93
edits
Changes
→Single Run Results
[[File:Large-Shay-cuda.png]]
It was my design choice, for reasons of wrapping being able to wrap my head around the logic, to launch 3 two-dimensional grids per image, one per colour channel, instead of preparing a 3 dimensional grid and somehow utilizing the z member. We pass the current iteration (z) to use in calculating the correct location for single dimensional representation of the image:
<nowiki>
