Princeton Institute for Computational Science and Engineering (PICSciE) Parallel Computing Workshop - Day 1

Does your computer have 8 cores? Do you spend much time waiting for your code to run, while 7 of those cores sit idle? If so, then this is the workshop for you! The instructors will spend three days teaching the basics of MPI and OpenMP, the two most common tools used to parallelize scientific/numerical codes. During the workshop, participants will also parallelize several single-processor codes, providing hands-on experience. By the end of the workshop, you will have the tools you need to parallelize your own code. During an optional fourth day, you can bring in any code(s) you are working on and start parallelizing them while the instructors are available to answer questions/provide guidance. This workshop is aimed at those who have had no experience with parallel programming, though some general programming experience is assumed. The sample codes are written in C, so we provide a primer/refresher for those whose C may be rusty. Instructors: Jonathan Sievers has worked in a variety of fields in cosmology, with a main emphasis on using HPC to study the cosmic microwave background and galaxy clusters. Much of the data he has used has come from some of the highest permanent telescopes in the world, located in the Chilean Andes at altitudes of over 5,000 meters. Dr. Sievers did his graduate work at Caltech and did a postdoc at the Canadian Institute for Theoretical Astrophysics before moving to Princeton University. Jonathan Dursi was part of the team that won a Gordon Bell Award at Supercomputing 2000 for work in massively parallel adaptive mesh computations of detonations in Type Ia supernovae. He has been a member of the US Department of Energy's Advanced Strategic Computing Initiative FLASH Project, and more recently a senior research associate at the Canadian Institute for Theoretical Astrophysics (CITA). He has used large-scale simulations to probe questions of combustion, gravitational instability, and magnetic fields in the contexts of supernovae, galaxy clusters, and protoplanetary disks. He is now an HPC Analyst at SciNet. Space is limited to 20 participants, so register today at the Training website, www.princeton.edu/training or contact Andrea Rubinstein at alrubins@princeton.edu .