Re: [lmi] upgrade to xmlwrapp-0.6.0

[Vaclav Slavik]

On Fri, 2009-04-24 at 02:04 +0000, Greg Chicares wrote:
> > That's strange, I measure ~40% improvement (I don't remember if it was
> > in input_test or only iterators-manipulating subset of it, though) with
> > MinGW's gcc-3.4. Did you use 3.4 or 4.3 in your tests? The improvement
> > in performance with gcc-4.x is so small it's only observable under
> > valgrind, but I did see real difference with 3.4 back when I was working
> > on this...
> 
> I can't see any such improvement in 'input_test'.

I'm sorry, my memory was a bit hazy: the improvement was in the test
that exercised iterators copying the most, i.e. the now-removed "Vector"
test. I reverted my tree to CVS 6 days ago (before your merged in
nodes_view changes) and was able to replicate considerable speedup with
mingw/gcc-3.4 (median of five runs):

                     with     without   ratio
    Vector:        419760     1085911   61.3% 

Adding boost::pool was the first attempt to improve performance,
nodes_view and node::elements() came afterwards and changed the picture.

> matter. But maybe the version of libstdc++ does matter? 

Generally speaking, it does. gcc4's libstdc++'s default allocator is as
good as boost::pool (at least on Linux).

Regards,
Vaclav