On Jul 31, 2012, at 11:19 AM, Daniel Wheeler wrote: > This might be crazy slow if this is done point by point. Might be better to > create the interpolated phi array beforehand in one shot and then do the > integral.
Another option is not to interpolate at all and use FiPy's cell-centered values: x, = mesh.getCellCenters() s = numerix.argsort(x) numerix.exp(-scipy.quad(phi.getValues()[s], 0, x[s])[0]) Because these are cell-centered values, there are probably edge effects to worry about. The sort is probably not necessary on a Grid1D, but it's safer. _______________________________________________ fipy mailing list fipy@nist.gov http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]