Re: [boost] [range] Performance benchmarks?

Subject: Re: [boost] [range] Performance benchmarks?
From: Thorsten Ottosen (thorsten.ottosen_at_[hidden])
Date: 2012-02-02 05:14:43

Den 02-02-2012 03:00, Dave Abrahams skrev:
>
> on Wed Feb 01 2012, Mathias Gaunard<mathias.gaunard-AT-ens-lyon.org>  wrote:

>> They are in Boost, and in particular in Boost.Range, which was being
>> asked about here.
>
> I understand.  Maybe Beman didn't know that Boost.Range ranges were
> iterator pairs.  In the end,
>
>    We're explicitly interested in what happens to efficiency when
>    algorithms are /implemented/ using range primitives rather than by
>    iterator increment, comparison, etc.
>

Well, the only way to answer that is by implementing Alexandrescu's 
range abstractions and start comparing it with BoostRange. Not trivial 
IMO. Then make sure that the underlyigng algorithms/loop-unrolling are 
exactly the same. So basically, a huge (but very interesting) work.

An advantage of range primitives is whever we need to store a predicate 
or functor; in Boost.Range, we need to store one in each iterator which
is not optimal (clever implementation of 
transform_iterator/filter_iterator would optimize this away for empty 
functors/predicates).

Again, I don't know what the speed difference will be between the two 
ideas; If the data being processed is large, I guess storing an extra
function in the end iterator is of no significance.

The iterator-based range library also seem to work well with the 
existing standard library;

-Thorsten