Supercomputer simulations of stucture formation are an emerging tool for understanding the physical properties that drive the formation and evolution of galaxies. The computer cluster built with this SCREMS grant will be used to run some of the world's largest Gadget-2 hydrodynamic simulations, incorporating our promising galactic outflow model that has enabled, for the first time, a fully dynamical investigation of the complex feedback cycle connecting galaxies and the intergalactic medium. These simulations will be compared to a wide suite of multiwavelength, multi-epoch observations of galaxies and intergalactic gas, in order to refine our models and identify missing physics. Additionally, this cluster will be used to run our advanced far-infrared radiative transfer code called Turtlebeach, which makes observable predictions for far-infrared emission from Gadget-2 simulations, for comparison to present observations and future data from Herschel and ALMA. The computing power provided by this grant will provide a new level of capabilities that will significantly improve the dynamic range of our models, likely leading to substantial advances in our understanding of galaxy formation. These science investigations will be complemented by a closely coupled program to expand training in supercomputing proficiency.
With ever-growing telescope power we can now detect galaxies back to when the universe was just a few percent of its current age. These images have revealed a Universe filled with galaxies coming in a wide range of sizes, shapes, and colors, and evolving through a myriad of physical processes including gravity, shocks, radiation, and chemical synthesis. Understanding how the galaxies we see came to be is a central challenge in astronomy that can now be addressed using powerful supercomputers. Supercomputer simulations evolve the universe from shortly after the Big Bang until the present day, and make predictions for the galaxy population that can be compared to forefront observations. Successful aspects of the model allow us to solidify our physical intuition, while areas that disagree with observations motivate improvements and refinements. Supercomputers are required to comprehensively model the sundry physical processes operating over a large range of scales in space and time. The supercomputer built using this SCREMS grant will enable some of the largest, most sophisticated simulations ever to be done, which are sure to yield new insights into the physical processes by which galaxies come to look the way they do. Furthermore, this supercomputer will be used to run our state-of-the-art tools for comparing these models to observations, particularly in the emerging arena of millimeter astronomy. Finally, this grant will enable the development of lesson units on supercomputing, to train the next generation of students to utilize this rapidly-advancing technology.
