here is a patch for the bn_mat_inverse() routine in /src/libbn/mat.c to support for orthogonal matrices which would be useful in my implementation of the pull routine.
just realised that the new patch added causes the benchmark tests to fail so appears there is a bug in code. Take a look at it today or early tomorrow to correct problem.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
here is the modified patch which applies smoothly in (/src/libbn/mat.c)
which supports for orthogonal matrices. But however when i run the benchmark tests it fails 3 of the tests. don't really understand what is going on.
Need some more eyes on it. I believe this would be useful in implementing the pull which will have to regain geometric transformations.
here are the three benchmark tests that are failing and here is the
modified patch in unified diff format for orthogonal matrix support of
bn_mat_inverse(/src/libbn/mat.c).
[Code]
Running the BRL-CAD Benchmark tests... please wait ...
+++++ moss
Frame 0: 355823 rays in 0.47 sec = 756380.62 rays/sec (RTFM)
Frame 1: 711646 rays in 0.93 sec = 766976.45 rays/sec (RTFM)
Frame 5: 11386336 rays in 14.14 sec = 805493.49 rays/sec (RTFM)
Frame 6: 22772672 rays in 28.33 sec = 803858.82 rays/sec (RTFM)
Frame 7: 45545344 rays in 55.88 sec = 815078.07 rays/sec (RTFM)
pixcmp pixels: 250916 matching, 3 off by 1, 11225 off by many
moss.pix: WRONG WRONG WRONG WRONG WRONG WRONG
+++++ world
Frame 0: 483779 rays in 1.42 sec = 339664.82 rays/sec (RTFM)
Frame 1: 967558 rays in 2.78 sec = 348032.81 rays/sec (RTFM)
Frame 4: 7740464 rays in 21.72 sec = 356355.31 rays/sec (RTFM)
Frame 5: 15480928 rays in 43.74 sec = 353948.07 rays/sec (RTFM)
pixcmp pixels: 245650 matching, 8 off by 1, 16486 off by many
world.pix: WRONG WRONG WRONG WRONG WRONG WRONG
+++++ star
Frame 0: 312053 rays in 0.99 sec = 314776.13 rays/sec (RTFM)
Frame 1: 624106 rays in 1.87 sec = 333886.60 rays/sec (RTFM)
Frame 4: 4992848 rays in 14.25 sec = 350428.56 rays/sec (RTFM)
Frame 5: 9985696 rays in 28.59 sec = 349227.71 rays/sec (RTFM)
Frame 6: 19971392 rays in 58.07 sec = 343894.57 rays/sec (RTFM)
pixcmp pixels: 262144 matching, 0 off by 1, 0 off by many
star.pix: answers are RIGHT
+++++ bldg391
Frame 0: 375758 rays in 1.41 sec = 267293.79 rays/sec (RTFM)
Frame 1: 751516 rays in 2.77 sec = 271297.68 rays/sec (RTFM)
Frame 4: 6012128 rays in 21.75 sec = 276379.10 rays/sec (RTFM)
Frame 5: 12024256 rays in 43.50 sec = 276423.57 rays/sec (RTFM)
pixcmp pixels: 262144 matching, 0 off by 1, 0 off by many
bldg391.pix: answers are RIGHT
+++++ m35
Frame 0: 655368 rays in 2.07 sec = 316345.39 rays/sec (RTFM)
Frame 1: 1310736 rays in 4.09 sec = 320639.67 rays/sec (RTFM)
Frame 3: 5242944 rays in 16.02 sec = 327181.69 rays/sec (RTFM)
Frame 4: 10485888 rays in 32.14 sec = 326255.47 rays/sec (RTFM)
pixcmp pixels: 259145 matching, 37 off by 1, 2962 off by many
m35.pix: WRONG WRONG WRONG WRONG WRONG WRONG
+++++ sphflake
Frame 0: 1889266 rays in 4.78 sec = 395428.08 rays/sec (RTFM)
Frame 1: 3778532 rays in 9.08 sec = 416224.06 rays/sec (RTFM)
Frame 2: 7557064 rays in 18.68 sec = 404658.57 rays/sec (RTFM)
Frame 3: 15114128 rays in 36.15 sec = 418042.75 rays/sec (RTFM)
pixcmp pixels: 262144 matching, 0 off by 1, 0 off by many
sphflake.pix: answers are RIGHT
The m_inv.patch file looks like it is reversed - it is subtracting the new
content, rather than adding it.
Which three of the benchmark tests failed?
