RE: ODE solvers with complex data

Thanks, that seems like the approach to take. It makes sense not to make
duplicate functions. I don't know the math behind all this well enough to
fully understand it, but I'm learning as I try to convert my source code.

Cheers,
Michael

> -----Original Message-----
> From: Brian Gough [mailto:bjg@xxxxxxx] 
> Sent: Friday, November 02, 2007 4:49 AM
> To: Michael Stauffer
> Cc: help-gsl@xxxxxxx
> Subject: Re: [Help-gsl] ODE solvers with complex data
> 
> 
> At Wed, 31 Oct 2007 20:29:38 -0400,
> Michael Stauffer wrote:
> > My first task with GSL is to convert some Matlab code to C. 
> The code 
> > uses the Matlab ODE solvers, and so I'm planning on using the GSL 
> > odeiv routines. However, the Matlab code uses complex data, but it 
> > seems the GSL odeiv routines only use double-type data. The 
> GSL odeiv 
> > system "function" routine specifies input and output as 
> type double. I 
> > figured that since GSL supports complex data types, there would be 
> > matching odeiv functions for complex types if it were 
> possible to use 
> > them.
> 
> For odes, you'll need to work with the real and imaginary 
> components as doubles in an array of length 2N.
> 
> We don't generally have complex versions of every function 
> when the real version can be used, as it leads to a lot of 
> duplication in the code.
> 
> -- 
> best regards,
> 
> Brian Gough
> 
> GNU Scientific Library -
> http://www.gnu.org/software/gsl/