Computational Astrophysics
Computation has become an essential tool in theoretical astrophysics, data analysis, and modeling, and Princeton is a world-leader in the development and application of numerical methods in astrophysics.
Researchers at Princeton use scientific computation to study an enormous range of physical processes. At the largest scales, N-body, hydrodynamic, and radiative transfer methods are used to study cosmological structure formation, galaxy formation, and reionization (Cen, Ostriker). This work has helped establish the modern theory of the Lyman alpha forest (Cen and Ostriker). At the smallest scales, particle-in-cell (PIC) methods are used to follow particle acceleration and microscopic instabilities associated with collisionless shocks in diffuse astrophysical plasmas (Spitkovsky). In between, a wide variety of numerical methods are used to understand core-collapse supernova explosions (Burrows), accretion onto compact objects (Burrows, Spitkovsky, Stone), gravitational fragmentation of molecular clouds and star formation (Stone), accretion disks (Stone), and the light scattering properties of interstellar dust grains (Draine), just to name a few.
Astrophysicists at Princeton do not merely run public domain codes, but rather they are leading efforts to develop, implement, and test new state-of-the-art algorithms in many areas. Important methods developed at Princeton include tree-mesh codes for collisionless dark matter (Bode & Ostriker), hydrodynamic and radiative transfer codes to study planet atmospheres (Burrows) and reionization (Cen), a variety of grid-based MHD and radiation hydrodynamic codes to study everything from star formation (Stone) to supernovae (Burrows), and PIC and gyro-kinetic codes to study plasma dynamics (Spitkovsky). Members of the department are collaborating in various projects led by F. Pretorius in the Princeton Physics department to develop codes to simulate dynamical spacetimes and black-hole mergers. In observational astronomy, Princeton is one of the leading centers for the development of the software analysis pipelines for SDSS, WMAP, ACT, and LSST (Lupton).
The department benefits from close ties with the Princeton Institute for Computational Science and Engineering (PICSciE), which houses one of the most powerful collections of high-performance computing systems at any university in the country. These systems are freely available for use by any on-campus researcher, and some of our graduate students have used several million cpu hours per year for their thesis work. Members of the department also have access to emerging petascale systems at DOE, NASA and NSF national supercomputing centers. Students at Princeton receive a solid education in numerical analysis and software engineering through courses offered in the department, and in colloboration with PICSciE. A certificate in scientific computation is offered through the graduate school.
