[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] [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

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • You may highlight syntax and preserve whitespace by using [code] tags.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]


>>
It's funny because they are all jeets?

more like tragic
>>
I am confused, I don't get it
>>
>>106736239
uh lots of jeets are CEOs in USA and ministers in UK. Stop pooing down, poo up.
>>
>>106736154
Senior devs do CS101 level programming and pawn off anything more difficult?
>>
>>106736364
isnt google ceo a jeet?
>>
>>106736154
>copy pasting from google
what is this 2005?
>>
File: 1667922488285995.jpg (46 KB, 600x533)
46 KB
46 KB JPG
>>
File: 1759137745421258.png (28 KB, 1536x864)
28 KB
28 KB PNG
C coder looking at python dict()
>>
>>106736239
Alright then big guy, prove your worth.
t. Indian
Challenge me genuinely btw, I am admittedly not very good at programming. I want to see what level I should be to deserve minimum wage, purely as a hypothetical(I'm happy in India, I don't want your jobs, honest).
>>
>>106736199
google it
>>
>>106736662
>>106736430
>>106736299
why did jeet deleet?
>>
>>106737090
The jannies got to him for racism probably.
>>
bump
>>
>>106736154
LOL
>>
>>106736662
I guess he's gone. Anybody else would like to step up? I really want to see all the programming skill and intuition /g/ touts that we Indians lack.
>>
>>106737090
how can she ban HOW CAN SHE BAN
>>
Thank God for the 100k trump visa tariffs
>>
>>106739599
our app frequently restarts, losing all active sessions, please fix it. it's hosted in a kubernetes cluster, i don't know the language, figure it out yourself
>>
>>106739748
I've never done that (I'm fizzbuzzer whose only experience with the web and networks is writing ftp clients and servers in C), but I'm sure I'll figure it out. Send me the .git saar
>>
>>106739802
we hired you to fix it, this is urgent, get to it
>>
>>
File: society-atlas.jpg (74 KB, 726x726)
74 KB
74 KB JPG
>>106739599
i just wrote this.
comments are for my own use so i dont waste time when ill come back to this code. i dont have much experience with intrinsics.

#include <immintrin.h>
#include <stdio.h>

#ifndef __NOUVO_DEFINES_VECTOR_ATOI_H__
#define __NOUVO_DEFINES_VECTOR_ATOI_H__

const __attribute__((aligned(32))) __m256i VECTOR_ATOI_BIAS_VEC = {
(long long)0x3030303030303030ULL,
(long long)0x3030303030303030ULL,
(long long)0x3030303030303030ULL,
(long long)0x3030303030303030ULL
};

const __attribute__((aligned(32))) __m256i VECTOR_ATOI_CHALLENGE_ZERO_VEC = {
(long long)0x0ULL,
(long long)0x0ULL,
(long long)0x0ULL,
(long long)0x0ULL
};


#endif

unsigned int vector_atoi_period(char *str)
{

//substract '0' from all bytes. if anything is less than \0, it means its a separator.
//assuming separators are . (period), space, newline or null termination
__m256i chunk = _mm256_loadu_si256((const __m256i *)str);

__m256i shifted = _mm256_sub_epi8(chunk, VECTOR_ATOI_BIAS_VEC);
//cmpgt is the only available so were checking if 0 is greater than the shifted chunk
__m256i is_digit = _mm256_cmpgt_epi8(VECTOR_ATOI_CHALLENGE_ZERO_VEC, shifted);
unsigned int mask = _mm256_movemask_epi8(is_digit);

printf("vector_atoi_period : period mask : %u\n", mask);
return (_tzcnt_u32(mask));
}

int main(void)
{
char buffer[1000] = "999.abcdefsg";
printf("main : period value : %u\n", vector_atoi_period(buffer));
}