[patches:#202] Added functionality for Orthogonal Matrices in
bn_mat_inverse(/src/libbn/mat.c)
Status: open Created: Tue Jun 25, 2013 07:36 PM UTC by Anonymous Last Updated: Fri Jun 28, 2013 07:14 PM UTC Owner: nobody
here is a patch for the bn_mat_inverse() routine in /src/libbn/mat.c to
support for orthogonal matrices which would be useful in my implementation
of the pull routine.
attachment of patch in a zip file
just realised that the new patch added causes the benchmark tests to fail so appears there is a bug in code. Take a look at it today or early tomorrow to correct problem.
here is the modified patch which applies smoothly in (/src/libbn/mat.c)
which supports for orthogonal matrices. But however when i run the benchmark tests it fails 3 of the tests. don't really understand what is going on.
Need some more eyes on it. I believe this would be useful in implementing the pull which will have to regain geometric transformations.
Nyah,
The m_inv.patch file looks like it is reversed - it is subtracting the new content, rather than adding it.
Which three of the benchmark tests failed?
Hi Cliff,
here are the three benchmark tests that are failing and here is the
modified patch in unified diff format for orthogonal matrix support of
bn_mat_inverse(/src/libbn/mat.c).
[Code]
Running the BRL-CAD Benchmark tests... please wait ...
+++++ moss
Frame 0: 355823 rays in 0.47 sec = 756380.62 rays/sec (RTFM)
Frame 1: 711646 rays in 0.93 sec = 766976.45 rays/sec (RTFM)
Frame 5: 11386336 rays in 14.14 sec = 805493.49 rays/sec (RTFM)
Frame 6: 22772672 rays in 28.33 sec = 803858.82 rays/sec (RTFM)
Frame 7: 45545344 rays in 55.88 sec = 815078.07 rays/sec (RTFM)
pixcmp pixels: 250916 matching, 3 off by 1, 11225 off by many
moss.pix: WRONG WRONG WRONG WRONG WRONG WRONG
+++++ world
Frame 0: 483779 rays in 1.42 sec = 339664.82 rays/sec (RTFM)
Frame 1: 967558 rays in 2.78 sec = 348032.81 rays/sec (RTFM)
Frame 4: 7740464 rays in 21.72 sec = 356355.31 rays/sec (RTFM)
Frame 5: 15480928 rays in 43.74 sec = 353948.07 rays/sec (RTFM)
pixcmp pixels: 245650 matching, 8 off by 1, 16486 off by many
world.pix: WRONG WRONG WRONG WRONG WRONG WRONG
+++++ star
Frame 0: 312053 rays in 0.99 sec = 314776.13 rays/sec (RTFM)
Frame 1: 624106 rays in 1.87 sec = 333886.60 rays/sec (RTFM)
Frame 4: 4992848 rays in 14.25 sec = 350428.56 rays/sec (RTFM)
Frame 5: 9985696 rays in 28.59 sec = 349227.71 rays/sec (RTFM)
Frame 6: 19971392 rays in 58.07 sec = 343894.57 rays/sec (RTFM)
pixcmp pixels: 262144 matching, 0 off by 1, 0 off by many
star.pix: answers are RIGHT
+++++ bldg391
Frame 0: 375758 rays in 1.41 sec = 267293.79 rays/sec (RTFM)
Frame 1: 751516 rays in 2.77 sec = 271297.68 rays/sec (RTFM)
Frame 4: 6012128 rays in 21.75 sec = 276379.10 rays/sec (RTFM)
Frame 5: 12024256 rays in 43.50 sec = 276423.57 rays/sec (RTFM)
pixcmp pixels: 262144 matching, 0 off by 1, 0 off by many
bldg391.pix: answers are RIGHT
+++++ m35
Frame 0: 655368 rays in 2.07 sec = 316345.39 rays/sec (RTFM)
Frame 1: 1310736 rays in 4.09 sec = 320639.67 rays/sec (RTFM)
Frame 3: 5242944 rays in 16.02 sec = 327181.69 rays/sec (RTFM)
Frame 4: 10485888 rays in 32.14 sec = 326255.47 rays/sec (RTFM)
pixcmp pixels: 259145 matching, 37 off by 1, 2962 off by many
m35.pix: WRONG WRONG WRONG WRONG WRONG WRONG
+++++ sphflake
Frame 0: 1889266 rays in 4.78 sec = 395428.08 rays/sec (RTFM)
Frame 1: 3778532 rays in 9.08 sec = 416224.06 rays/sec (RTFM)
Frame 2: 7557064 rays in 18.68 sec = 404658.57 rays/sec (RTFM)
Frame 3: 15114128 rays in 36.15 sec = 418042.75 rays/sec (RTFM)
pixcmp pixels: 262144 matching, 0 off by 1, 0 off by many
sphflake.pix: answers are RIGHT
[Code]
But I don't see anything wrong in the code.
On 7/17/13, Cliff Yapp starseeker@users.sf.net wrote:
Related
Patches: #202
here is the patch i generated today