[postgis-devel] New node splitting algorithm for GiST

Alexander Korotkov aekorotkov at gmail.com
Wed Oct 12 12:07:24 PDT 2011


It's O(N*log(N)), or more particular O(M*N*log(N)), where M - number of
dimensions.
My experimens show that difference in CPU usage with linear algorithm is
insignificant.

------
With best regards,
Alexander Korotkov.

On Wed, Oct 12, 2011 at 10:59 PM, Paul Ramsey <pramsey at opengeo.org> wrote:

> Is the algorithm still O(N) or is it O(N^2) ala Guttman?
> P.
>
> On Wed, Oct 12, 2011 at 11:58 AM, Alexander Korotkov
> <aekorotkov at gmail.com> wrote:
> >
> "A new double sorting-based node splitting algorithm for R-tree", A. Korotkov
> > http://syrcose.ispras.ru/2011/files/SYRCoSE2011_Proceedings.pdf#page=36
> > I believe it's applicable to not very high dimensions (3 or 4). On 5 and
> > more dimensions guttmann quadratic algorithm could become more effective
> > than my.
> > ------
> > With best regards,
> > Alexander Korotkov.
> > On Wed, Oct 12, 2011 at 10:49 PM, Paul Ramsey <pramsey at opengeo.org>
> wrote:
> >>
> >> And, in general terms, do you think this approach would be applicable
> >> in higher dimensions?
> >> P
> >>
> >> On Wed, Oct 12, 2011 at 11:49 AM, Paul Ramsey <pramsey at opengeo.org>
> wrote:
> >> > Thanks for this, could you give me a reference to your paper too?
> >> > P.
> >> >
> >> > On Wed, Oct 12, 2011 at 10:52 AM, Alexander Korotkov
> >> > <aekorotkov at gmail.com> wrote:
> >> >> Hi!
> >> >> Heikki Linnakangas recently commit my patch with new node splitting
> >> >> algorithm for GiST to PostgreSQL master branch:
> >> >>
> >> >>
> http://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=7f3bd86843e5aad84585a57d3f6b80db3c609916
> >> >> Since it can significantly accelerate index search, you might be
> >> >> interested
> >> >> to apply same changes to PostGIS.
> >> >> Discussion threads are here:
> >> >> http://archives.postgresql.org/pgsql-hackers/2011-10/msg00158.php
> >> >> http://archives.postgresql.org/pgsql-hackers/2011-09/msg00576.php
> >> >> ------
> >> >> With best regards,
> >> >> Alexander Korotkov.
> >> >> _______________________________________________
> >> >> postgis-devel mailing list
> >> >> postgis-devel at postgis.refractions.net
> >> >> http://postgis.refractions.net/mailman/listinfo/postgis-devel
> >> >>
> >> >>
> >> >
> >> _______________________________________________
> >> postgis-devel mailing list
> >> postgis-devel at postgis.refractions.net
> >> http://postgis.refractions.net/mailman/listinfo/postgis-devel
> >
> >
> > _______________________________________________
> > postgis-devel mailing list
> > postgis-devel at postgis.refractions.net
> > http://postgis.refractions.net/mailman/listinfo/postgis-devel
> >
> >
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-devel/attachments/20111012/305928bd/attachment.html>


More information about the postgis-devel mailing list