Computational Cosmology on the Petascale

We will promote 100x algorithmic improvements to accompany the hardware acceleration from 10 Teraflops to a Petaflop.  This will embody our past developments in adaptive time integration, force evaluation and a unique renormalization approach to the evolution of structure. Unfortunately, higher dynamic range makes it increasingly difficult to achieve the scaling required for ambitious problems.  We must build an adaptive hydro code that scales as well as our adaptive gravity code by integrating tree and grid methods. We will simulate large cosmological volumes with sufficient resolution to capture the evolution of the smallest galaxies through to the largest clusters.  We will study the formation of galactic dark halos seeing lumps, streams and interactions between these and the baryonic components, critical for all proposed dark matter detection schemes. We will model the formation of present day galaxies, from spirals to ellipticals, studying the role of their large scale environment and focusing our effort on the proper modeling of small scale physics, namely molecular cloud and star formation.  We will follow the formation of the first stars through the formation of globular clusters – the oldest remnants of star formation prior to the reionization of the Universe.  We will develop the analysis and visualization software needed to maximize that impact in collaboration with a few international groups and shared within our extensive collaboration network.

• George Lake, University of Zurich
• Prof. Ben Moore, University of Zurich
• Dr. Joachim Stadel, University of Zurich
• Prof. Romain Teyssier, University of Zurich

• Dr. Jonathan Coles, University of Zurich
  
• Christine Moran, University of Zurich
• Sarah Nickerson, University of Zurich
• Dr. Marcus Wetzstein, University of Zurich