Actions to Meet the Science-Driven Needs for Science Computing at the Space Telescope Science Institute

摘要

By the next five years, our staff will be involved in a number of new projects and new astronomical missions that will require very demanding computing architectures both in terms of hardware and software (e.g. LSST, ALMA, JWST). Based on our evaluation, the most promising path ahead to address and satisfy these science-driven needs for our science staff relies on High Performance Computing (HPC), using Central Processing Unit (CPU) clusters or Graphics Processing Units (GPUs), depending on the application at hand. Storage and collaborative tools are two other top areas in ever greater demand and will continue to represent two top areas for IT to focus on.