LLNL logo Lawrence Livermore National Laboratory

Center for Applied Scientific Computing

A Brief History of CASC

The Center for Applied Scientific Computing (CASC) was founded in March 1996. At the time, the National Energy Research Supercomputer Center (NERSC) and the Center for Computational Sciences and Engineering (CCSE) had just moved from LLNL to Lawrence Berkeley National Laboratory. Though initially stunned by the move, nearly all NERSC and several CCSE computational and computer scientists chose to stay at LLNL.

CASC was conceived as a new home for applied mathematics and computer science research at LLNL. To ensure its future success, it was important that CASC have strong ties to LLNL and its programs. Therefore the “CASC model” became — a world-class, in-house research group with strong internal ties to LLNL programs to solidify its relevance and strong external ties with academia to ensure that CASC was always bringing in new ideas and personnel to perform cutting-edge research. CASC became a division of what was then the Scientific Computing and Communications Department within the Computation Directorate. It had 12 original members and was predicted to be a 20-person research organization.

Meanwhile, the Accelerated Strategic Computing Initiative (ASCI) program was just starting out, and CASC was there to fill the demand for positions. CASC was instrumental in many ASCI milestones, and its algorithms and software are still used in key ASC codes today. In its first five years, CASC was a fast-growing organization and focused mainly on enabling large-scale numerical simulations. Its early successes led to additional funding and support, allowing it to grow to almost 100 people after only five years.

Some of CASC's earliest projects: scalable linear solvers (hypre), adaptive mesh refinement (SAMRAI), and subsurface flow (PARFLOW) are still among its most successful. Both hypre and SAMRAI impacted many different Laboratory programs, and although CASC discontinued its participation in PARFLOW many years ago, the PARFLOW project is still providing valuable contributions. Over the last several years, CASC has stayed at roughly the same size, but its research scope has expanded to include data science research. Other successful projects include Overture (object-oriented tools for solving PDEs in complex geometries), Ardra (neutron and radiation transport), large-scale visualization, and Sapphire (scientific data mining).

CASC employees have helped LLNL research teams win several prestigious honors including Gordon Bell Awards for excellence in using high-performance computing to solve scientific and engineering problems. R&D Magazine, which hands out the annual R&D 100 awards to recognize the 100 most technically significant new products and innovations, has recognized CASC excellence numerous times.