-UR-05-8307, Los Alamos National Laboratory, Los Alamos, New Mexico, USA, October 2005. [pdf]
  • Brian W. Barrett. Return of the MPI Datatypes. ClusterWorld Magazine, MPI Mechanic Column, 2(6):34–36, June 2004. [link]
  • Thomas Naughton, Stephen L. Scott, Brian Barrett, Jeffrey M. Squyres, Andrew Lumsdaine, Yung-Chin Gang, and Victor Mashayekhi. Looking inside the OSCAR cluster toolkit. Technical report in PowerSolutions Magazine, chapter HPC Cluster Environment, Dell Computer Corporation, November 2002.