[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: 1685128747012.png (1018 KB, 1280x811)
1018 KB
1018 KB PNG
What is the point of using C if it isn't even that fast?
>>
File: 1669877876487290.png (19 KB, 904x564)
19 KB
19 KB PNG
>>93679619
it gets the job done
>>
>>93679673
/thread
>>
>>93679619
Okay, I'm a zigger now
>>
>>93679673
Aren't you tired of posting this image that will never become a meme
>>
>>93679619
this guy is a hack fraud you know,
don't believe anything he says
>>
>>93679778
source?
>>
>>93679673
fpbp
I hate modern software
>>
>>93679619
This retard has a track record of blatantly fabricating information and I don't know what this benchmark is but the idea that the winner would beat C++ by over 6.5x is too absurd to take at face value.
Have you ever wondered why he doesn't work at Microsoft anymore? Hmm?
>>
>>93679619
It's probably because Windows doesn't have Linus' binary blob which accelerates C programs.
>>
>fastest languages
>ignores assembly entirely
what a fraud
>>
File: IFYOUONLYKNEWG.png (459 KB, 535x571)
459 KB
459 KB PNG
>>93679619
>>
>>93679727
My zigger!
>>
File: formerly chuck's garage.png (58 KB, 1285x410)
58 KB
58 KB PNG
>>
>>93680566
meant for >>93679818
>>
>>93679619
Yes, again, this kind of laissez faire view in regards to one's philosophy works only when you have enough egoistic resources. However, when you are forced through ego depletion to find alternate ways of finding those resources and changing to different modes, you will be forced to question your worldview and you will need to have it stable, as otherwise, you won't be able to derive the consistency that facilitates the extraction of those resources.
That is, superficiality works only as long as you worldview is not questioned. However, when you see serious flaws in it, you can't be superficial anymore as otherwise, you are just lying to yourself and creating illusions.
Functions are important. Having only a general vague idea doesn't do anything as long as you don't know how to apply it. In order to apply it, you also need the specifics. The general is enough only in the phase of understanding or accommodating yourself to a concept, but not in the phase of application or integration of those concepts.
And no, intellectual thought isn't try hard as long as your primary goal is to understand. You don't do it for show, but for the goal of understanding.

I also don't think that I was unclear in my message. Even if I was, noone brought any criticism that addressed my point. Instead, I see only the nirvana fallacy in that someone denies something just because it's not perfect or he doesn't understand it so he immediately thinks that it is flawed.
It is not up only to the translator to make the receiver to understand. The latter needs to piece the details and nuances together to understand the idea presented.
That's kinda of a childish view in regards to knowledge.

MBTI is not an excuse to bring excuses towards the way of how you think and act, but it is a tool to be more aware.
Projection I see.
>>
>>93680154
>>93680379
the measurement is based on speed of the programs in practice. he has half a million subscribers and he gave the one year to come up with as efficient as possible solution for calculating primes. There were people using assembly but it wasn't optimized enough to win the challenge.
Video: https://www.youtube.com/watch?v=pSvSXBorw4A
Github repo: https://github.com/PlummersSoftwareLLC/Primes
>>
i was gonna learn c but someone told me i would have to learn how to do data structures and shit on my own and i was like nigga idk how to do that
>>
>>93680643
just use C++ or Java.
>>
>>93679619
This guy has the eyes and slow meandering slur of a major alcoholic
>>
Meanwhile in reality.

// prog.rs
pub fn fib(n: i64) -> i64 {
if n <= 1 {
return n
}
return fib(n-1) + fib(n-2)
}

fn main() {
println!("{}", fib(45))
}


// prog.c
#include <stdio.h>

long long fib(long long n) {
if (n <= 1) {
return n;
}
return fib(n-1) + fib(n-2);
}

int main(void) {
printf("%lld\n", fib(45));
}


$ rustc -O prog.rs -o prog
$ time ./prog
1134903170
./prog 4,23s user 0,00s system 99% cpu 4,267 total
$ gcc -O2 prog.c -o prog
$ time ./prog
1134903170
./prog 2,45s user 0,02s system 99% cpu 2,478 total
>>
>>93679619
The direct party involved might beg to differ. You are only bound to disagree because your incomprehensible way of explaining crap was challenged. Good communicators know how to simplify their message. That is, be definition, a try hard approach to appeal to those who might not be familiar with it (and intimidated by the term INTJ). I see that a lot in younger people who rely on "seeming smart," but has no way of relaying their thoughts effectively. If you can't simplify it, you probably have no idea what you are talking about.

