Hi Robert,
Have you had a chance to look at this uniform buffer object patch? I
can take another go at it if other changes to uniforms are making it
hard to reason about this patch and its effects. However, I think it
is a big improvement.

Tim

On Sat, Jul 14, 2012 at 10:03 AM, Tim Moore <[email protected]> wrote:
> Hi,
> This submission changes the treatment of uniforms, uniform blocks, and
> uniform buffer objects to provide a lot more flexibility. All the
> information about a uniform (offset, array stride, etc.) is now
> available to the user program. Blocks don't have to be assigned a
> binding point at shader program link time anymore, so code can
> postpone the decision about what binding points to use until it knows
> what blocks are actually used in the shader program. You are no longer
> restricted to std140 layout for uniform blocks, but you can still use
> it and interrogate OSG for the offsets if you don't want to deal with
> the std140 alignment rules.
>
> Since the uniform data isn't available until the shader program is
> linked, often in another thread, a mechanism is provided to test if
> the data is available and request that the program be linked if it
> isn't. This implies that the program doesn't have to be in a state set
> in the scene graph until its offsets have been determined.
>
> This is a fairly big patch IMHO, so in addition to the files I've
> provided a sequence of patches that implement the new functionality
> step by step.
>
> Tim
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to