In game dev, you sometimes avoid holding on to pointers for more than a single frame or operation because of this. Especially since many things get created or destroyed over time, and we want to keep our arrays densely packed to reduce cache misses.
We are trying a database like approach for retrieving references to things in our current project. It adds a small performance penalty, but it’s far from a bottleneck and it’s been nice to work with
Did you ever work on a game development project where you would have rather done what you did versus using Unity?
Would you want to work in a way where performance didn't matter?
How could you tell the difference between the following two scenarios:
- A specific kind of higher performance was important.
- Higher performance wasn't important: you'd get the same reviews, sell the same copies, whatever graphics improvements you got didn't matter, and maybe you even finish your game 2x sooner because you weren't writing an engine, all else being equal. But because it was intellectually stimulating to you to deal with performance, you committed code every day, which is better than nothing.
Yes. Sorry for the low brow comment, but fuck unity. I really dislike the fact that if someone wants to become a gamedev in 2023, their only practical options are to write everything themselves, learn unity, or learn unreal engine. I’ve been putting together an alternative. https://github.com/shawwn/noh
It’s frankly amazing to work with a production grade engine that compiles in three minutes. I wouldn’t trade it for all the complexity in unity, regardless of how many extra copies I’d sell. But I realize I’m in the minority.
I think it matters to love the tech stack you use, and gamedev has gone out of fashion mostly because no one loves any of the stacks. They deal with them, which is different.
We are trying a database like approach for retrieving references to things in our current project. It adds a small performance penalty, but it’s far from a bottleneck and it’s been nice to work with