Vulkan

Fancy rendering techniques dominates the content of computer graphics blogs. Today, though, I am not going to talk about the drawing itself, but about the "boring" part: the bridge that turns assets: (in my case USD hydra) per-prim scene graph into the flat, GPU-ready arrays my batches want. For the past year that bridge was the ugliest code in the project – three overlapping bookkeeping mechanisms, data processed one prim at a time, and a lock around almost every access.

Recently I finally decided to solve the synchronization system in my Vulkan renderer. It is something that I really want to do for a long time, to me it should be part of the FrameGraph design in the application, where you declare your individual render passes and the resources it uses (read/write). The synchronization system itself handles the resource transitions, where it should change layout from? Whether we need to setup a semaphore etc.