Robert Osfield

Making things simpler and more efficient is something I'll always be drawn to like a month to flame. Except in this instance done right we'll see improvements rather than being...

@AnyOldName3 Did you do any performance benchmarking when making these changes? I've been playing with different performance tests today with bigger models I have but the CPU overhead is pretty...

Thanks for the details and model. FYI, I've edited my previous post thanks. This is what the profiler makes of cubes.glb on my Kubuntu 24.04 + Geforce 1650 system: ~~~...

I am still wandering around looking for a coherent path forward on the topic of dynamic state, I have lots of ideas for different parts but nothing that brings everything...

@AnyOldName3 Are there any vsgExamples that relying upon the vsg::SetScissor + vsg::SetViewport push/pop now in vsg::CommandGraph + vsg::SecondaryCommandGraph? I'm wondering about when these might be needed, I'm presuming it's cases...

I have now checked in my first experimental changes built on top of the dynamic-viewport branch: https://github.com/vsg-dev/VulkanSceneGraph/tree/dynamic-viewport2 The first commit promotes GraphicsPipelineState to a StateCommand and sets the slot number...

I've just been searching the VSG codebase for places where the PipelineStates as set and got a whole bunch of hit in the CompileTraversal where there are various methods for...

Thanks for the input. My current inclination is see how much we can pare things down to stuff that is essential for normal usage, and if there is possible usage...

I had a few minutes available so I created a dynamic-viewport3 branch from dynamic-viewport2: https://github.com/vsg-dev/VulkanSceneGraph/tree/dynamic-viewport3 And quickly removed the code in CompileTraversal that passes on the ViewportState on as the...

Taken the scalpel to some CompileManger/CompileTraversal methods to drop the ViewportState entries as I think these are not required. 9d17041fe0a871da90c7023a99b5aa3aacc18b5b