Re: gobject n_preallocs

[Tim Janik <timj gtk org>]

On Thu, 16 Jan 2003, Owen Taylor wrote:

>
> tom <tom dbservice com> writes:
>
> > i  made a improved test... (file below)
> >
> > with n_prealloc = 1000:
> > created:  ~200000
> > freed: ~1500
> >
> > and n_prealloc = 0:
> > created: ~150000
> > freed: ~1500
> >
> > if you look at this you see that setting n_prealloc even decreases
> > performance...
>
> Nothing very unexpected about this. n_preallocs makes GObject
> use GMemChunk, and normal malloc is known to be faster than
> GMemChunk

right, assuming glibcs malloc implementation, using GMemChunk is
actually a complicated way of saying "i want to waste a bit of
performance and a lot of memory" and GType's n_prealloc enables
that for object allocations.

> except on systems with highly broken malloc()
> implementions.

owen, what's your take on ignoring the memchunk implementation once
we detected glibc, and simply revert to malloc/free in that case?
we'll have to keep an extra list of things to free upon
memchunk destruction, but that's still much better than the
current wastefull aproach.

>
> Regards,
>                                         Owen

---
ciaoTJ