i have no problems with indians btw, im just protecting my honor since im a /g/ poster
>>
>>106739876 cont
this should be THE fastest way to find a non-num char's position in a string, on x86-64, with avx2 and bmi instructions
>>
>>106739896 cont of cont
>non-num char's position
actually, separators that i use in the dataset
whose value is < to '0', ofc.
namely: period, space, or newline (and all whitespaces but these dont appear in my dataset in this context)
>>
File: msft-ai-lmaoo.jpg (129 KB, 828x983)
129 KB
129 KB JPG
oh, its a humor thread?
lets have some humor then
>>
File: inclusive.png (131 KB, 700x1000)
131 KB
131 KB PNG
>>
>>106739916
if you are not checking the length of str, you presumably control your input. then why is your input in such a shitty format?
>>
File: C-aliens.png (136 KB, 640x463)
136 KB
136 KB PNG
>>106740119
human readable, not shitty
it makes it trivial to inspect the integrity of data
>>
File: basedjak-linoos.png (1.25 MB, 2248x1054)
1.25 MB
1.25 MB PNG
>>106740119
also len is gonna be solved with null termination.
0 is smaller than '0' after all
all one has to do is to over-allocate the size of an __m256i (or __m256i - 1 if one includes classical null termination into the size)
>>
File: sudo.png (215 KB, 1600x2000)
215 KB
215 KB PNG
>>106740119
and and, concerning the format, im dling jsons from the web, so originally its in human-readable form anyways
and this programs job is to turn said data into arrays of floats, basically
thats why it has to be fast. maybe not *that* fast, but i decided ill indulge myself
>>
>>106739876
I mean, I asked for a challenge, not a demonstration but I'll still critique it I guess.

First of all, the obvious one that you cannot use this on strings less than 32B (or atleast 32Bytes of valid buffer.) Since you're using loadu, it means you can pass unaligned strings in there which can additionally cause a GP(0) fault since the unaligned load can cross over a page boundary to some page that is not available to you. And extension of the same critique is that if you naively extend it to accept arbitary strings, it cannot test for strings with length that is not a multiple of 32B either, so your 1000B buffer is not a good length choice.

Secondly, there are important symbol characters after 0x30 in ASCII and the logic breaks for extended schemes like ANSI altogether.

Also,
>non-num char's position
This accepts any character in range 0x30 to 0xAF, or 0x30 to 0x7F ASCII, plenty of non-num characters in there. Reminds me of some of the hacky hacks I do, for eg to avoid division, if the domain is constrained enough.

I don't see what was a challenge here though. I'm trying to judge what I need to be good at to deserve minimum wage like that anon said. Also, C intrincs are ugly, holy shit, glad I hadn't tried them before. I'm sticking with assembling objs in nasm.
>>
File: buffers-g.png (180 KB, 865x740)
180 KB
180 KB PNG
>>106740244
>I really want to see all the programming skill and intuition /g/ touts that we Indians lack.
thats what i showed you. i hope.

also i know what you wrote as critique
its alr been addressed here >>106739916
and here >>106740211

idk what challenge.
write a fizzbuzz in branchless only.
thats a fun test of ones skills in c
>>
File: maxresdefault.jpg (55 KB, 1280x720)
55 KB
55 KB JPG
>>106736662
>t. Indian
Your country's average IQ is around 75.
There are too many Indian immigrants in tech, that they statistically cannot all have an average of 100.

There's no point in you proving you, nor as the "champion" of your race.
If you understand this intrinsically, you will find us on the same page.

I feel sorry for the real ones out there, who may have a harder time, with all that's going on.
But most of them... you should know it, or you're likely one of them, I guess.
>>
File: zuck-droids.png (494 KB, 720x693)
494 KB
494 KB PNG
>>106740244
aaand yeah
obv its ascii only.
my charset is ascii only
>>
>>106740189
by opening it in editor.exe? you even had to oversize the buffer because otherwise the algorithm chatgpt spat out would try to read your inefficient int representation outside of it
>>
>>106736534
she retarded tho
>>
>>106740315
its a playground. i just shoved a random number thats sure to be big enough for my needs
>muh inefficient representation
you shouldnt be saying things like that if you dont see how human readable form is easy to verify.

ill tell you, for argument sake and just to wave my humongous dick around:
if data gets corrupted youre gonna see random characters appear in random places
you cant check things this way if your data is all binary
shrimple ass fucking crustaceans
>>
>>106740351
>you cant check things this way if your data is all binary
https://en.wikipedia.org/wiki/Error_correction_code
>>
>>106740422
>reading comprehension : F-
>>
>>106740449
>error correction: editor.exe
>>
>>106740280
I didn't see that though. I do this all the time?

>alr been addressed
Not entirely. The point of contention still remains that the load can be unaligned and cross over a page boundary that isn't served to you by the mmu. It will crash for real strings towards the end of your available contiguous addressing space.

>>106740288
I am genuinely below average in my peers and I'm not even from some big gated institution or anything, most of us were rural randos with okay grades in high school. My CGPA through college was 6.5. All I'm saying is that it's probably not Indians who're bad at programming, I'd say from experience that most likely, it's you genuinely picking up the worst of us through some elaborate coincidence, systemic corruption, or unfathomable conspiracy.