It's amazing how you profess that I did not bring anything into the discussion when there were only two entities involved; me and JambiChik. I don't think I have to elaborate how I pretty much made her day. You, on the other hand, care only about how you are perceived. Always on the defensive when someone challenges your unpolished theories, which were mostly mainsteam regurgitations of the MBTI, not really needing further context nor confusion from your unpolished communication style.

Anecdotal? lol. Okay. Wisdom comes from experience; the data that you work with comes from previous application. You offered nothing but terminologies from the MBTI. Regurgitating crap is usually a sign of no application. You can learn all about MBTI as much as you want, as I did in my early 20s, but if you can't effectively communicate your message, especially if you have nothing but technical terms and dry concepts, you are probably just addicted to consuming information.
>>
>>93680566
That's a pretty retarded thing to say.
>>
>>93679818
He owned a company that was sued for selling scamware.

https://www.atg.wa.gov/news/news-releases/attorney-general-s-office-sues-settles-washington-based-softwareonlinecom
>>
>>93680566
holy based
linuxcels seething
>>
>>93680566
so true. damn i didnt know this dude was so based
>>
>>93680695
>Fibonacci is more real than prime numbers calculator
you can't make this shit up
>>
>>93680746
I used that example because it's easy to translate the code between languages verbatim to get a faithful speed comparison, which is something people clearly don't do for a lot of those programming language benchmarks. There is no way to explain the results otherwise.
>>
>>93680914
A problem with microbenchmarks is that the results are kind of pseudorandom. GCC uses this heuristic when optimizing, LLVM uses that heuristic when optimizing, and in some cases one works better and in other cases the other works better.
In larger programs this averages out, all the random micro-differences add up to similar stats for both compilers. But in small programs it's down to which compiler happens to handle the one teeny-tiny case you wrote better.
If I compile the C code with clang it runs in 4.04 seconds. But in the real world clang is not actually 65% slower overall than GCC.
>it's easy to translate the code between languages verbatim to get a faithful speed comparison
Is verbatim translation actually interesting, though? In the real world different languages are written in different ways. For example C++ programmers use std::string_view much more sparingly than Rust programmers use &str, but they use uninitialized memory more. (Both can give significant speedups.)
You can translate code that uses &str to std::string_view kind of verbatim, and you can translate uninitialized memory in C and C++ to MaybeUninit in Rust with a bit of patience, but that's not how real-world software is written. The performance of real-world software is based on how people write code in reality.
>>
>>93680914
You are free to add optimizations to C prime code to prove us wrong :)
>>
>>93679619
> david plummer
how fucking retarded must you be to take any advice from a conartist and shitware developer that was fined by his own state and forced to shut down his business? this guy is an abysmal failure, and going by this picture..
>>93680511
he is aging badly. hopefully he'll be dead soon.

>>93679766
hello, tranny.
>>
>>93680731
he tried to delete it from wikipedia too but it was all undone and it's permanently there now. he's such a fucking worm. he seriously thought
> oh, nobody will ever know about this. i will open a youtube channel and post on reddit because i used to work for microsoft! people love microsoft, right?
and he got btfo.
>>
File: file.png (141 KB, 1580x458)
141 KB
141 KB PNG
>>93680731
https://en.wikipedia.org/w/index.php?title=Dave_Plummer&action=history
i'm responsible for "added further information of david's lawsuit, with citations and link to press release from washington's attorney general's office".

