Re: [osg-users] Fwd: Re: Color array optimization

2010-04-12 Thread Vincent Bourdier

Hi,

Le 07/04/2010 20:45, Jason Daly a écrit :
Yes, BIND_PER_PRIMITIVE is bad.  BIND_OVERALL is fine.  I don't know 
of any Optimizer settings that do this, but it shouldn't be hard to 
write one.


Just one last question ...
Why does BIND_PER_PRIMITIVE is wrong ? what about BIND_PER_PRIMITIVESET ?

Thanks.

Regards,
Vincent
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] Fwd: Re: Color array optimization

2010-04-08 Thread Vincent Bourdier

Hi,

Thanks for your answers, I just made it and it runs well.

Regards,
   Vincent.

Le 07/04/2010 20:45, Jason Daly a écrit :

Ulrich Hertlein wrote:

Hi Tim,

On 7/04/10 15:12 , Tim Moore wrote:
   

 osg::Optimizer attempts to optimize for performance, not memory.
 The optimization you're proposing will have a negative impact on
 OpenGL performance.

How do you figure? Vincent would like to replace BIND_PER_VERTEX colors
with a BIND_OVERALL color where possible. That is a very common thing to
do in OSG and is, as I understand it, well supported even in OpenGL 3.0+
and OpenGL ES 2.0.
 


Looking at that again, I may have confused that with BIND_PER_PRIMITIVE.
   


Yes, BIND_PER_PRIMITIVE is bad.  BIND_OVERALL is fine.  I don't know 
of any Optimizer settings that do this, but it shouldn't be hard to 
write one.


--J


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
   
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] Fwd: Re: Color array optimization

2010-04-07 Thread Ulrich Hertlein
Hi Tim,

On 7/04/10 15:12 , Tim Moore wrote:
 osg::Optimizer attempts to optimize for performance, not memory.
 The optimization you're proposing will have a negative impact on
 OpenGL performance.
 
 How do you figure? Vincent would like to replace BIND_PER_VERTEX colors
 with a BIND_OVERALL color where possible. That is a very common thing to
 do in OSG and is, as I understand it, well supported even in OpenGL 3.0+
 and OpenGL ES 2.0.

Looking at that again, I may have confused that with BIND_PER_PRIMITIVE.

/ulrich
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] Fwd: Re: Color array optimization

2010-04-07 Thread Jason Daly

Ulrich Hertlein wrote:

Hi Tim,

On 7/04/10 15:12 , Tim Moore wrote:
  

osg::Optimizer attempts to optimize for performance, not memory.
The optimization you're proposing will have a negative impact on
OpenGL performance.

How do you figure? Vincent would like to replace BIND_PER_VERTEX colors
with a BIND_OVERALL color where possible. That is a very common thing to
do in OSG and is, as I understand it, well supported even in OpenGL 3.0+
and OpenGL ES 2.0.



Looking at that again, I may have confused that with BIND_PER_PRIMITIVE.
  


Yes, BIND_PER_PRIMITIVE is bad.  BIND_OVERALL is fine.  I don't know of 
any Optimizer settings that do this, but it shouldn't be hard to write one.


--J

___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


[osg-users] Fwd: Re: Color array optimization

2010-04-06 Thread Vincent Bourdier


Hi Robert,

The optimization I am currently working on is the good usage of binding.
Some of our code generate color array with a color by vertex, even if
the color is the same for each vertices... keeping only one color using
the binding overall, we can obtain a considerable file size reduction.

Thanks for your answer.

Regards,
   Vincent.

Le 06/04/2010 17:42, Robert Osfield a écrit :

 Hi Vincent,

 On Tue, Apr 6, 2010 at 8:31 AM, Vincent Bourdier
 vincent.bourd...@gmail.com   wrote:


 Looking for optimization I'm currently searching a color Array optimizer.
 Something modifying the color array and its binding when possible, of course
 without modifying the render of the geometry.

 Is there something like that ? I didn't found anything in the
 osgUtil::optimizer documentation.


 There isn't a specific color array optimization.   I'm rather curious
 about what you'd optimize about a colour array.

 Robert.



___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] Fwd: Re: Color array optimization

2010-04-06 Thread Tim Moore
On Wed, Apr 7, 2010 at 3:23 AM, Ulrich Hertlein u.hertl...@sandbox.dewrote:

 Hi Vincent,

 On 7/04/10 2:04 , Vincent Bourdier wrote:
  The optimization I am currently working on is the good usage of binding.
  Some of our code generate color array with a color by vertex, even if
  the color is the same for each vertices... keeping only one color using
  the binding overall, we can obtain a considerable file size reduction.

 osg::Optimizer attempts to optimize for performance, not memory.
 The optimization you're proposing will have a negative impact on OpenGL
 performance.

How do you figure? Vincent would like to replace BIND_PER_VERTEX colors with
a BIND_OVERALL color where possible. That is a very common thing to do in
OSG and is, as I understand it, well supported even in OpenGL 3.0+ and
OpenGL ES 2.0.

Tim
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org