Statistics – Computation
Scientific paper
Mar 2004
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2004aps..marv38001n&link_type=abstract
American Physical Society, March Meeting 2004, March 22-26, 2004, Palais des Congres de Montreal, Montreal, Quebec, Canada, MEET
Statistics
Computation
Scientific paper
Adaptive mesh refinement (AMR), developed by Berger and collaborators for resolving regions of large gradients in numerical solutions of hyperbolic problems, has been adapted to computational cosmology. The formation of galaxies and large scale structure in the universe is caused by gravitational instability amplifying seed perturbations laid down in the early universe. The nonlinear growth and interactions of these pertubations across a vast range of length scales makes this a challenging numerical problem and a good candidate for AMR. Our extensions to AMR include adding a particle-mesh algorithm to model collisionless dark matter and stars, self-gravity, cosmic expansion, multispecies reactive hydrodynamics, ionization kinetics, radiative heating and cooling, and a heuristic treatment for star formation. We have implemented a portable, parallel 3D code which scales well to hundreds of processors. We present sample applications to early galaxy formation which achieve an unprecendented spatial dynamic range of one trillion in localized regions of the simulation volume. We discuss planned future developments and applications to the physics of the early universe. This work has been supported by the NSF Partnerships for Advanced Computational Infrastructure program through the NCSA Alliance.
No associations
LandOfFree
Adaptive Mesh Refinement in Computational Cosmology does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Adaptive Mesh Refinement in Computational Cosmology, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive Mesh Refinement in Computational Cosmology will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1621764