This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: Making some minor contributions to GSL
- From: Gerard Jungman <jungman at lanl dot gov>
- To: linas at austin dot ibm dot com
- Cc: gsl-discuss at sources dot redhat dot com
- Date: Tue, 13 Jan 2004 15:56:00 -0700
- Subject: Re: Making some minor contributions to GSL
- Organization: Los Alamos National Laboratory
- References: <20040113121703.B39552@forte.austin.ibm.com>
On Tue, 2004-01-13 at 11:17, linas@austin.ibm.com wrote:
>
> I have some code that might be appropriate for inclusion with GSL,
> and was curious to gauge interest, and understand the hurdles
> involved.
>
> Polygamma functions, Abramowitz & Stegun chapter 6.4 -- these
> are derivatives of the gamma (factorial) function. The algo
> I have is fairly straightforward ... and is accurate to about
> 1e-14 or so for "small" orders and arguments. (The "multiplication
> theorm" breaks down above order 30 or 40 or so, i.e. for arguments
> greater than a 100 or 200 or so. I'm not sure why.)
I have never made a study of robustness of the GSL
implementation, in gsl_sf_psi_n(). There are only
the test cases in specfunc/test_sf.c. Is there
a specific deficiency in the GSL implementation?
If you have some test cases that break it, please
send them so I can check it out. If you have code
that can plug a hole in the current implementation,
then that would be greatly appreciated.
Thanks,
G. Jungman