Hi,I wonder if there is a reason that gsl_multifit_nlinear_fdf does not include a pointer to a function that computes both the function and its Jacobian at the same time, because this is often more efficient than doing it separately.
gsl_multiroot_function_fdf does have this option.Or can we assume that, e.g., whenever the function for the Jacobian is called, the function itself is called with the same parameters before so that one can reuse intermediate steps internally?
Cheers, Olaf