i hate shitware authors so much that it's fucking unreal. david needs to be shot for scamming people in this way. these are parasites on society's ass that need to be eradicated from the universe. they're just greedy and selfish devils.
>>
>>93679673
Dude how often will you spam this shitty imagine you've been debunked a hunderd times by now, Rust does support UTF-8 perfectly fine.
>>
>>93680643
Just learn Rust. Trust me I've picked it up this year and the developer experience is fantastic.
>>
>>93681512
nobody here on this board has ever expressed any genuine interest in rust except for paid shills and people that don't actually code. go home, reddit.
>>
>>93681252
David didn't code it. It was an open competition
>>
>>93680566
Seriously?
>>
>Con artist
>neo-g sucks his penis
bravo
>>
>>93679619
In this thread: Fizzbuzzers jealous of an elder Autist.
>>
^ failed troll, he shitposted before, not new IP also he failed to comprehend that the boomer didn't code a single line
>>
>>93680710
Is it untrue?
>>
I don't trust lang benchmarks
they are full of vectorization, intrinsics and shit that's outside of the regular/boring programming experience
>>
>>93682593
the first few releases in the early 90s yes, but it was just linus' hobby project then
>>
>They're not. If anything, it's a test of the authors, not the languages, and Zig was given a cheat through the scoring system. It ran on half the hardware threads which gave it turbo frequencies, but then they multiplied the per-thread speed by the number of hardware threads, not cores, meaning it only ran on 16 cores but got a 32 multiplier.
>To determine once and for all ... Yeah, lol, no, not even close. The results are so flawed, I'm baffled that they were presented in the first place. That Zig solution had a lower number of passes than the Rust solution, because it ran on half the threads (1 per physical thread, instead of 2 per thread like the rest) and it got divided by 16 instead of 32. And I don't see the issue of having SIMD enabled for Rust and Zig but not enabled for C and C++ being addressed. There's one thing that this shows very clearly: getting an objective list of the fastest programming language is impossible. Different languages might be better in different scenarios. Also, like it was said, especially if you know the target system and if you REALLY need performance, assembly is the one to go. Zig, Rust, C and C++ all very good choices as well. Java too (was not that surprised to see it there), though for some (me) it's ugly AF, I'd rather struggle with assembly than Java. But for corporate projects Java is very welcomed.

From the comment section.
>>
>>93683059
jesus fucking christ
>>
File: draper_drunk_laughing.jpg (26 KB, 345x336)
26 KB
26 KB JPG
>>93683055
>but it was just linus' hobby project then
It still is.

>>93683059
If it looks too good to be true it probably is.
>>
>>93680566
>Doesn't actually answer the question
Wincels cope. Your OS is a mess bloated with spyware. Unlike Linux.
>Le 3kb binary blob vs the whole OS being close
Your point?
>>
>>93679619
this fat autistic nigger and his bullshit helped make windows a fucking pile of garbage that can't be moved between machines without editing hundreds of registry edits. he should never be allowed near a computer again.
>>
instead of shilling shitlangs why not just get a job and do something productive?
>>
>>93679673
C is the least "gets the job done" language there is. Python "gets the job done", and its arguable that "gets the job done" isn't even a good thing.
>>
>>93681429
as many times as faggots make this same thread each day
>>
>>93681429
>implying "works on my machine" == "deboonked"
>>
File: Fibonacci.png (92 KB, 1088x992)
92 KB
92 KB PNG
>>93681429
>>
>>93684523
I think the problem with the code on the left is that you use 1..=40 instead of 1..41. Inclusive iterators need an extra flag for the exit condition. If you changed this one thing they'd run equally fast for sure
>>
>>93684523
so true
>>
File: funky.png (310 KB, 1974x1393)
310 KB
310 KB PNG
>>93680695
Seems like gcc is doing something funky
https://godbolt.org/z/v5vzY4vdf
>>
>>93679673
fpwp
C doesn't have utf-8 support at all, you need 3rd party (or 1st party of course) code for it.
Just because trannies can't program doesn't change this fact (nor that rust does, in fact, support utf-8 natively).

