Hi Robert/all,
Sorry for the hijack....
I appear to be missing posts on the mailing list?
Never got Roberts reply on this post, I have noticed this happening for a
few weeks with intermittent posts not coming through to the mailing list,
mostly Roberts replies but  I don’t think it's just Roberts.
Is there a problem, I have done the obvious like check junk-mail etc..

Regards

Martin.


-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Riccardo
Corsi
Sent: 04 July 2010 23:07
To: OpenSceneGraph Users
Subject: Re: [osg-users] StateSet sharing for max performance

Hi Robert,

I actually tested the 2 solutions and I can confirm there's only a
negligible difference in the cull traversal time, drawing time is just
the same.
So far solution 2 fits much better with my application, so I can pay
the little overhead.
Thanks,
Ricky

On Thu, Jul 1, 2010 at 17:01, Robert Osfield <[email protected]>
wrote:
> Hi Ricky,
>
> The cull performance will be very slightly higher for method 2 as the
> cull traversal will have to push/pop more state, but the draw
> traversal will be identical between the two.  There will be so little
> difference in measurable performance I would suggest going for what
> makes most sense for your application.
>
> Of course one should test this... it'd be easy to benchmark the two
> configuration to make sure that there isn't a measurable difference.
>
> Robert.
>
> On Thu, Jul 1, 2010 at 3:57 PM, Riccardo Corsi
> <[email protected]> wrote:
>> Hi All,
>>
>> in my scene I have many drawables that share the same settings/glsl
>> program - the needed params to render the different drawables are
>> already passed in as vertex attributes at geometry level.
>> I'd like to know it there's any difference, with respect to
>> performances, if I share the common StateSet either one of the
>> following methods:
>>
>> 1. share the common SS in a single geode
>>
>> geode1 - SS1
>>   |
>>   |--------------------------------
>>  |              |               |
>> draw1      draw2        draw3
>>
>>
>> geode1 - SS2
>>  |
>>  |--------------------------------
>>  |              |                |
>> draw11      draw12      draw13
>>
>>
>>
>> 2. share the same instance of the SS (let's call them SS1 and SS2) at
>> drawable level:
>>
>> geode
>>  |
>>
 |--------------------------------------------------------------------------
--------------------------------------------
>>  |                       |                        |
>>  |                       |                        |
>> draw1--SS1      draw2--SS1        draw3--SS1      draw11--SS2
>> draw12--SS2        draw13--SS2
>>
>>
>> The second method would be preferable to me, as every drawable
>> represents an independent object, and I might need to change the way
>> it's rendered at runtime (see hilight, etc...).
>> I'd like to know if there are some performance penalties with respect
>> to method 1.
>>
>> Thanks a lot.
>> Ricky
>> _______________________________________________
>> 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
>
_______________________________________________
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

Reply via email to