A&A has some of the most demanding computing problems in the physical sciences and uses roughly 20% of the world supercomputer time. A&A research has been the source of innovative new algorithms (e.g., the ``tree code'') and data mining techniques (e.g., the Sloan survey) and has developed an array of sophisticated techniques for the analysis of image data to derive physical quantities. Numerical solutions, such as n-body simulation and hydrodynamical modeling, play a special role in A&A research because we are unable to do controlled experiments in any other way. Moreover computational astrophysics is a field of ever growing importance to make quantitative comparison between theory and observation.
In support of computational A&A research we have installed a 20-node ``beowulf'' computer network and will acquire a large shared-memory parallel computer through the successful PSciNet initiative (§ 3.1). Computational astrophysics needs to be developed through a new course at the undergraduate level (§ 2.2.1), and in the graduate program.