>>93679619
This is clearly a trick, whatever that is. But C has always been a shitlang only ever used because that's what came with unix, and unix was given out for free to unis. Even up to the early 2000's, it was common to use inline asm anywhere you needed performance, for example. C was neither the first, nor the best, nor the fastest high-level language. Nor was it the first high-level language used for an OS, nor was unix such a first OS (burroughs 5000 and multics were both earlier and both written in high level languages. PL/I, for example, was earlier than C, far more featureful than it, and far faster to boot).
>>
>>93679619
Isnt this the guy who worked on Vista? I would run with a bag over my head for the rest of my life in shame.
>>
>>93684930
gcc here unrolls the fib fully. The fucky implementation is rust, it unrolls, but only on one side, and it does an indirect call (slow) on the non-unrolled side and does a very weird rewrite (check if above 3 instead of check if <= 1, which requires analysis that you'd think is quite sophisticated).
>>
>>93684590
the code is right there, why not run it for us.
>>
>>93684590
The problem is that the one on the left is recursive and the one on the right is O(1) closed form.
>>
>>93679619
Why would I listen to a boomer vlogger for programming advice.
>>
>>93684523
Is there anyone actually retarded enough to get baited by this?
>>
>>93680566
based
>>
>>93679619
I don't know what you're talking about. Nothing can be faster than c. Except hypothetically tachyons.
>>
>>93686829
kek
>>
>>93679619
it's reliable. no other language has been used as long and widely.
>>
>>93680566
What binary blob is he talking about that Linus has access to the source code for?
>>
>>93686897
Lisp has been used longer and more widely by some definition of widely. Also infinitely more reliable.
You could also say fortran in the same veins.
>>
>>93686829
Hand writing it in assembly (if you're good enough).
>>
>>93680566
big if true
>>
>>93686399
Yes, the first time I saw this image on /g/ it got 20 replies.
>>
>>93681410
I applaud your dedication and service anon
>>
>>93686960
Probably the ones on your hard drive
>>
>>93679673
how do you fuck up this badly? is it the language or the devs?
>>
>>93680695
Anyone who writes fib this way, and isn't demonstrating dynamic programming in school should be fired.
>>
>>93684590
No, the problem is that:
1. He used a different algorithm between the two (Rust one is exponential time, C one is constant time)
2. He compiled Rust with optimizations off and compiled C with O3
3. His benchmark includes the time of the Rust compilation but not the C one
>>
>>93679619
>easy
>stable
>popular
>>
File: 1-1comparison.png (118 KB, 910x1910)
118 KB
118 KB PNG
>>93686012
okay
>>
>>93683059
This doesn't explain the gap between C and Rust though.
>>
>>93690876
>the issue of having SIMD enabled for Rust and Zig but not enabled for C and C++
If you see a 3× difference between these languages you should know that trickery is afoot
>>
>>93690949
There is a 6x difference. But even so, how is it the boomers fault? He didn't submit answers, so clearly it's the C programmer's fault (they had one year!)
>>
And so the truth is as it ever was:
>The compiler is all that matters for compiled languages.
You are all silly boys. I hope you are all having fun being silly boys.
>>
>>93690968
Drawing conclusions from that is on him. Not something you would do if you wanted to give an informative evaluation of these languages (but absolutely something you would do if you wanted to produce clickbait)
>>
>>93691033
Only if you're writing toy programs
>>
>>93690876
The order of winners changed just a few days after he published the video if you read the comments, and that for the MT which he presents.
Someone needs to go into the data and compute a new tier but ignoring garbage like Zim.
>>
>>93680612
>he has half a million subscribers and he gave the one year to come up with as efficient as possible
This proves nothing.
Whi cares how many retards are subscribed to rhis cunr?
>>
>>93685696
Posting image since for some reason hiroshimoot will mark this post as spam.
>>
>>93679619
C is comfy and works. C syntax is very simple and very documented. Also, C is still the fastest, go on, zig isn't the fastest, this guy is literally a windows dev, are you trusting him?
>>
>>93679619
>C, C++ and Java have the same speed
opinion discarded
>>
>>93688546
I appreciate the non answer. But next time, if you don't know, just don't bother
>>
>>93693348
nocoder detected
>>
>>93679619
Something is seriously fucked here. What combination of compiler flags and algorithms caused C/C++/Java to be neck and neck while Zig and Rust flew past them?
>>
>>93679619
>dead-inside-eyes' garage
stop posting this faggot "benchmarks", it's garbage ragebait to generate trafic as everyone and their mom will spend time to prove him wrong.
>>
>>93696657
He is literally the only programmer on youtube which doesn't have a faggot city dweller voice while also being legible
>>
>>93693348
You had one year to post more efficient C solution
>>
>>93696781
And it's still braindead bait, so if that's the best you've got then just ignore all youtube programmers
That's what I do, works out great
>>
>>93696781
does not change the fact that his benchmarks are horse-shit.
i'll never forget the /g/ saga on the prime sieve about 2-3 years ago
>>
>>93696781
Dumbest shillpost itt
>>
>>93680566
seems like dave is just like the average poster on 4chan
>>
>>93691299

