Re: [Bug-gsl] Two possible bugs

[Brian Gough]

claudio writes:
 > Hi
 > 
 > 1) if I try to compile a my program that uses gsl    
 >    with icc (intel c compilers) with -DD options I obtain this
 >    message (icc version 7 and 8) :
 > 
 > /usr/include/gsl/gsl_linalg.h(457): error: expected a ")"
 >   gsl_linalg_balance_columns (gsl_matrix * A, gsl_vector * D);

Hello,

If -DD redefines the "D" in the prototype as a macro D => 1 maybe that
causes problems.

 > 2) If I compile a code with icc 8 (there is no problem with 7)
 >  the subroutine single value decomposition doesn't work
 >  it conitnue to run to infinity

Can you provide a send example program as test case? (e.g. maybe it
depends on the matrix). Perhaps you can use some printf statements to
see why it is failing to converge, comparing your compiler with GCC.

 > Ps. A question I would like to contribute to GSL
 > How I sould I do? I  wrote some subroutine and I'm also
 > a clever webmaster

Helping to solve this svd problem would be very useful

regards

-- 
Brian Gough