Re:Show HN: Rust Implementation of Conway's Game of Life
by brundolf on Wednesday April 08, 2020 @02:09PM
(#22815412)
Yeah, though my goal with this particular project was to strike a balance between performance and readability, not squeeze out the absolute maximum performance. I.e., I didn't bother with any optimizations that entered the territory of "tricks", where they started to make the code less intelligible. For my use-cases (the configurations included in the repo), the current implementation is not limited by speed (and I didn't even turn on compiler optimizations, which in my experience can give a 20x speed improvement).
[ Parent
| on HN ]
Re:Show HN: Rust Implementation of Conway's Game of Life
by e12e on Wednesday April 08, 2020 @02:41PM
(#22815696)
Oh, I don't mind that at all. All these ideas would need to be be benchmarked anyway :)
But ever since implementing a naive (backed by 2d array) game of life, I from time to time play with the idea of trying for something much higher performance. I'm not sure if this one scales to 30fps at 4k for example (maybe it surpasses that on modern hw!). That's the kind of target that'd be fun to aim for. Quite possibly using the frame buffer for state is more efficient anyway.