Statistically, if you have a group of half a million people who have a shared interest in programming videos, a decent sized handful of them will be knowledgeable about assembly language. Among those, at least a few will be good. While the theoretical fastest solution to any computing problem will always be found within assembly, if it can't be produced by any of half a million people, it's not an ideal language choice for developing fast applications. There are only so many unicorns capable of writing high performance assembly programs with no bugs.
>>
>>93679619
Unlike Rust, Zig actually gives you real benefits over C.
>>
>>93696836
What was it?
>>
>>93680612
So they aren't running the same algorithm? lmao, fuck off.
>>
>>93697166
No. But that means C devs are incompetent
>>
File: file.png (215 KB, 1002x677)
215 KB
215 KB PNG
>>93697001
I propose that 100% of people can write high-performance, bugfree assembly code, and that the core problem is that the development process we use is fucked.
>>
>>93680379
Assembly is a low level machine language, it's not really comparable to the rest. You can write an application in it, but it's going to be like cutting down a tree with a chefs knife
>>
>>93697387
Only people who've never used any kind of aassembly could possibly say something like that.
>>
>using math operations to test the speed of languages
retarded
memory performance is what matters

The Zig version was the "fastest" becaus the metric he used was per-thread performance in the multicore category. The Zig implementation used comptime to make it easy to adjust every possible parameter, so they had a highly optimized version for every possible metric. The other languages focused on the single categories like single core throughput and total multicore throughput.
>>
>>93679619
What the fuck is zig?
>>
>>93697444
>zig made it easy for someone to do something that worked, it's not fair :(
>>
>>93697026
one redditor named bruce went on a crusade to prove that rust was the absolute fastest language so he cloned dave garage repo and added garbage implementations of the prime sieve in C++, C# and others then went on here to shill his garbage repo.
there was no rust implementation, as he was (still is if still alive) a massive moron, he wanted someone to write one for him, someone did, and it was slightly slower than the C and C++ version so he went full-denial mode and wrote some rust code containing the "sieve" aka some hardcoded variables and obviously it was incredibly fast as all it did was doing nothing.
another dude and me wrote two compile-time versions of the sieve which ran even faster than the rust implementation, obviously.
he then found a faster than runtime C++ version in rust on the internet, he refused to publish sources as it was obvious that this moron weren't good enough to write it, we got the binary, reversed it and found the og author so we made fun of this brucetard
he then spammed /g/ and reddit for days with innane threads about how rust is godsent and Cnile are morons, yadda yadda.
he then deleted his repo once redditors told him to fuck off then he went back to shill for gnome and wayland, which is what he was known for beforehand.
>>
>>93697337
I don't want to draw graphs
>>
File: dumbfuckjuice.jpg (226 KB, 563x651)
226 KB
226 KB JPG
>>93680566
kek. That would be the most hilarious GPL violation in the history of GPL.
>>
>>93697712
I argue that you should. If you drew a graph, you would spend less overall time developing your program, AND there would be few to no bugs in them.
>>
>>93697642
That's fucking hilarious.
> he went back to shill for gnome and wayland, which is what he was known for beforehand.
Any relations to drew segfault by any chance?
>>
>>93688632
It's the user.
UTF8 won, everything and everyone agrees except retards who want to talk shit about Rust.
>>
>>93698258 For me it's DEC SIXBIT (6-bit ASCII)
>>
>>93679619
I've seen this middle aged guy for the first time maybe a year ago and now he looks like a grandpa, what the fuck happened?
>>
File: 1543602264592.png (45 KB, 888x390)
45 KB
45 KB PNG
>>93684523
Jokes aside, I implemented the same retarded algorithm in c and it is still faster.

