Volume 7, Issue 1 -
Spring/Summer 1999

Volume 6, Issue 3
Fall 1998

Volume 6, Issue 2
Spring/Summer 1998

Volume 6, Issue 1
Winter 1998

Volume 5, Issue 4
Fall 1997

Volume 5, Issue 3
Summer 1997

Volume 5, Issue 2
Spring 1997

Volume 5, Issue 1
Winter 1997

Volume 4, Issue 4
Fall 1996

Volume 4, Issue 3
Summer 1996

Volume 4, Issue 2
Spring 1996

Volume 4, Issue 1
Winter 1996

Volume 3, Issue 4
Fall 1995

Volume 3, Issue 3
Summer 1995

Volume 3, Issue 2
Spring 1995

Volume 3, Issue 1
January 1995

Volume 2, Issue 4
October 1994

Volume 2, Issue 3
July 1994

Volume 2, Issue 2
April 1994

Volume 2, Issue 1
January 1994

Volume 1, Issue 4
October 1993

Volume 1, Issue 3
July 1993

Volume 1, Issue 2
April 1993

Volume 1, Issue 1
January 1993

TWO CULTURES: ACADEMIC PARTICIPATION IN THE DoD HPC MODERNIZATION PROGRAM


The Department of Defense (DoD) has established the High Performance Computing Modernization Program (HPCMP) to establish a world-class computing infrastructure to support the defense applications community. The HPC Modernization Office reports to Dr. Anita Jones, Director of Defense Research and Engineering and a former head of the CRPC External Advisory Committee. Recently the HPCMP has awarded four contracts for Major Shared Resource Centers (MSRCs) at

  • The Army Corps of Engineers Waterways Experiment Station, Mississippi The Aeronautical Systems Center, Wright Patterson Air Force Base, Ohio The Naval Oceanographic Office, Stennis Space Center, Mississippi
  • The Army Research Laboratory, Maryland

Among them, these four sites will have an impressive array of supercomputing equipment of a variety of architectures, including vector multiprocessors, scalable parallel systems, and networks of high-end scientific workstations. When interconnected via the high-bandwidth Defense Research and Engineering Network (DREN), these centers will offer some of the world's most powerful computation facilities to military computer users at installations all over the United States.

Of course, powerful computer hardware is not, by itself, enough. To be successful, these centers will need the best emerging software for use of high-performance computers. Recognizing this, the Defense Department has allocated fully a quarter of the annual budget for Major Shared Resource Centers for Programming Environment Training (PET). This program includes plans for DoD users to attain greater expertise as they are introduced to leading-edge tools and techniques for parallel computing.

Jones has stated on many occasions that she wants to get the academic research community deeply involved in this program to ensure that it remains at the forefront of HPC developments. Hearing her message, the CRPC and many other academic institutions and researchers participated as team members in MSRC proposals. The CRPC is part of the winning PET teams at three of the four MSRCs. We also expect to collaborate with the university team at the remaining center. We welcome the opportunity and challenge to bring new ideas to a vital user community.

The opportunity offered by involvement in the PET program is obvious. Many significant computations are done daily at the MSRCs, ranging from surface and groundwater flows to radar cross-sections. The DoD community needs truly high-performance computing to solve these problems. Key technologies pioneered by CRPC researchers and other academics, such as domain decomposition solvers and out-of-core dense linear algebra routines, will be vital to providing these capabilities.

The challenge is to bridge the very different cultures of DoD and academia to fulfill the HPC Modernization program opportunity. The military's goals focus on the end product: getting accurate simulations done before the results are outdated. Contrast this with academic researchers, who are more interested in experimenting with new techniques. From slightly different viewpoints, DoD users would like PET to provide robust production tools to solve physical problems, while the researchers are more comfortable producing experimental programs and measuring the behavior of machines. For the program to succeed we need to find a middle ground between using academic computer scientists as support programmers and using military application developers as rudimentary testbeds. Resolving this conflict will require creative thinking on both sides, possibly along with the involvement of commercial firms.

In spite of these challenges, it is essential that both sides persist in their determination to work together as partners. We must bridge the culture gap by understanding each other's goals and requirements. To succeed, DoD Modernization will need to mobilize the best academic computer scientists to think about and help solve defense problems. Computer and computational scientists need to see application developers as collaborators rather than consumers and work with them to build a better software infrastructure for the Defense Department and the nation. Only if this spirit of collaboration is achieved can the HPCMP achieve the lofty vision that Jones has established for it.


Table of Contents