My first experience with Lemmy was thinking that the UI was beautiful, and lemmy.ml (the first instance I looked at) was asking people not to join because they already had 1500 users and were struggling to scale.
1500 users just doesn’t seem like much, it seems like the type of load you could handle with a Raspberry Pi in a dusty corner.
Are the Lemmy servers struggling to scale because of the federation process / protocols?
Maybe I underestimate how much compute goes into hosting user generated content? Users generate very little text, but uploading pictures takes more space. Users are generating millions of bytes of content and it’s overloading computers that can handle billions of bytes with ease, what happened? Am I missing something here?
Or maybe the code is just inefficient?
Which brings me to the title’s question: Does Lemmy benefit from using Rust? None of the problems I can imagine are related to code execution speed.
If the federation process and protocols are inefficient, then everything is being built on sand. Popular protocols are hard to change. How often does the HTTP protocol change? Never. The language used for the code doesn’t matter in this case.
If the code is just inefficient, well, inefficient Rust is probably slower than efficient Python or JavaScript. Could the complexity of Rust have pushed the devs towards a simpler but less efficient solution that ends up being slower than garbage collected languages? I’m sure this has happened before, but I don’t know anything about the Lemmy code.
Or, again, maybe I’m just underestimating the amount of compute required to support 1500 users sharing a little bit of text and a few images?
Rust is not to blame, but that code that has been written in Rust might be to blame.
The algorithms used have more effect than the language used, and Rust might make using certain algorithms more painful and thus steer programming towards less efficient algorithms. (This is true in general, but I don’t know if this is what has happened with Lemmy.)
Look at salvo [diesel] coming it at #200+ on this benchmark1, lots of programming languages have at least one framework that is faster on the microbenchmark. This isn’t especially meaningful, but it does show that, let’s say, a feature rich framework in Rust might end up being slower than a Python framework that’s laser focused on the specific use case.
There’s a catch here, something I read someone mention on Hacker News and I agree. Python is easy when you don’t care about performance; the moment you need to worry about it, all the easiness gets thrown away.
Everything is easy when you don’t care about performance.
Have you ever used py-spy? It’s an excellent profiler for Python code (written in Rust 😉). It can attach to a running process and tell you what line is taking the most time. Seems pretty easy to me. (Which is not to say Python can achieve optimal C speed.)
I don’t think there’s such an easy profiling tool for C or Rust? But I’d be happy to be proven wrong here.
Go solve 20 or 30 Project Euler problems. All of them are solvable in less than a second using Python (or any language). Write your solutions in C or Rust and you will soon see that a naive or brute-force solution in Rust will literally never finish (the heat death of the universe will come first), but a clever and efficient solution in Python takes less than a second.
This is why I say algorithms matter more than language. There’s like 2 or 3 orders of magnitude to be had by choosing the fastest language (which is to say, Rust might be 1000 times faster than Python in some cases), but there’s like 10 or 20 orders of magnitude to be saved using the right algorithms sometimes.