Parallel Algorithms for Cosmological Statistics at NCSA
Benjamin D. Wandelt
Award year: 2003-2004
The recent release of the preliminary analysis of cosmic microwave background (CMB) data taken by the Wilkinson Microwave Anisotropy Probe (WMAP)  made front page news. The parameters of modern physical cosmology are now known better than ever before. However, the results that were published by the WMAP science team  are only a first taste of what is possible if the full statistical information in this data set were extracted. Therefore, the WMAP data set presents golden opportunities: 1) it is currently the most precisely measured cosmological data set for testing the fundamental assumptions of the standard model of cosmology (inflation, adiabatic Gaussian perturbations) on the largest scales and at the earliest moments in time. 2) It contains qualitatively new, untapped information on these properties of the early Universe which will only be accessible with sophisticated analysis techniques. 3) Tantalizing hints of departures from the standard model on super large scales need to be studied using exact methods which allow a detailed assessment of statistical significance.
The goal of this proposal is to take advantage of these opportunities by building on preparatory work by me, my students and my postdoc over the past year and combining our expertise and long standing involvement in CMB satellite missions with the computing excellence and facilities at NCSA. I demonstrate in this proposal that this combination of assets will enable us to overcome the challenges that currently prevent the application of more sophisticated methods to CMB data sets and allow us to build production quality parallel implementations of our advanced computational methods. We will apply them to forthcoming semi-annual releases of the WMAP data set. As a result, these codes will be uniquely capable of handling the complexities of future CMB data sets, promising an enduring role for NCSA at this fast-moving frontier of cosmology.