Re: [eigen] Specializing max_coeff_visitor for some number types

[Marc Glisse <marc.glisse@xxxxxxxx>]

On Mon, 7 Apr 2014, Marc Glisse wrote:

> As you can see in LU, there is a subtlety I hadn't thought of: max(abs) was 
> returning a RealScalar, not a Scalar. That's not a big problem when we only 
> want to compare it with 0.

Note that we end up computing twice as many abs() as before, because we 
store the coefficient in the visitor and not the result of abs. Avoiding 
it would be possible, adding a new unary functor that defaults to 
scalar_abs_op, storing its result_type in the visitor, calling it before 
passing the result to better_coeff_op, and making better_coeff_op work on 
this new type. But that's starting to get a bit complicated for such a 
simple feature...

--
Marc Glisse