Alexander Nasonov wrote:
Alexander Nasonov wrote:
David Abrahams wrote:
The tiny prototype that started all this is at: http://lists.boost.org/MailArchives/boost/msg53032.php
Thanks, Dave. This gives me an idea.
Now I don't think I found a good idea, forget it. But I've tried another idea: If all functions have same arity, signatures of several functions can be deduced in one call. This improves performance for huge function sets.
Great! :) But I don't really understand what you're talking about. Too much missing detail :(
In the end of my post you'll find numbers for algorithm "unique" based on typeof and mpl::set with and w/o this optimization. Unique signatures are at 1, 2, 10, 39 and 200.
And of course I don't know what the numbers mean either. -- Dave Abrahams Boost Consulting http://www.boost-consulting.com