You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think we can leverage change computation to make things faster. Here's the idea
in round N, we have computed the set Delta of sprites/tile that WILL change (between round N and N+1)
in round N+1, we use the changed set of locations (based on Delta) to compute the set of sprites that are PotentiallyAffected and for which we should run rules; sprites outside of PotentiallyAffected do not need to be visit at all
The above requires a spatial index to find the sprites in the 3x3 neighborhood of a given sprite (does it really?).
Once we have paid for computing Delta_N, computing Delta_N+1 should be cheaper.
See Bejeweled. This is due to use of allSprites in too many places (I think).
The text was updated successfully, but these errors were encountered: