At Mon, 11 Aug 2008 18:55:04 +0200,
Carlos Medrano wrote:
> I would like to use linear algebra functions, for instance
> gsl_linalg_cholesky_decomp
>
> The problem is that I am using float numbers (gsl_matrix_float
> etc). It seems that these functions only accept double
> (gsl_matrix). Are there some linear algebra functions specific for
> float?
Only the standard BLAS functions, gsl_blas_sgemm etc. For other
functions we only support double.
> If not, which is the best way to manage this situation? Just
> copy to a double matrix and then copy back the result?
Yes, or work in double precision.
--
Brian Gough
GNU Scientific Library -
http://www.gnu.org/software/gsl/
