Boost logo

Boost :

From: Joerg Walter (jhr.walter_at_[hidden])
Date: 2002-07-01 00:40:52


----- Original Message -----
From: "Benedikt Weber" <weber_at_[hidden]>
Newsgroups: gmane.comp.lib.boost.devel
To: <boost_at_[hidden]>
Sent: Sunday, June 30, 2002 2:19 PM
Subject: [boost] Re: uBLAS and expression templates

> Joerg
>
> > It's a proven (but not well known ;-) technique to reintroduce
> > temporaries at the right places into the expression, i.e. writing A *
> vector
> > (B * v).
>
> I am glad there is at least a possibility to force the creation of a
> temporary. However, the user should not have to be concerned about such
> details. I never did anything with expression templates and I only
recently
> read about them in the context of the Spirit parser. However, I think it
> should be possible to change the design so a temporary is automatically
> introduced where appropriate (say for the prod() functions). I urge you to
> think about a possibility to change that. It would be a real benefit for
the
> library.

Ok. After discussing this issue intensely with Mathias (thanks, Mathias!),
I've added overloads like the following:

    template<class M, class E1, class E2>
    NUMERICS_INLINE
    M
    prod (const matrix_expression<E1> &e1,
          const vector_expression<E2> &e2) {
        return M (prod (e1, e2));
    }

With these new overloads, we'd recommend to write your original sample in
the following way:

    t.restart ();
    for (int run = 0; run < 1000; ++ run)
        r.assign (prod (prod<matrix<double> > (A, B), v));
    std::cout << t.elapsed () << std::endl;
    t.restart ();
    for (int run = 0; run < 1000; ++ run)
        r.assign (prod (A, prod<vector<double> > (B, v)));
    std::cout << t.elapsed () << std::endl;

A couple of questions remain open:
- Should we rename the original ET' based prod() functions to something like
defer_prod()? - Is it possible to determine a sensible default for the type
of the temporary (for example via some traits classes)?

[snip]

Regards

Joerg


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk