In order to help evaluate all of the recent NMG work by Richard, myself, and others, I've written a script that tests whether objects can be successfully converted to NMG and BoT forms within a given timeframe while tracking stats. The sh/conversion.sh script is pretty trivial to use (sh/conversion.sh path/to/file(s).g) as you just specify file(s) and it'll then automatically attempt to convert every object.

Here's the output summary for the previous 7.16.10 release (before a lot of recent NMG/LIBRT/LIBBN changes) on all of our own sample geometry:

   Files:  31
 Objects:  9244
Failures:  155 NMG, 185 BoT
NMG conversion:  98.3%  (9089 of 9244 objects)
BoT conversion:  98.0%  (9059 of 9244 objects)
  Success rate:  98.2%
Elapsed:  5382 seconds
Average:  0.6 seconds per object

That's allowing 5min per conversion and is about what I'd expect. Here's the count when we only allow 5sec per conversion:

Failures:  155 NMG, 185 BoT

Notable already is that 1) the overwhelming majority of objects convert within just a couple seconds, 2) there's a bug converting NMG objects into BoTs even though that should be a direct translation, and 3) the failures are far too many given this is our sample geometry (we should be 100%).

The script is useful to anyone working on the NMG or LIBRT/LIBWDB facilities for tessellating and converting. Specifically, it should tell you if a change has verifiably made things better or worse, faster or slower, with greater or fewer failures.

I'm running the statistic again on the main code using HEAD now that several changes are incorporated but anyone can run the scripts with a simple: ./sh/conversion.sh path/to/file(s).g

Cheers!
Sean

------------------------------------------------------------------------------
Increase Visibility of Your 3D Game App & Earn a Chance To Win $500!
Tap into the largest installed PC base & get more eyes on your game by
optimizing for Intel(R) Graphics Technology. Get started today with the
Intel(R) Software Partner Program. Five $500 cash prizes are up for grabs.
http://p.sf.net/sfu/intelisp-dev2dev
_______________________________________________
BRL-CAD Developer mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/brlcad-devel

Reply via email to