yes, I got similar timing values. The testing is quite difficult. The results are almost equal when I change both versions to search only for one point of a way, so I am positive that it works, but there are still differences. I can't prove it, but I assume that both versions are wrong, because both don't find info for some points which lie inside the quadtree.
I assume those are points that lie exactly on a boundary AND on a quadtree bbox.
I'd try the following: if a point is within the quadtree , but no info is found, the algorithm should try again with a point next to the original.
> Date: Fri, 13 Jan 2012 19:20:16 +0100
> From: wmgcnfg@web.de
> To: mkgmap-dev@lists.mkgmap.org.uk
> Subject: Re: [mkgmap-dev] [Patch v3] LocationHook with new Quadtree
>
> Now some performance facts using my standard 15 tile map:
>
> r2168: LocationHook time: 44,8s
> patch: LocationHook time: 11,5s
>
> So the time for the LocationHook is reduced to 1/4. Great!
>
> But:
> The tiles differ very much between r2168 and the patched version. So
> there is some investigation work required to check if the patched
> version does what it should do.
>
> WanMil
>
> > Hi,
> >
> > now complete and based on r2167.
> >
> > Gerd
> >
> > http://gis.638310.n2.nabble.com/file/n7181669/locationHook_speedup_v3.patch
> > locationHook_speedup_v3.patch
> >
> > --
> > View this message in context: http://gis.638310.n2.nabble.com/Patch-v3-LocationHook-with-new-Quadtree-tp7181669p7181669.html
> > Sent from the Mkgmap Development mailing list archive at Nabble.com.
> > _______________________________________________
> > mkgmap-dev mailing list
> > mkgmap-dev@lists.mkgmap.org.uk
> > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
>
> _______________________________________________
> mkgmap-dev mailing list
> mkgmap-dev@lists.mkgmap.org.uk
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev