Re: [Help-gsl] LM nolin fit problem

[Joakim Hove]

Jim Cadien <address@hidden> writes:

> There is no problem compiling, but when it runs, I get an error,
> occuring on the call to
>
> gsl_multifit_fdfsolver_set (s, &f, &x.vector);
>
> gsl: vector_source.c:45: ERROR: index out of range
> Default GSL error handler invoked.
> Aborted
>
> I slightly modified the vector_source.c code to have it print the
> values i and v.size.  The number of points in my data is 15.  The
> value of v.size is 15, but i=3542.  I really have no idea what i
> represents, and therefore no idea what it should be or how it got to
> be that value.

Well,

your modified vector_source.c tells you that you are trying to access
element i=3542 of a vector with only 15 elements - clearly an index
out of range.

As to the source of the problem, it is probably in your LM code, and
without access to it, no one can help you.


Joakim



-- 
Joakim Hove
hove AT ift uib no                 /    
Tlf: +47 (55 5)8 27 90            /     Stabburveien 18          
Fax: +47 (55 5)8 94 40           /      N-5231 Paradis           
http://www.ift.uib.no/~hove/    /       55 91 28 18 / 92 68 57 04