Here we can verify the parallel solution reads the initial pixel values and applies the rotation correctly:

It was my design choice, for reasons of being able to wrap my head around the logic, to launch 3 two-dimensional grids per image, one per colour channel, instead of preparing . It was my initial thought to launch a 3 dimensional single grid and somehow utilizing utilize the z memberto mimic 3 dimensions. We I should also try to accomplish this in a single grid to compare the results. Instead, we pass the current iteration (z) to use in calculating the correct location for single dimensional representation of the image: