I'm working on a raycasting system that involves a 31x31x31 grid. For each of the 5,766 directions, I check 15 points in that grid to determine if any of those points are solid. When I find a solid point, I mark all subsequent points as shadowed. While each ray operates efficiently in nanoseconds with some optimizations, running this process 5,766 times really adds up, especially when I need multiple rays. I've been looking for alternatives to this brute-force approach but haven't had much luck. Any suggestions?
2 Answers
It seems like you're building a line of sight system! One thing to think about is whether you need to cast rays from every cell simultaneously. For instance, if you have three cells in a row where the middle one is solid, you might end up shadowing the surrounding cells unnecessarily. Instead, consider checking rays more selectively based on the grid layout before executing full checks.
Have you thought about using colliders to cover larger areas? If you can define big chunks as solid, you could skip the raycasting altogether if there's no collision detected. That could significantly cut down on the number of checks you need to perform.
What do you mean by colliders? Are you suggesting to create zones instead of checking every ray?
Exactly! If the middle cell is solid, you might only need to handle it once rather than checking all nearby cells. A clever pre-check could save you time!