I may be proving my personal skill here, but I'll just be adding an earnest admission here that I'm genuinely below average in skill, knowledge and problem solving here (of people who actually do computers, not people who watch a youtube video and apply for jobs).

That's why I'm always baffled when /g/ posts examples of incredibly stupid Indians, which aren't even the norm here, who somehow get H1Bs and a life abroad (I don't even want it, desu), while we struggle to get domestic jobs here.

>Average iq is 75
The methodology on that paper was shady as fuck and a fuck ton of people interviewed were barely literate and extremely poor (a significant percentage of our population is).

If you believe that someone who has barely attended school till 5th grade, that being years ago in their life, confronted with IQ problems suddenly, will perform well on problems of a kind they've never had to busy their minds with, that's just stupid. So yeah, it's not, or atleast yours would be close too if your country was in a state like ours, given that IQ is precisely that measurement.

>>106740303
Have a "Based" on that one.
>>
>>106740499
>It will crash for real strings towards the end of your available contiguous addressing space.
addressed.
i over-allocate
i have 32 null terminating bytes at the end of my buffer in my actual code
i dont even need thatmuch, but its simpler than analyzing my data and then getting fucked by a one off mistake that will appear once in every 10 million iterations
>page boundary
i dont have the luxury to have my data aligned.
i have a string, and i parse it, thats it.

theres a u after the load in _mm256_loadu_si256
its the unaligned load so there wont be any issues on that front
>>
>>106740499
>it's you genuinely picking up the worst of us through some elaborate coincidence, systemic corruption, or unfathomable conspiracy
Nepotism.

There's the paradox, of wanting to "help third world countries" develop, wherein you should not take their talent away, because the country will develop at a slower pace, if at all and not even worse, with all that brain-drain going on.

There's also a distinction to be made between an individual and a (theoretically) abstract agglomeration of individuals.

I have this thing about me, where I'm an autist, weird, smart, dumb, I don't really know, but I seem to be different, and I have tried to be... anti-racist is maybe the wrong term, I tried to "operate without prejudice," and the results are:
Everyone is factually happier if I "operate with prejudices on" instead of off, seemingly, but let's just say, 99.999% are too... incapable of breaking that fourth wall, to fully understand what I must imply.

I'm saying, Jose wants me to treat him like a Jose, Ranjeet like a Ranjeet, Emily like an Emily, and Bob like Bob, and attempting to treate everyone under some generic minimalist "typedef human being" construct leads to... less good results.

Yes, this is a very elaborate
>instructions unclear
>got my dick stuck somewhere
I need as insurance.

Cricling back around,
>always baffled when /g/ posts examples of incredibly stupid Indians, which aren't even the norm here, who somehow get H1Bs and a life abroad
it seems you too, my friend, have an interest in making out the bullshitters and scammers, and letting merit reign once more... :)
>>
>>106740576
The issue is precisely that is accepts unaligned strings (actually not really, but I'm just making the assumption that if you were to accept only aligned strings, you'd have some mechanism in place to ensure that on a lddq or load instruction for example, or your program will crash enough to warn you), with loadu it's subtle because towards the end of a contiguous addressing space on the page boundary, which will happen rarely, the loadu will generate a GP(0) too.

But yes, if you're overallocating by atleast 31 bytes, then it's fine.

That still feels a little hacky to me to be honest but mitigations are costly in this case so I'm fine with that. Just don't use this to write any function where the size of the string is not known to be in some overallocated bound. It's perfectly fine to slow down for safety in general string processing. Your strings aren't very diverse though. Specialized application, so you're all good.
>>
>>106740646
If there's no brain drain here, then all the brain retained goes to waste and is not utilized at all. So, yeah, unfortunate conditions but the worst part of it is that our name is dragged through the mud with it.

And I don't expect you, or anyone, to treat me as some generic template homo sapiens when you have enough information to distinguish that I'm an Indian. I just would rather that the idea of an "Indian" that people have, align with the reality of being one. Like I can only judge what the average H1B is like with the memes posted here, but I'm not even sure if that's accurate or heavily cherry picked again.

