What is the idea behind thinking about the two differences? Is it similar to the idea of using several for loops vs using one big (nested?) for loop?
Staffi-geng
@grafour one big difference between these two implementations is how much memory we need; if we supersample multiple triangles onto a high-res grid, then we have to keep all these values in memory before performing all the averaging calculations at the end.
Alescontrela
If you aren't memory-constrained I image it's faster to supersample all N triangles onto a high-res grid, since this operation can be parallelized.
What is the idea behind thinking about the two differences? Is it similar to the idea of using several for loops vs using one big (nested?) for loop?
@grafour one big difference between these two implementations is how much memory we need; if we supersample multiple triangles onto a high-res grid, then we have to keep all these values in memory before performing all the averaging calculations at the end.
If you aren't memory-constrained I image it's faster to supersample all N triangles onto a high-res grid, since this operation can be parallelized.