HLRB Project h007y
Core-Collapse Explosions
Max-Planck-Institut fuer Astrophysik
Proposing Institution
Max-Planck-Institut fuer Astrophysik
Project Manager
Dr. Hans-Thomas Janka
Karl-Schwarzschildstr. 1
85741 Garching
Abstract
Core-collapse supernovae, i.e. the explosion of stars with masses larger thaneight times the mass of our Sun, are among the most energetic events in theUniverse. Furthermore, core-collapse supernovae are the major source for theproduction of elements heavier than iron. These heavy elements are admixedduring the explosion to the interstellar matter and can be later found inthe subsequent generations of stars. Thus supernova explosions are important for the chemical evolution of stars and galaxies in the Universe.However, the details of the explosion mechanism are still unclear and numerical simulations are needed to investigate this issue. Recently, the Garching supernova group developed a numericalcode that allows one to study the explosion mechanism of core collapsesupernovae with a refined spectral treatment of the neutrino physics anda detailed description of neutrino-matter interactions in spherical symmetryas well as multi-dimensional environments (seeRampp & Janka A&A 396, 2002 ; Buras et. al, A&A 447, 2006).The high dimensionality of the transport problem and the long evolution timethat requires hundreds of thousands of time steps, however, have the drawback of high computational costs for a two-dimensional simulation.With the requested ``test-account'' on the HLRB IIsupercomputer of the Leibniz Rechenzentrum we want to test the performance(e.g. scalability) of our code on a large number of processors (i.e. N > 128) for an OpenMP and hyprid OpenMP-MPI version of the code. Since our code is already successfully applied on a 112 processor SGI Altix system, major optimization workfor the HLRB II hardware architecture is not expected. Theexperience we will gain with the ``test account'' will be helpfulfor judging the perspectives and defining the needs of a later``full'' project proposal.
