The time which it takes to obtain a shock normal from a given data set depends for the most part on three points: the number of data points included in the fit; the resolution of both the solution and the plot grid; and whether the C backing code has been compiled.
The number of points included with a computation depends both on the time intervals selected and whether the upstream and downstream data are to be processed using MIRROR SYMMETRY or through a PERMUTATION of the data. It is always better at first try to use MIRROR SYMMETRY especially if the C backing code is not installed.
By far the bulk of the processing time occurs in establishing the surface solution plot. Since this plot is not essential to the solution other than to establish the number of minima, there is no reason to produce it at very high resolution. A 10x10 solution or even a 15x15 solution resolution is generally sufficient. Without the C backing code the grid and solution resolution are best left identical using a 5x5 grid at the smallest.