We use a systolic N-body algorithm to evaluate the linear stability of the gravitational N-body problem for N up to 1.3 × 10^5, two orders of magnitude greater than in previous experiments. For the first time, a clear N-dependence of the perturbation growth rate is seen, µe ∼ lnN. The e-folding time for N = 10^5 is roughly 1/20 of a crossing time (Refer to PDF file for exact formulas).
Department, Program, or Center
School of Physics and Astronomy (COS)
Astrophysical Journal 2002 November 20 vol. 580, issue 1, p. 606-609
RIT – Main Campus