Mario 🔒 on Nostr: when it comes to complexity of a given problem, there are multiple classes (best ...
when it comes to complexity of a given problem, there are multiple classes (best known is probably P and NP, standing for "solveable in (Non-deterministic) Polynomial time").
most common algorithms are either in P or in a less complex class, as computing more complex ones would become really inefficient really fast.
if i have i.e. an matrix multiplication algorithms for sth like lighting computation in a video game we could use a naive implementation with a timecomplexity of Θ(n³) (bc the algorithm has 3 for loops, iterating through the entire input), regardless of the size of the matrices as this would just be a constant factor which doesn't scale as hard as i.e. an additional matrix.
so i think that most if not all computer games are in the same complexity class and therefore should be able to be optimized in a way that all games could have the same performance on a given hardware
most common algorithms are either in P or in a less complex class, as computing more complex ones would become really inefficient really fast.
if i have i.e. an matrix multiplication algorithms for sth like lighting computation in a video game we could use a naive implementation with a timecomplexity of Θ(n³) (bc the algorithm has 3 for loops, iterating through the entire input), regardless of the size of the matrices as this would just be a constant factor which doesn't scale as hard as i.e. an additional matrix.
so i think that most if not all computer games are in the same complexity class and therefore should be able to be optimized in a way that all games could have the same performance on a given hardware