Re: [Bug-gsl] Problem with gsl_sf_legendre_sphPlm

[James Casey]

Please disregard my last email.  I found the problem.  PyGSL returns a
complex variable, but sphPlm(x) is not actually complex.  The
imaginary part is just completely wrong but after some checking it
turns out that the real part is actually correct, and is checked in
the gsl package.  I just have to ignore the imaginary part that is
returned.  I am just have a bad math day I think.

Cheers,
James Casey

On Sun, Jun 1, 2014 at 12:28 PM, James Casey <address@hidden> wrote:
> Hello,
>
> I am running GSL version 1.15.  I have originally ran into the problem
> with pygsl and verified that the problem is actually in GSL by running
> the C libraries.  When running:
>
> gsl_sf_legendre_sphPlm(l,m,cos(t))
>
> I have problems for large l, and small t.  Specifically, with l = 192,
> m = 1, and t = 0.01,  the values of the spherical harmonics are
> completely wrong.  The complex component is several orders of
> magnitude too large and the real part is greater than one, which
> should not happen for a normalized function.  I haven't tried to look
> into the code myself, yet, but I wanted to know if anyone else had
> reported this or if anyone else was experiencing similar problems.
>
> Cheers,
> James Casey
>
> --
> James Casey, Graduate Research Assistant
> Center for Relativistic Astrophysics
> School of Physics - Georgia Institute of Technology
> 837 State St. Atlanta GA, 30332-0430



-- 
James Casey, Graduate Research Assistant
Center for Relativistic Astrophysics
School of Physics - Georgia Institute of Technology
837 State St. Atlanta GA, 30332-0430