I'm working on voxel terrain which uses a chunk system. One component is scanning the area around the player to determine which new chunks to add as they move. I took the easy way and wrote the code in this structure:
# view_pos varies and can be anything
var chunk_size = 16
var view_dist = 64
var view_pos = Vector3(0, 0, 0)
var mins = Vector3(view_pos.x - view_dist, view_pos.y - view_dist, view_pos.z - view_dist)
var maxs = Vector3(view_pos.x + view_dist, view_pos.y + view_dist, view_pos.z + view_dist)
for x in range(mins.x, maxs.x, chunk_size):
for y in range(mins.y, maxs.y, chunk_size):
for z in range(mins.z, maxs.z, chunk_size):
var pos = Vector3(x, y, z)
if pos.distance_to(view_pos) <= view_dist:
# Add a new chunk here
This works but I want to optimize my loop for performance, given it can induce notable slowdowns with a lot of chunks. As you can see I preform my loop in a cubic area encompassing the view, yet only care to add new chunks if they're within a spherical radius encompassing that cube: Every point that falls within the virtual cube but outside the virtual sphere encompassing it is a wasted iteration.
How do you suggest doing my x / y / z loop, or better off a single loop in a 1D array converted to 3D positions, so that each iteration is automatically a valid point within a sphere?
