Yes, Paul: equals is the image output. ...
and the kd-tree should be equal, i summed up for each node the depth and
also for the leaves. everything is equal.
/but the perf not
... i working with float for all values. i don't understand why it's not at
all equal the nbr of tests. but the performance difference isn't not only
that, because the difference of the nbr of test done and the performance
difference are not similar.
*****************************************************************************
Visual Studio 2003
*****************************************************************************
KD-Tree...[23348:20]
KD-Tree Build Time = 0.248399s
+ ObjList = 23348
+ Nodes = 14139
+ depth sum = 13724
+ Leaves = 8049
+ depth sum = 7657
Intersection Time = 0.000254502s
Rays = 1
Nodes Checks = 79
Bounding Checks = 122
Triangles Checks = 49
Linehits = 6
Intersection Time <osg> = 0.0215715s
Linehits = 6
SpeedUp = 84.7596
Ray Trace Time = 3.94728s
Rays = 786152 (199K/s)
Nodes Checks = 24806690
Bounding Checks = 49252143
Triangles Checks = 16609007
Viewer...
Scaling image from (600,800) to (512,1024)
*****************************************************************************
Visual Studio 2005
*****************************************************************************
KD-Tree...[23348:20]
KD-Tree Build Time = 0.256906s
+ ObjList = 23348
+ Nodes = 14139
+ depth sum = 13724
+ Leaves = 8049
+ depth sum = 7657
Intersection Time = 0.000297803s
Rays = 1
Nodes Checks = 79
Bounding Checks = 122
Triangles Checks = 49
Linehits = 6
Intersection Time <osg> = 0.0203945s
Linehits = 6
SpeedUp = 68.4831
Ray Trace Time = 4.59982s
Rays = 786152 (170K/s)
Nodes Checks = 24806705
Bounding Checks = 49252212
Triangles Checks = 16608961
Viewer...
Scaling image from (600,800) to (512,1024)
2007/10/15, Paul Melis <[EMAIL PROTECTED]>:
>
> Adrian Egli wrote:
>
> > Hi all,
> >
> > same code, allways default project settings, release, same CPU/RAM/...
> > windows native and os/x native dual boot....
> >
> > what you think about this !!!! results are equal, but computing time
> ....
>
> With "results" I assume you mean the image output? As the number of tree
> nodes checked and triangles intersected is obviously not equal. Which
> should explain at least some of the differences. Are the Kd-trees equal
> among the different platforms?
yes i still check out the issue, and i don't understand why the kd-trees may
is
Paul
>
> >
> > ************************************************************
> > MAC OS/X XCode
> > ************************************************************
> > Ray Trace Time = 2.97638s
> >
> > Rays = 787307 (264K Rays/s)
> > Nodes Checks = 24939199
> > Bounding Checks = 49512447
> > Triangles Checks = 16703960
> >
> >
> > ************************************************************
> > Visual Studio 2003
> > ************************************************************
> > Ray Trace Time = 3.77131s
> >
> > Rays = 786152 (208K/s)
> > Nodes Checks = 24806690
> > Bounding Checks = 49252143
> > Triangles Checks = 16609007
> >
> > ************************************************************
> > Visual Studio 2005
> > ************************************************************
> >
> > Ray Trace Time = 4.62109s
> >
> > Rays = 786152 (170K/s)
> > Nodes Checks = 24806705
> > Bounding Checks = 49252212
> > Triangles Checks = 16608961
> > --
> > ********************************************
> > Adrian Egli
> >
> >------------------------------------------------------------------------
> >
> >_______________________________________________
> >osg-users mailing list
> >[email protected]
> >http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
> >
> >
>
> _______________________________________________
> osg-users mailing list
> [email protected]
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
--
********************************************
Adrian Egli
_______________________________________________
osg-users mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
