I know that the Map-Reduce model is a common model of parallel computation, perhaps some sort of standard way(?). I also know that graphic cards are specifically built for parallel computation. However when reading about the graphic pipeline I am not sure if there is a point at which the Map Reduce model is used. If not, what kind of parallel computations (just the general model of them so I could read more about it) are done during the typical rendering process of modern graphics cards? If it is used, where would that be on the below pipeline?
This model isn't appropriate for rendering games, so it's not used in game engines. However, it can be implemented on GPUs. F.e. if you are going to implement Spark / relational algebra on GPUs, it may be a part of implementation.