[vtkusers] Trouble with ModifiedBSPTree

Biddiscombe, John A. biddisco at cscs.ch
Wed Feb 29 15:30:04 EST 2012


Rounding error/tolerance?

There's one place in the ModifiedBSPTree code where a > and >= (or < and <=) can make a small difference to the tree generation, and possibly also to intersection tests.

If you have a test that you can easily send to me, I'll have a look at it. There are a couple of places where for optimization reasons I used < in place of <= (etc) and wondered secretly if it would ever make a difference.

Why it is different on two machines which are otherwise similar I can't really guess - it ought to be reproducibly wrong on both or neither really, but I'm still happy to check it out anyway.

JB

-----Original Message-----
From: vtkusers-bounces at vtk.org [mailto:vtkusers-bounces at vtk.org] On Behalf Of David Doria
Sent: 29 February 2012 14:37
To: vtkusers at vtk.org
Subject: [vtkusers] Trouble with ModifiedBSPTree

I work about half time on two different computers, both running Ubuntu 11.10. On one of them, my code that uses ModifiedBSPTree to perform a bunch of ray/mesh intersections works perfectly (in a test case of a
10x10 grid of rays aimed at a plane, all of the rays intersect the plane, as they should). On the other computer, however, about half of the exact same rays on the exact same plane do not intersect! (there doesn't seem to be a pattern to which rays in the grid do not intersect). I realize this is a pretty vague problem statement, but can anyone think of a reason why the results would be different on different computers? The computers are both using git master VTK.

Any thoughts of why this would happen or what else I can do to troubleshoot are appreciated!

David
_______________________________________________
Powered by www.kitware.com

Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html

Please keep messages on-topic and check the VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtkusers



More information about the vtkusers mailing list