Rationale

In 2001, the DoE Office of Science embarked on an ambitious new program called ``SciDAC - Scientific Discovery through Advanced Computing.'' One of the goals of SciDAC is to develop and deploy the computer science tools that will enable computational scientists with DoE mission-relevant applications to take full advantage of terascale computing platforms. While a large number of computer science-centered projects received funding in so-called ``Integrated Software Infrastructure Centers'' (ISICs), the approach to scientific visualization was not integrated across the program. Only a small number of computational projects include specific visualization tasks supported through SciDAC. It is clear that many DoE computational scientists, who use NERSC or other OASCR (Office of Advanced Scientific Computing Research) computational resources have visualization requirements that are currently not met.

Objective

We are organizing a one-day workshop dedicated to the identification of crucial scientific data visualization needs. The workshop will not be concerned with presenting state of the art in scientific data visualization. Instead, the workshop will consist of computational scientists and engineers who will present their visualization requirements for current and next generation projects. We will invite leading computational scientists and engineers from across the DoE Scientific Computing programs, with the purpose of having them present their visualization needs. Application scientists and engineers will be asked to provide one-page summaries of their data visualization requirements. We will collect these requirement summaries in a white paper and make them available to DoE for further planning of the visualization program in MICS (Mathematical, Information and Computational Sciences Division), in particular for potentially growing the program. The white paper should be seen as a requirements document similar to the "NERSC Greenbook" for computational resources.

Date and Location

We propose to hold the one-day workshop adjacent to the next NERSC User Group (NUG) meeting. Many key NERSC users will be attending the NUG meeting, and we would invite them to stay for an extra day to participate in the one-day workshop. The NUG meeting will take place in 3-4 June 2002, at both the OSF in Oakland, and at LBNL in Berkeley.

List of Invitees

We plan to invite application scientists and engineers and scientific data visualization experts to participate in the one-day workshop. Potential DoE computational scientists and engineers to speak at the workshop include:

 Kwok Ko, SLAC-HENP (accelerator)
 Rob Ryne, LBNL-HENP (accelerator)

 Peter Nugent, LBNL (astrophysics)
 Ed Siedel, AEI (astrophysics)
 Tony Mezzacapa, ORNL-HENP (astrophysics)

 Bill Tang, PPPL-OFE (fusion)

 Mathew Maltrud, LANL (climate modeling)
 Warren Washington, NCAR (or John Drake, ORNL)

 Phil Colella, LBNL-OASCR (combustion)

 Daniel Rokhsar, LBNL (Bioinformatics)

 ??? (environmental mod1eling ???)
 ??? (nano science)
 ??? (chemistry)
 ??? (computational neuroscience)
 

Potential DoE and university- and industry-based visualization researchers to participate in the workshop include:

Jim Ahrens, Los Alamos Nat'l Lab.
Tom DeFanti, University of Illinois at Chicago
Mark Duchaineau, Lawrence Livermore Nat'l Lab.
David Ebert, Purdue University
Herbert Edelsbrunner, Duke University
Pat Hanrahan, Stanford
Chuck Hansen, University of Utah
Chris Johnson, University of Utah
Kenneth I. Joy, University of California-Davis
Arie E. Kaufman, SUNY Stony Brook
Kwan-Liu Ma, University of California-Davis
Raghu Machiraju, Ohio State University
Nelson L. Max, Lawrence Livermore Nat'l Lab.
Robert J. Moorhead, Mississippi State University
Gregory M. Nielson, Arizona State University
Valerio Pascucci, Lawrence Livermore Nat'l Lab.
Jarek Rossignac, Georgia Tech
Will Schroeder, Kitware
Oliver G. Staadt, University of California-Davis
Rick Stevens, Argonne Nat'l Lab.
Samuel P. Uselton, Lawrence Livermore Nat'l Lab.
Amitabh Varshney, University of Maryland at College Park
John Shalf, Lawrence Berkeley Laboratory
 

There will be opportunities during the workshop for application scientists and engineers and visualization researchers to discuss jointly the visualization requirements. Discussion will be directed clearly at the refinement of visualization needs. The aim will be to ensure that the final visualization requirements lists will be comprised of issues that truly require innovative visualization approaches.