>letting merit reign
I wouldn't call my aim that noble. I just want that skilled people do actually get the bare minimum of opportunities, every meritocratic step over is just a bonus.
>>
>>106740720
alr taken into account.
>hacky
thats the usecase for c ¯\_(ツ)_/¯
>Just don't use this to write any function where the size of the string is not known to be in some overallocated bound.
i was gonna enforce that with a special type but actually fuck it
im the sole user of my code, i dont need to child-proof it.
i do comment it though. ive done my share of rewrites because i completely forgot what goes where after leaving a project for a while
>>
>>106740720
>>106740812 cont
>foregetting shit
and thats the reason for the comments here
the fact that i use a "greater than" check instead of a "less than" confuses the fuck out of me. i mean, comparatively.
i like having very explicit code, and when i cant, i supplement it with comments

sometimes i worke for 14 hours, i can be drunk and/or baked, or be sleep deprived
i want explicit, easy to read code.
and its in that way that i meant "not being experienced with intrinsics"
the instructions are still "tough" to read for me.
its not at an instinctive level like standard c
>>
bump
>>
>>106736199
/g/unior moment
>>
>>106740332
>retarded
yeah it's python
>>
>a pajeet is doing code review in a humor thread
literally incapable of behaving like a normal human being
>>
>>106736534
#include <unordered_map>

Done.
>>
>>106741796
I realised it said C rather than C++ right as I hit send and if I acted quickly I could have cancelled the post, but I decided instead to decrease the quality of this thread and the site as a whole.
>>
>>106739876
how did you learn this? Did you just study the relevant intel manuals or are the any good examples/ guides/ blogs that you learned from? I know the lemire guy has some material, otherwise I have no idea.
>>
>>106740499
>If you believe that someone who has barely attended school till 5th grade, that being years ago in their life, confronted with IQ problems suddenly, will perform well on problems of a kind they've never had to busy their minds with, that's just stupid. So yeah, it's not, or atleast yours would be close too if your country was in a state like ours, given that IQ is precisely that measurement.
i believe that's the point? that's what an "average" means

i cant speak for others but india being so fucked up with so much inequality is one of the main reasons i don't really trust anyone from there. humans are wired to absolutely hate injustice, and societies with vast inequality can only avoid sliding into instability by inventing systems to internally justify the injustice, like the caste system.

if you're indian, you're either going to be downtrodden, disenfranchised and malnourished, or you're going to be someone who is complicit in the aforementioned in remaining so. you don't consider these indians being low iq as all that important to the nation's image, because you don't consider them as real people to begin with. maybe you can prove me wrong but egalitarianism and the the typical indian psyche just don't seem very compatible.
>>
>>106741971
ask chud gpt.
it sucks donkey cock through the ass at programming but its a quite good search engine
its gonna throw semi-working shit at you but at least you will have a starting point
you then have to check what the manuals say
and use your own brain to find the optimal solutions
or at least thats what i did. cant ascertain its the meta, but it works for me
>>
>>106740499
>>106742035
actually i didnt really read your post and had just assumed it to be the typical "but muh caste" drivel so i may have misread the spirit of what you were trying to say
>>
>>106741971
>>106742091 cont
thats what i did in this case
i asked chud gpt for an atoi, but with vectorization
it provided a shitty solution i reworked to remove redundant instructions
then i asked it to refine the code
it speat out kinda useless shit at me even more useless than the previous solution
so i rewrote it to suit me
picrel.
except, as you see, i didnt go with the minmax solution as instructed
>>
i cant stress that enough
the chatbots are basically a search engine
theyre effing brainless
but theyre still correlation machines, so youre gonna get *something* useful
and the output is only as good as your prompts
you have to verify, you have to refine, and to do that you still need at least a decent understanding of the theory behind what youre doing
>>
>>106742149
I see, thanks. Learning by doing it is, then.
>>
>>106742239
always is.
you create much stronger associations when you apply the theory youre learning as you go vs just passively ingesting information.
not only you also activate visual memory and mechanical memory to an extent, improving sheer information retention
but also you work with the properties of the ideas youre dealing with, enabling comprehension of the subject on a deeper level
its also more compact information-wise to just try out something vs have it explained to you in words.
bigger bitrate kind of deal, youre getting information through multiple channels
>>
>>106736154
All of the senior software engineers I've known had beards (like actually not stubble) and where white.
>>
great humor thread
>>
File: apple-energy.jpg (152 KB, 1009x339)
152 KB
152 KB JPG
>>106742366
95% of them turn into general tech discussions anyways
still, have a you and a meme
>>
>>106742413
They added sideloading (also FIY Job's approved sideloading while he was alive. AFAIK it's been Cook who's pushed back.)
>>
>>106736534
>>106741796
search.h
https://linux.die.net/man/3/tsearch
>>
>>106736662
>>106740244
>>106740499
>fellow jeet?
>>
File: fkwwcvpeo4g91.jpg (109 KB, 1280x1083)
109 KB
109 KB JPG
>>106740812
>that's the uscase for c
True enough

>>106740896
You did great. Just keep practicing them, and ideally read chapter 9 through 14 at the very least if you want to work with AVX2
https://cdrdv2.intel.com/v1/dl/getContent/671200
Temporal and Non-Temporal data writes in particular would help you a lot if you go down the path of systems programming more.

