Page 15 of 15

Re: [GSoC 2012] Off-Screen Particles project - Failed

Posted: Sat Nov 24, 2012 10:23 am
by syedhs
Looks like the project is confirmed dead...

Re: [GSoC 2012] Off-Screen Particles project - continuation

Posted: Wed Nov 28, 2012 11:21 pm
by Karol Badowski 1989
I'd like to send this message after that pronounced update is ready and working. This is the hardest change in project (even if final change effect will be visible only in one place between 2 passes), but most important one, because gives the highest acceleration.

After exam session that is comming up in end of january I'll have time for project again. I really hope I'll finally be able to finish replacing current stenciling to hardware one in main project that week. However I really would like to say it after it is really done and working without mistakes, not before it is done.

Configuration of stencil buffer flagues to set is ready, also tets of this is done on entire objects rendering (however here it needs towork only selectively - that is the hard nut to crack). It still needs to be active only in two moments (just before and just after "discard" operation - and this is the thing that most likely needs material swap or render group swap in listener that will recognise specific passes by their identificators).
My biggest concern is whether "discard" operation really will turn out to work (that means - do not activate change of stencil buffer bits for discarded pixels). I can see if it really brings supposed effect after I manage to finish preparing it.

Hardware stenciling works simply in simple OpenGL programs - I am only 85% sure that there would be expected effect here. If not ' Ill continue looking for other solution how to use this technique.

Re: [GSoC 2012] Off-Screen Particles project - continuation

Posted: Wed Jan 23, 2013 9:40 pm
by Karol Badowski 1989
I've started revisioning material (by creating documentation) - to refresh everything and prepare for next update.

Today I've started with creating presentation - I'll translate it to English language version
(this link will be substituted)
: Off Screen Particles: description of rendering order

I'll update this post with new links as new documents are ready.


http://http.developer.nvidia.com/GPUGem ... _ch23.html confirms now clearly that discard operation will help me avoid stencil writes even if PS for specified pixel was accessed (so mentioned solution should be working properly).