next up previous
Next: Computation of convergence times Up: Methods Previous: Methods

Numerical integration

All integration was performed with a custom-written implementation of the 4th-order adaptative stepsize Runge-Kutta algorithm (Press, 1992), with $ 10^{-3}$ relative accuracy. Source code is available at http://cinquin.org.uk/ada/ada_blas_runge_kutta.html. The data was plotted using GMV or gnuplot.