The instructions are very obtuse, because there are a lot of things that should be possible with SIMD instructions, so to keep the number of instructions manageable plus some other reasons, the instruction sets have to be "minimized".

One thing you can do is define a _mm256_cmplt_epi8 macro that jist flips the arguments.

I have PCMPLTB as a macro instruction alias in my NASM common include.

>>106741971
Manuals are best. I wouldn't argue against anon's chatgpt methodology, but I don't trust an AI for jackshit and so I certainly will not recommend using it for anything. The manual (linked above) has everything you'll need, including written summaries and explanations on how to use things.

>>106741742
I didn't start that line of discussion but,
I'll take that as a compliment

Have meme
>>
File: 1739403548029185.webm (1.78 MB, 370x360)
1.78 MB
1.78 MB WEBM
>>106742035
The point is that if I flashbang my grandfather, who hasn't studied beyond 8th, or anybody similar, with an IQ quiz, and logical questions that he never had to deal with, he will perform poorly.

/sci/ will cope but IQ tests need a baseline level of familiarity to be worth anything and they can be trained for beyond that.

I'm not saying that those people are not important,- I love all my compatriots, they're very important to me, I'm just saying that the iq comparison with most other countries, especially developed and educated ones, as it being some fundamental, environment independent statistic, is simply not fair.

>>106742035
>>106742101
Yeah lmao. I do not follow at all. I love my countrymen and I could not give less of a damn about their castes. I can appreciate when people use caste to uplift their community, pride in your heritage and whatnot, but if you use it to outcompete or drag someone down, that's abhorrent behaviour.

>>106742658
Hello
>>
>>106739983
Thank you.
>>
>>106741805
>>106741796
fucking retard
shame on you
>>
I can't stop thinking about "Learn to code" written above the gates of Auschwitz from the last thread.
>>
File: images(4).jpg (23 KB, 576x432)
23 KB
23 KB JPG
>[BEGIN PROGRAM: JERRY SEINFIELD ON STAGE]
>JERRY: What's the deal with female Rust programmers?
>[EXEC AUDIENCE LAUGH]
>JERRY: They're not female..
>[EXEC PAUSE FOR EFFECT]
>JERRY: ...and they're not programmers!
>[EXEC AUDIENCE LAUGH]
>JERRY: You know they say that rust is memory safe...
>[EXEC PAUSE FOR EFFECT]
>JERRY: ...yet they seem to have forgotten they are men!
>[ALTER AUDIENCE MODIFY LAUGH OBSCENE]
>[EXEC AUDIENCE LAUGH]
>JERRY: They go on and on about the "Code of Conduct",
>[EXEC JERRY GESTURE DOUBLE QUOTES]
>JERRY: as if we are pirates sailing the high seas, trying to avoid mutiny...
>[EXEC PAUSE FOR EFFECT]
>JERRY: But wasn't it pirates who thought having women on board was bad luck?
>[EXEC AUDIENCE LAUGH]
>JERRY: And their mascot is a crab! A crab!
>JERRY: Well we know where crabs belong!
>JERRY: At the bottom of the sea!
>[EXEC AUDIENCE LAUGH]
>[LOAD SOUND seinfield_bass.wav]
>JERRY: Walk the plank!
>[FREEZE FRAME]
>[PLAY SOUND]
>[END PROGRAM]
>>
>>106744812
/g/ humor, srs bsns
>>
bump
>>
>>106745569
im gona stil this program and come to america dumb white devil
>>
>>106736154
retarded format, but it isn't wrong. I honestly don't know what's going to happen when all the boomers retire and retards need to inherit mloc C++ codebases, especially since said boomer corps pay like shit, have low visibility and their products suck.

