10

Go's a few times slower than C++ or Rust according to this benchmark. But not compared to shit shitshow of a language.

https://benchmarksgame-team.pages.debian.net/...

Comments
  • 11
    Ah yes, bechmarksgame, the site that doesnt measure AFTER the jvm spinup cycle 👍
  • 8
    If you compare a compiled language you should also mention the time it takes to compile. Although that will probably still be less than 200s.

    Also with the right libraries python would probably be a lot faster.
  • 5
    @FrodoSwaggins dont care didnt ask plus you are white
  • 3
    @Lythenas Sure that's true. But even Node.js outperformes Python
  • 5
    @yellow-dog JVM has overhead of its own that I find disturbing enough
  • 2
    @ScriptCoded yeah i can see that
  • 0
    Is this with numpy?
  • 7
    Not that the benchmark makes sense because you just don't use Python directly for heavy computation. Instead, you identify these maybe 10% of the program that account for 90% of the running time and do that e.g. in C.
  • 2
    That site is utter garbage. Its benchmarks are biased.
  • 1
    @yellow-dog I'm talking about it's resource usage. Sorry for being unclear.

    @Fast-Nop @Charon92 Yeah sure. I get it. But still people use it for production servers. Not just for training their AI models or whatever.

    @kescherRant Yeah seems like that. Now I know. Sorry
  • 2
    @ScriptCoded Why would a production server use heavy math computation like Mandelbrot?
  • 1
    A few facts to benchmarks:

    - Time and / or resource usage might not matter at all, the use case does
    - the use case is also important for the question of distribution (eg compiling - optimize for specific processor - not mass distributable if not all machines are the same)

    - a Benchmark without detailed listing of the environment, steps taken and all variables in eg configuration is garbage
    -- C programs can be slow as fuck when you simply change the optimization to an I386 processor ;)
    -- Cold vs Hot
    -- How many times? AVG / Median / Std deviation?

    ... And so on.

    Just by looking at the side I see a pile of crap trying to look smart.
  • 1
    @ScriptCoded Instagram are running Django on their servers, and it’s not even the fastest web framework in Python 🤷‍♂️ There are some early stage experimental asynchronous Python frameworks that are faster than Node.js as well.
  • 0
    @Lythenas should you also factor in the time it takes to install the jvm?

    Thats just bs. Compile time should be considered if you’re comparing development speed, not runtime speed.
  • 0
    This brings back memories i remember there was a huge thread like this when devrant just started!!!

    .NetCore FTW
  • 1
    @FrodoSwaggins i recommend you read this post by yours truly https://razorsh4rk.github.io/rant/...
  • 0
    @FrodoSwaggins Yep
  • 0
    How does pypy perform?
    Also you save a lot of development time using python, which can be used to optimize your data structures and algorithms. That is probably a bigger factor than the language.
  • 2
    <sarcasm>I consider such things as much because I am always writing some Netflix level code in which speed of execution from my backend needs to be as speedy as possible, aint not latency problems from my side no siree, and my clients? optimized to the core. JS is for putos and I do as little css trickery as possible. I use the fastest shit possible</sarcasm>

    All jokes aside, how much of this matters? even on enterprise level development (in which my soul gets crushed on the daily thank you very much) the great majority of applications you could possibly write on the most basic of lamp stacks(just an example) would suffice.

    Plus, spin up a rest api in go vs doing it in the "much slower" Django Rest Framework or heck even laravel, chances are it won't matter as long as best practices and fast delivery of the end product happens.

    Need speed? ain't shit beating C or C++, but I venture to say that for most anything will do.
  • 2
    @AleCx04 and for my next trick I will see how much of what I said above is true on a TCL based web app I will build cuz I hate life and everyone surrounding me.
  • 2
    @AleCx04 update on the TCL web based app: fuck that
  • 0
    @AleCx04 i actually have to write speed critical code, so far i have been able to outperform my coworkers' rust code with scala so 🤷‍♂️🤷‍♂️
  • 0
    @yellow-dog we all have to write speed critical code, my point is that it is not netflix/fb/google speed critical you get my point
Add Comment