C
#include <stdio.h>

int fibonacci(int n) {
if (n == 0) {
return 0;
}
if (n == 1) {
return 1;
}
return fibonacci(n - 1) + fibonacci(n - 2);
}

int main(void) {
for (int n = 0; n <= 40; n++) {
printf("%d\n", fibonacci(n));
}
return 0;
}

rust
fn fibonacci(n: i32) -> i32 {
match n {
0 => 0,
1 => 1,
_ => fibonacci(n - 1) + fibonacci(n - 2),
}
}

fn main() {
for n in 0..=40 {
println!("{}", fibonacci(n));
}
}
>>
>>93698050
*Drew The Noble
>>
>>93698029
My writing alone is so ugly that I can't read what I wrote 2 days ago
Let alone drawing
>>
>>93698700
That's just GCC vs LLVM, compile the C one with clang and they're equally fast
>>
>>93698775
Drew ignoble
>>
>>93698793
I have had success applying this technique by using graph drawing software instead. I also didn't write code by hand, but instead, once done with the high-level design, I wrote pseudocode blocks that I linked together in a kind of lower-level design graph. At both levels, this allowed me to think about and identify in advance areas that had commonality that wasn't apparent at a glance, which allowed me to greatly reduce system complexity. It also allowed me to structure my thoughts in a way that I could completely separate 'writing the code' and 'thinking about the program'. I vastly credit this aspect for the fact the resulting code was bug-free on the first try (1500 lines, C code, implementing a B-Tree-based bespoke compressing keystore).

Also you might actually not need to be able to read your own drawing, because the act of putting it on paper has an effect on memory, so you'll be able to recall what you wrote down/were thinking/did.

In any case, I encourage everyone to give it a try for any non-trivial piece of code (don't bother if the code is trivial, though, ti's just a waste of time obviously).
>>
>>93698812
The fact that rust only has one implementation and it's llvm is rust's own fault, though.
>>
>>93699091
Hmmmm how do I know what kind of graphs to draw? Say I am making a graphical client for ftp in gtk or swing
>>
>>93699129
There's a bootstrap compiler (mrustc), there's a GCC frontend that's making great strides and has been accepted for upstreaming once it's done, and there are other codegen backends for rustc.
But that aside, what's the performance difference between GCC and LLVM? Not big on real-world programs, right?
>>
>>93679727
Zigger power. We wuz da real jews
>>
>>93681410
i'm pinging ur ip rn
>>
File: Untitled.png (27 KB, 651x372)
27 KB
27 KB PNG
>>93693348
Chads code on windows primarily and only use linux to test ports.
>>
>>93699160
The most appropriate graph style may or may not depend on the person. For me when I use this method, I use basically arbitrary symbols to separate logical units, and I do the separation mostly based on flow (so I represent looks, conditionals and one-off 'computation blocks' and combine these elements to express the overall design, using arrows to show how after A we go to B and so forth). I use an empty rectangle with a number indicator to express parallel computation. When a certain block is too complicated to express inline, I first use one block with the name of the thing that will fit there, then express on a different graph the thing itself.

