DIRECTOR'S PERSPECTIVE
As we look back over the two years since NERSC relocated to Lawrence Berkeley National
Laboratory, we feel a great sense of accomplishment. Some of the early successes are
easily forgotten, so it is worthwhile to remind ourselves that together we achieved more
than even the biggest supporters of the move envisioned. First, we accomplished the move
with no disruption of service. I know of no precedent for a working supercomputer center
the size of NERSC relocating, much less maintaining continuity. Second, while
maintaining our parallel vector systems, we brought two new massively parallel processing
systems into production, vastly increasing our computing capability. The evaluation and
acceptance of the Cray T3E-900 was particularly difficult and time-consuming, straining
our already stressed resources. These early difficulties were well worth
the effort -- today
NERSC's T3E-900 is the largest unclassified computing resource in the U.S., and with
innovations such as checkpoint restarting, we have demonstrated that a highly parallel
machine can perform at high production standards. Third, we began implementing our new
vision of what a high-performance computing center can be, providing computational
science as well as computer science expertise to our clients, and conducting computational
and computer science research to accelerate the pace of scientific discovery.
Many elements contributed to NERSC's early success, but it would not have been possible
without the highly competent and dedicated effort of the NERSC staff. This transition has
not been easy, but we have maintained our high level of client service and achieved
significant scientific results, as summarized in this report. I am grateful to our DOE Office
of Energy Research sponsors for their endorsement of our ambitious plans; to our clients,
in particular ERSUG and EXERSUG, for their continued support in difficult times, for
their willingness to embrace new technologies, and for their patience during occasional
glitches; and to everyone at Berkeley Lab who worked hard to make NERSC welcome at
its new home. My special thanks and congratulations, however, go to the NERSC staff for
their skill, dedication, and tireless efforts to make NERSC the best scientific computing
resource in the world.
But this is just the beginning. In an environment where technology changes every 18
months, NERSC cannot afford to stand still. We are beginning the crucial decision-making
process for our next major system procurement, evaluating rapidly changing technologies
and industry trends as well as architecture, hardware, and software issues. We are
discussing new computational science initiatives that will enhance our contributions to the
Energy Research scientific mission. We have ongoing efforts to strengthen our
organization and to ensure the proper balance and integration of production and research.
And we continue to build our computational and computer science expertise. These are
more difficult tasks than we faced during the startup phase. Success will come more slowly
and less visibly. We must continue working together in the spirit of our startup phase to
assure NERSC's continuing success. I am counting on the continued support of our clients
and stakeholders, as well as the dedication of the NERSC staff, to bring us closer to our
vision of NERSC as a worldwide leader in high-performance computing.
Horst D. Simon
Next Page
Back to Table of Contents