Computational Cosmology at Berkeley

Many physical systems lack the symmetries necessary to allow analytic treatement, and direct numerical simulation is the best or only means of studying such systems. For this reason computers are key to answering many questions in physical cosmology and our understanding of the universe is rapidly growing as computational tools and capabilities advance to meet this challenge. Cosmology is no longer a data-starved discipline, rather we are required to cope with a data deluge. Whether it be data from the cosmos or from simulations, the ability to store, organize and process large amounts of data is key to progress in our endeavors.
Numerical methods are crucial in
  • Simulating the nonlinear collapse and subsequent dynamics of dark matter --- the dominant component of the mass in the Universe.
  • Complex multi-physics models involving gravity, hydrodynamics, star formation, feedback, ...
  • Creating mock catalogues to optimize experimental design, test and calibrate analysis pipelines and methods, and tease insight from TB of data.
  • Developing and applying advanced statistical tools to analyze data with high precision (e.g. the CMB)
  • Modeling physical processes in objects used to trace the expansion of the universe (e.g. supernovae)
Many of the current observational and simulated data sets are huge (Terabytes of data) so that the analysis of the simulations and comparisons with observation and theory requires focussed and specialized effort; experts in data analysis are needed in order to capitalize on the amazing potential of the numerical and observational data once it has been reduced, this again involves intensive computing.

Many of the cosmologists at Berkeley are heavily involved in developing and using computational methods in the ways described above. One focus is developing the Nyx code. Other local computational cosmology pages include

However, there are also many excellent public codes which are used by the group as well, for example:

Back to BCG home page.