WTF

WTF.

Rust beats C++ on the shootout now?

Other urls found in this thread:

benchmarksgame.alioth.debian.org/u64q/compare.php?lang=rust&lang2=gpp
benchmarksgame.alioth.debian.org/u64q/program.php?test=mandelbrot&lang=rust&id=3
benchmarksgame.alioth.debian.org/u64q/which-programs-are-fastest.html
github.com/opal/opal/issues/941.
benchmarksgame.alioth.debian.org/u64q/go.html
twitter.com/NSFWRedditImage

...ok, found out why. It has the fastest performance on the regex test by far because it has macros that create optimized regex code at compile time. That pulls down the geometric mean used in the test.

benchmarksgame.alioth.debian.org/u64q/compare.php?lang=rust&lang2=gpp

Wtf so rustsjws are cheating

Java beats C# in almost all tests
yet Java gets way more hate for being "slow".

Rustcucks have take over the shootout and are gaming the benchmarks to trick the masses. They won't accept improved versions of C or C++ benchmarks that out-perform Rust. Many of the C/C++ versions are crippled, using shitty third-party libraries, bad algorithms, etc. that just don't compare to what the tricks they're using for Rust.

No. You just don't know how to read the chart.

Everybody is cheating at that thing. Java is doing great at the k-nucleotides benchmark because they were allowed to use a specialized hashmap instead of the one in the standard library.

Those tests are inaccurate. If you open any of C# examples you'll find out they are using classes instead of structs, using foreach loops instead of for loops, and plenty other shit I cant think off the top of my head.

...

I can't tell if that's really perl7 or if it is a joke and it's rust

i like csharp

me too, do you have skype =) ?

>guise its totally gonna replace C with something memory safe!
unless and until this gap is it never will.

kek

That's not rust or perl

Why don't you put your money where your mouth is and submit faster solutions?

it's rust
benchmarksgame.alioth.debian.org/u64q/program.php?test=mandelbrot&lang=rust&id=3

>java
>C#
C++ is better.

Rust managed to get there in such a short time. I have no doubt that it'll eventually be as fast as C.

This is so incompetent that I can't even

I hate rust wtf Go is for God

protip: less than one order of magnitude does not matter

is f# really that slow?

>that idiomatic rust code

macros should have never let out of lisp.

source?

why would I? To show retarded nigger cattle that C# is faster than Java? rofl, if you base your language choice based on benchmarks performed by some random then you deserve to be retarded

also why I don't get why people want templateHaskell to succeed

Yes and no. Java has JIT compilation and should perform well on this text, but doesn't.

Also, Rust is still faster on every benchmark than Go, Java, or C#.

C; C++, and Fortran on array code, are the only languages that don't get left in the dust by the rust.

Literally nobody uses arrays

>Rust
>2017
>using a SJW language

for certain things.

Language arguments are stupid. You use whatever makes your task easiest or whatever the company you work at uses.

(Serious interest) Why not?

You do if you use Fortran. Good support for fast parallelized operations on arrays is literally the reason why you use it. That's why it does the best on the N-body benchmark.

>i dont know how to read a box-and-wisker plot
wew

Quite sure they run their tests on Mono which is fucking dumb. I still expect Java to be faster overall.

You're dumb.

It seems weird there's such a gap between C and C++. You'd think the C++ people would just copy the C code and call it theirs.

Also, what a shitty box plot. If everything's between 0-5x the baseline, what's the point of a shitty logarithmic scale that cuts off the bottom of the results.

>0 times the Base time

>between

The point is there are points below 1x that are cut off for no good reason, not that anything runs in no time.

Disgusting

Some people value security and productivity more than raw performance in some benchmarks game.

If you want performance you use arrays.

Just look at the raw numbers then.

Arrays are the fastest collection because of how CPU cache lines works. Please stop posting and go work on your JS/Java/C++ shit or you can kill yourself retard

That was my reaction as well. Looking at this thread most of Sup Forums doesn't either.

Rusted many autist's jimmies

iirc special processing for literal pre-fixes which just happens to match that particular problem.

Not instead-of.
As-well-as the one in the standard library.

>autist
Wow, ableist much?

