I updated v.voronoi in trunk r40286 and devbr6 r40288 to use much less
memory, it's also a bit faster. There is still work to do on it, but it
does much better now. The memory bottleneck is now no longer the
triangulation but topology building. Please test.
And another comment from Martin about the algorithm used in the current
I started studying Fortune sweepline algorithm. As soon as I make any
significant progress with the implementation, I will inform you. As a
matter of fact, original version of v.voronoi/delaunay uses a Fortune
algorithm from netlib, which is quite a crude implementation. I am going
to create a new code from scratch. However, I will use the previous code
as a guideline. Even though I am going to implement the same algorithm,
new code should be much more efficient, after I do some polishing.
grass-dev mailing list
[hidden email] http://lists.osgeo.org/mailman/listinfo/grass-dev