acd asp atd cgd esig hao mmm rap scd
NCAR FY97 Annual Scientific Report Scientific Computing Division

SCD Director's Message

FY97 was a particularly challenging year for SCD. In the spring of 1996, we completed a procurement that would have brought one of the world's most powerful supercomputers to NCAR: the NEC SX-4. That machine would have enabled U.S. atmospheric scientists to address problems that are currently intractable and that will remain intractable until comparable computing power is available. Also, NCAR would have been able to provide U.S. atmospheric scientists with computing capability that is comparable with what their international peers will have during the remainder of the decade.

In the summer of 1996, the Department of Commerce (DOC) launched an anti-dumping investigation centered on Japanese supercomputers. At that point, NCAR realized that the NEC SX-4 might never be available in the U.S., and thus adopted the following strategy at the beginning of FY97:

  1. Prepare to switch to highly parallel, Distributed Shared Memory (DSM) microprocessor systems, e.g. SGI Origin, HP Exemplar, etc.,

  2. Continue acquiring U.S.-manufactured Parallel Vector Processor (PVP) systems, and

  3. Seek partnerships with national and international organizations.

In keeping with 1), during FY97 NCAR acquired a 64-processor DSM for test and evaluation. In keeping with 2), during FY97 NCAR installed two PVPs -- a Cray C90 and a Cray J9se. Following the DOC's recent determination of dumping, NCAR will continue with this strategy.

In mid-FY97, the NCAR Director requested that each NCAR Division prepare a Strategic Plan that encompasses FY98-02. Accordingly, SCD assessed each component of its mission with respect to goals, the budgetary and technological environment in which we must operate, and strategies for achieving our goals in the anticipated environment.

Broadly, SCD's goal is to enable the best computational atmospheric and related research in the world by being a recognized leader in providing, applying, and advancing high performance computing technologies and expertise. The environment for the planning period includes rapidly changing technologies, a spectrum of requirements, and limited budgets. To achieve our goal, we envision a hierarchy of resources including computing, storage, networking, models, data, and analysis tools that are suitably distributed among SCD and the community it serves. Our strategies include close collaboration with scientists, manufacturers, and peer organizations to assess and apply new technologies; use of standards; and identification/elimination of legacy systems.

Table of Contents

FY97 Science Highlights

Educational Activities

 FY97 Publications

Community Service

Staff, Visitors, and Collaborators

NCAR | UCAR | NSF | NCAR FY96 ASR | NCAR FY97 ASR