[Charlotte.PM] Followup - Help me tune this for speed
drewhead at drewhead.org
Wed Aug 10 07:21:42 PDT 2005
-----BEGIN PGP SIGNED MESSAGE-----
I posted on Perl Monks a week(ish) ago for help with this problem.
Received several good replies. I was able to get my test data run down to
about 3 minutes from ~17. Significant, but not good enough.
Most significantly I got a lesson in math. It was asserted to me
that what I was trying to do was a Delaunay Triangulation. Never hard of
that one so off for some google + reading went I. It turns out this is
what I'm trying to do. Unfortunately there doesn't seem to be a perl
implementation of Delaunay Triangulation at the present time. However I
was able to find an implementation called qhull http://www.qhull.org.
Qhull blows everything I've done away, computing all the unique
neighboor edges in less than a second, even on my 2001 point data. Perl
just forks the process. The example cited for this on Perl Monks is
pretty much the code I'm using. At these speeds there is little incentive
for me to continue looking for a pure perl solution.
Now that I can build all the stats I ever wanted it's time for me
to figure out how to contruct a graphic file showing what the set of
coordinates look like.
Drew Dowling Drewhead http://www.drewhead.org
drewhead at drewhead.org | | WWW / \ Alpha Phi Omega
Concord, North Carolina | | Nimat / \ Gamma Lambda
CLEMSON UNIVERSITY ALUMNI | | Apatschin /_____\
TIGER BAND! | VGAP4 Hosting at http://vgap.drewhead.org
-----BEGIN PGP SIGNATURE-----
Comment: Public key available at http://www.drewhead.org
-----END PGP SIGNATURE-----
More information about the charlotte