Originally posted by JimO'Connor
I think I have a workable optimization idea for the debris to cut its computational cost by about 70%. Goldfish shipped, but that doesn't mean we're done with it. Same for MA and MAT, by the way.
Cool
While you're at it, can you take half the debris particles and let one thread of excution handle them, while a second thread handles the remaining particles? Maybe the left and right sides of the tank is possibly the way to group them.
This implementation would take better advantage of my dual CPUs and I like that
Almost as much as my daily glass of beer