Not Mono.
Dot Net Core.

my coworkers still believe PHP faster than java because awt/swing

Static analysis tools provide both of them. Rust is literally only reinventing the wheel with a bloated syntax

program time / fastest program time always >= 1

My bad then, I got confused by some other benchmark

.net core is actually even worse than mono though.

This is 1 of 4 charts shown with the same scale. The log scale is needed for Python programs.

Sure, why not. 1 second, 8 seconds, what's the difference? Totally not noticeable.

It's a macro for rust which defines for_vec with 2 possible forms:
- for_vec!(in_each [a, b, ...] do i fn j)
- for_vec!(i fn j)
The first form repeatedly runs the infix 'fn' between 'i' and 'j' for each element in '[a, b, ...]', returning nothing (the macro is completely wrong or written that way on purpose).
The second form calls the first form with '[a, b, ...]' being [0, 1, ..., 7].

Those programs were faster run with Dot Net Core than run with Mono.

In practice, rust programs like ripgrep, colenchyma and leaf handily beat the C equivalent in terms of speed. Benchmark speed ultimately doesn't matter much, even more so in the case of benchmarkgame which is awful in so many ways.

There are no good static analysis tools and even if there were rust does not let shit programs undermine good programmers by allowing them to use the code in production which is objectively a good thing.

The proof that there are no good C static code analysis tools is simple: linux exists.

Not instead-of.
As-well-as the one in the standard library.

So there must be something peculiar about the implementation that shits on mono for no good reason.

SJW scale is from highest-to-lowest:
- Ruby
- JS
- Rust
- Python
- Java
- C#
- D
- C++
- C

So you say.

your scale is broken

Show me the correct one

Highest: Rust, JS
Mid: rest
Lowest: HolyC

Ok, that seems reasonable

>muh arraylist is not array, it's a list

Can sjws even into systems programming or are all Rust ones """"""'software evangelists""""""" I mean Rust is pretty hard to get into.

nah, but they are probably most likely to focus on "codes of conducts" and community bullshit like that. All rust coders are mozilla cucks that listen to electro-swing and has pink hair.

no versions
dat unecessarily log/scaled y-axis
no reference code
explain to us again what the c++ vs rust means in that chart again and why they are in that order?

>[citation needed]
>have a (You)

I sure don't care about who is using a certain language.

Because I don't think like an SJW, you see.

why doesn't rust die already?
it's even more verbose and shitty than c++

CAN'T BUST THE RUST!

what kinds of pics you save?

Everyone knows ctards are retarded but even you're special amongst them.

what...the...hell...was...that?!

cd /food/burger
cat cow.asm

> JS
> SJW

Bullshit. Substantiate your claims.

not an explanation but sure

...

benchmarksgame.alioth.debian.org/u64q/which-programs-are-fastest.html

because geometric mean

if you had bothered to read the article you would know

"You can see that the order would be different if it was based on the median scores instead of the [pdf] geometric mean scores. "

also clearly stated

"Measurement is not prophesy"
The performance of a benchmark, even if it is derived from a real program, may not help to predict the performance of similar programs that have different hot spots.

>boy Sup Forums likes to run with stuff while that stuff warns against it

Looks like that cow had a prolapsed arse.

- Rust
- JS
- Ruby (US)
- Python
- Haskell
- Clojure
- Java
- C#
- Perl
- CL
- Ruby (JP)
- OCaml
- D
- C++
- C
- Tcl/Tk
- C+=
- Hoon
- Forth
- HolyC
- Nock

Weird how on average those languages get worse as you go down the list

>Weird how on average those languages get less hip as you go down the list

Check the reactions to that discs thing. Also it'd web, web is an sjw nest.

>that discs thing
What?

I'm gonna become a Rust commiter and sneak in master/slave and he/she terminology into the comments. Take that, you bunch of pearl clutching hand-wringers.

Won't work. They'll catch it in code review or find it later and dox you.

Why Ruby though? It's a nice language, and I whish I could use "lmao everything is and object" more at work than "lmao quadruple indentation" Python. Ruby code is nice to read and easy to maintain, and just as fast as Python or JS.

Find the dicss github repo to see what I mean (assuming it still exists)

>Why Ruby though?
Take a gander at this classic: github.com/opal/opal/issues/941. It also shows the benefits of anonymity.

> (OP)
Go beats Java at literally everything that matters

benchmarksgame.alioth.debian.org/u64q/go.html

go rollerblading

Top kek