Improved algorithm for calculating the Chandrasekhar function Original Research Article

Theoretical models of electron transport in condensed matter require an effective source of the Chandrasekhar image function. A code providing the image function has to be both accurate and very fast. The current revision of the code published earlier [A. Jablonski, Comput. Phys. Commun. 183 (2012) 1773] decreased the running time, averaged over different pairs of arguments image and omega, by a factor of more than 20. The decrease of the running time in the range of small values of the argument image, less than 0.05, is even more pronounced, reaching a factor of 30. The accuracy of the current code is not affected, and is typically better than 12 decimal places.