I thought I had a backup of the graph for the system I described but I guess it's all on my work devices after all.
>>
>>93699180
Is this a joke? It's quite big, gcc can be as much as 30% faster on even pretty normal code. On intel hardware you can go even faster using icc.
>>
>>93699241
t. nocoder
>>
>>93699411
If you want your code to reach the most people possible, you code on windows. Then for the 2% of people that aren't, you use a linux distro to test a port. It's hardly my fault you chose an os that used so little that there's more people using things like freedos or some toy os they made in their basement over yours.
>>
>>93699241
Why would developers write code on the same platform their customers use? Ridiculous.
>>
>>93680566
Oh so that's where it's from.
>>
>>93699508
t. nocoder
>>
>>93699386
I thought it was typically in the 3% range? Obviously worthwhile if both can build your program but not enough to disqualify a language, especially since it'll be drowned out by semantic differences.
I'm sure there are normal-looking microbenchmarks where it's 30%, but microbenchmarks gonna microbenchmark.
>>
>>93699586
>in the 3% range?
For microbenchmark, that's true. Real world programs compound those 3%'s in unexpected ways so you end up with much worse performance sometimes, and somewhat better performance sometimes (very rare but does happen with clang, sometimes of the order of ~5-6% faster than gcc).
>especially since it'll be drowned out by semantic differences.
Yes, and because it was designed post-llvm with the ability to be tuned straight for it (as opposed to the gcc model for instance), it could be faster on llvm than a gcc compiler frontend would be (we'll see). Ultimately you can only judge a "language's speed" by the implementations available for it and anything else is just armchair bullshit because it's not that simple (tm).
>>
>>93699411
>>93699580
Grow up and move out your parent's basement. And take a shower.
>>
>>93699645
>Ultimately you can only judge a "language's speed" by the implementations available for it and anything else is just armchair bullshit because it's not that simple (tm).
Oh, I completely agree with that, don't worry. It sometimes comes up in really stupid ways in arguments about Python and Lisp.
But when you see a 60% performance difference and you know for a fact that it's purely down to GCC vs LLVM then you might as well forget about the precise figure and point to that instead. It's important context.
>>
>>93699662
make me
>>
>>93699662
I have a job, car and house. When will you graduate middleschool, kid?
>>
>>93698050
I don't think so also he is a kangaroo that live with his mom
>>
>>93679619
This is all literally irrelevant. The real metric that determines how good a language is depends on just one thing: how easy it is for Sirs to do the needful in the given language. Languages like Python, PHP, C#, Java, Javascript are the top contenders.
>>
>>93702690
That's not goodness, that's popularity.
Goodness is almost exactly anticorrelated with popularity.
>>
>>93682611
>>93681410
>>93681252
obsessed samefag
>>
>>93703707
You have no idea what samefag even means, dumb redditard.
>>>/r/eddit/
Not him by the way. And he's right.
>>
File: 1681397389902624.png (5 KB, 306x119)
5 KB
5 KB PNG
>>93703707
Paranoid schizophrenic...
>>
>>93702727
>Goodness is almost exactly anticorrelated with popularity.
citation needed
>>
>>93703896
source: me
>>
>>93683055
So it no longer has a binary blob?
>>
>>93704021
It does, which is why linux-libre exists. But none that linus has the source for.
>>
>>93686012
>>93690750
damn, the rust tranny shut you up, huh, anon?
>>
>>93703947
Sir, are you an atheist?
>>
>>93703848
yeah bro youre in the kool kidz klub now you left le plebbit with all the normies
>>
File: huegfraud.png (51 KB, 600x600)
51 KB
51 KB PNG
>>93704072
>b-beteefoooh
I'm a no code that was just genuinely curious which is better.
>>
>>93705494
>I was j-just pretending to be retarded
>*posts troon shit*
>>
File: gjgdkg.png (22 KB, 573x508)
22 KB
22 KB PNG
>>93705565
>I'm retarded
>so you where merely pretending
anon congratulations, you are dumber than me.
I recommend playing undertale.
seems right up both our alleys.
>>
>>93707012
dial 8, don't die l8.
>>
>>93679619
i dont know about what you're talking about OP
but this jackass made a DIY light show device and promised he was going to make a followup video on how to code the light sequences and never did

fucking tease, fuck this guy honestly



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.