Petaflops Architecture WorkShop

April 22 through 25, 1996
Oxnard, California

Sponsors:
Advanced Research Projects Agency
Department Of Energy
National Aeronautics and Space Administration
National Security Agency
National Science Foundation

PAWS '96 was a federally sponsored workshop which identified opportunities and directions for future research in very high performance computing systems structures and technologies. The design of this, and other petaflops related workshops, is the achievement of practical petaflops scale computing in the nearest possible time-frame. The Proceedings from this Workshop are not yet available.

PAWS'96 Proceedings Production Schedule

Goals

Supporting Materials and Information

Organizers

Who Participates ?

Go Back to PETA
Please contact Larry Picha if you have any questions or comments about PAWS or this WWW page.
Design/Author: Larry Picha
Authorizing Official: Thomas Sterling, Program Chair
Last Revised: 01 MAY 96 (ldp)

GOAL
The purpose of PAWS'96 is to set initial directions for early inquiry into potential petaflops scale architecture and to identify key issues and objectives that must be addressed by future advanced computing research programs towards the goal of achieving effective petaflops performance. The focus of PAWS'96 will be approximately ten proposed systems approaches which will be considered from the perspectives of architecture, applications, and system software.

APPROACH
The four day workshop will consist of a series of meetings to conduct a detailed examination and evaluation of a set of proposed architecture concepts for near petaflops performance. The proposed architectures will be derived from those selected as part of the NSF solicitation for point-design studies of very high performance systems. A small group of experts in key aspects of high performance system structure, operation, and use will be convened to expose the potential and implications of the proposed architecture concepts. The workshop will include the following activities:

Conduct architecture design reviews of selected candidate architectures from the NSF sponsored point-design studies and determine their respective execution models.

Determine the programming model of each candidate architecture and examine them from the perspective of their potential for executing a specified set of applications.

Evaluate the ways in which the proposed architectures will limit the burden on system software and determine the requirements of each architecture for support from system software and tools.

Document the findings, open issues, and recommendations of the workshop in time to provide a driver for the following Petaflops System Software Summer Study to be conducted in June, 1996. Release preliminary conclusions of the workshop to the general community on the "world wide web".

PARTICIPANTS
Participation will be by invitation. These will include Investigators of the NSF point-design studies, experts in the broad field of high performance computing, and representatives of the sponsoring agencies. Approximately 50 participants are anticipated at PAWS'96 including facilitating staff. Non-government employees attending PAWS'96 will be reimbursed for travel expenses including an initial registration fee. NSF Investigators costs will be covered as part of their grants.

ORGANIZERS

Workshop Chair:
Thomas Sterling, NASA Goddard Space Flight Center

Co-Program Chairs:

Geoffrey Fox, Syracuse University
Peter Kogge, University of Notre Dame

Program Committee:
Ian Foster, Argonne National Laboratory
Fred Johnson, NIST
Paul Messina, California Institute of Technology
Rick Stevens, Argonne National Laboratory

Steering Committee Chair:
Paul Smith, HPCC NCO and NASA

Steering Committee:
George Cotter, NSA
Bill Feiereisen, NASA Ames Research Center
Bob Lucas, ARPA
John Van Rosendale, NSF
Gil Wiegand, DOE

Please contact Larry Picha if you have any questions or comments about PAWS or this WWW page.