2014-05-26 10:12 GMT+02:00 Vladimir Batov
Joel de Guzman wrote
On 5/25/14, 6:34 PM, Jeroen Habraken wrote:
This makes me wonder. How fast is Convert BTW? How does it compare to the examples in that benchmark? Boost is performance hungry! It will be a real cause of concern if it is slower than the benchmarks.
"convert" has no benchmarks of it own... because it does nothing. :-) It's a "manager" class rather than a "worker" class. It's merely an interface to coordinate/unify and uniformly deploy (or quickly swap on when-needed basis) varying conversion facilities. So, with the "coerce"-based converter "convert" performs as fast as "coerce". I actually ran tests with lexical_cast and others: naked lex_cast vs. convert+lex_cast. No performance degradation. I mention it in the docs.
Can this be true? Your converter concept returns a (small) bool. The API returns something like optional<T> (likely something significantly bigger than bool). This alone could add to performance hit. Worth measuring, at least. Regards, &rzej