[Clam-devel] TokenDelay patch

[David García Garzón]

That's what we did originally, Sam, but according Xavi's profiles the reserve 
didn't work as intended, at least on windows stl.


On Dissabte 05 Maig 2007, Sam Roig wrote:
> yes.
>
> why not use an stl vector insted and call reserve() to make sure the
> final vector will have enough capacity for whatever your are pushing
> on to it?
>
> then push_back() will only resize if the capacity needed wasn't reserved
> first.
>
> On 4/21/07, David García Garzón <dgarcia at iua.upf.edu> wrote:
> > On Divendres 20 Abril 2007, Xavier Amatriain wrote:
> > > On Fri, 2007-04-20 at 19:51 +0200, Pau Arumi wrote:
> > > > This can't break anything, ho ahead, commit it. We can improve it
> > > > later. Like leaving only one representation and get rid of
> > > > #ifdefs
> > >
> > > Will do..
> > >
> > > > However, I don't see what operations are more efficient using a
> > > > vector. Can you explain?
> > > > As I see it, when you copy, say, an spectrum its buffers are
> > > > copied in both cases.
> > >
> > > It's not about the copy, which does exist in both cases, but about the
> > > rellocation of memory. Every time you do a push or a pop, the queue
> > > resizes itself... or so does my profiler say ;)
> >
> > uau! thats an ugly backstab from the STL, I think we did a preallocation
> > to avoid this, but it wasn't working. Anyway, why a CLAM::Array instead
> > of a std::vector ;-)
> >
> > --
> > David García Garzón
> > (Work) dgarcia at iua dot upf anotherdot es
> > http://www.iua.upf.edu/~dgarcia
> >
> > _______________________________________________
> > Clam-devel mailing list
> > Clam-devel at llistes.projectes.lafarga.org
> > https://llistes.projectes.lafarga.org/cgi-bin/mailman/listinfo/clam-devel



-- 
David García Garzón
(Work) dgarcia at iua dot upf anotherdot es
http://www.iua.upf.edu/~dgarcia
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.clam-project.org/pipermail/clam-devel-clam-project.org/attachments/20070506/ffd82269/attachment-0003.pgp>