[Help-glpk] Re: glpk and ATLAS/GSL

[Andrew Makhorin]

>I meant moderately sized non-sparse LP problems.
>If you have lots of memory and several processors
>that can be a big advantage.

I think it would be more reasonable to have a dense version of the
simplex method rather than to replace existing basis maintenance
routines. However, I am not sure that that could be a big advantage.

>If i will write a
>gsl_solve_LP(gsl_matrix *A, gsl_vector *b, gsl_vector *c)
>function for { Ax>b, min(cx) } LP problem, is it really 
>necessary to reallocate and refill A in a lpx_* way ?

Yes, it is necessary, because glpk uses its own data structures.

Andrew Makhorin