I was at a company where I had to babysit people and teach them how to resolve gcc linker errors and other weird shit. This is the number one reason alone that Rust will win since its build system is no-nonsense and doesn't bend over to retarded legacy decisions.
>>
>>106740244
pajeet sama..... I kneel.
>>
>>106742304
>you create much stronger associations when you apply the theory youre learning as you go vs just passively ingesting information.
I fully agree, but at the very beginning I think getting an overview from someone knowledgeable and maybe a collection of tips and tricks is more efficient than learning everything from the ground up from manuals. Else you risk picking up some really bad habit if you have zero guidance, like yandere dev never learning that for loops exist.
You'll still have to put in the work, but an expert can guide you to do it more efficiently.
>>106743803
That manual is useful and contains what I was looking for. thanks.
>I wouldn't argue against anon's chatgpt methodology, but I don't trust an AI for jackshit and so I certainly will not recommend using it for anything.
Same, even if correct I'd be afraid of acquiring yandere-dev tier misconceptions for a topic where there is relatively sparse writing about online. I only use it when I already know enough about a topic to be able to verify it's output.
>>
>>106743803
based. thanks for the book anon
>>
>>106746530
yeah, ofc
if you have guidance its all the better
doesnt exclude experimental(?) learning though
the latter is a learning method, in structured learning or otherwise
and that method can be applied across all disciplines
>>
bump
>>
File: images(3).jpg (18 KB, 463x280)
18 KB
18 KB JPG
>>106745569
>ALTER AUDIENCE MODIFY LAUGH OBSCENE
KEK
>>
>>106736398
>swe is about complex algorithms and being a leet haxor
Retards like you are the reason we have faggy as leetcode exams as an interview step for a job where you center divs. 90% of the job is just clawing requirements out of the retarded Stacy PM and then plugging in apis like Lego blocks.
>>
bump
>>
>>106747083
>>106747784
Why?
>>
>>106746530
>I only use it when I already know enough about a topic to be able to verify it's output.
I'm a bigger luddite personally, but I suppose that's fair.

>>106746548
Yw.
After you've made yourself familiar with the ISA, you can also look at the Optimization manual as an additional resource

https://www.intel.com/content/www/us/en/content-details/821612/intel-64-and-ia-32-architectures-optimization-reference-manual-volume-1.html

There's also an intrisics guide, if you plan on writing code in C often, which I can't attest to the quality of, because I've never used it, but the ISA documentation and Optimization manual from Intel are really good and should generally apply to most AMD cpus as well (AMD's documentation is somewhat lacking in comparision).

And if you need guidance, you should just ask someone on /g/, I guess.
>>
bump
>>
>>106748047
what's your dayjob?
>>
>>106748047
thanks anon
thats very valuable manuals you gave me
for some reason i never thought about actually snooping around the web to find these kinds of resources
but now i actually saw whats in them
its hard to deny their usefulness
>>
File: stevejobless.png (543 KB, 1010x758)
543 KB
543 KB PNG
>>106748631
Post humourous images, not bumps

>>106748722
Closer to unemployed than stable, honestly. Hard to find a job that isn't being a spreadsheet jockey at a stationary store. Turns out that designing chips in Kicad is not a marketable skill in this economy afterall.

>>106749234
I hope you find use in them.
>>
>>106749631
>I hope you find use in them.
i certainly will, i dont know everything
and this is the good stuff that you gave me
>>
>>106749747
It's standard stuff. Good stuff comes from problem solving at looking at some old demoscene code imho.
>>
File: bigmac.jpg (34 KB, 640x640)
34 KB
34 KB JPG
>>
>>106750034
Lmao! Is this a Linux joke?
>>
>>106749997
yeah
ngl, the demoscene is some next level shit
https://www.youtube.com/watch?v=bD1wWY1YD-M
actual fukken C/++ virtuosos
>>
>>106750176
>Nostalgia Nerd
HE'S GONNA TAKE YOUR WIFE TO THE TEST SO YOU DON'T HAVE TO
>>
File: 1707591679708548.gif (976 KB, 272x244)
976 KB
976 KB GIF
>>106750176
You should check their code, it's often cooler than the effect and stay in touch with Amiga and IBM PC scenes as they're more Assembly Centric than Windows Demos.
>>
>>106750034
That's dumb but made me chuckle.
>>
File: fucking phone posters.jpg (351 KB, 1200x1286)
351 KB
351 KB JPG
>>
we are really redeeming with this one sirs
>>
>>106750296
yeah
>hex as boundary in loops
these people dream in machine code
ill have a good look at that, but i need to be fresh to have even a small chance at understanding this kind of stuff
sInt ISqrt15D(sU32 r)
{
sU32 e = sCountLeadingZeroes(r);
sU32 x = r<<(e+1);
sU32 i = x>>24;
sU32 f = (x&0x00ffffff);

x = ISqrtTab[i] + Mul24(ISqrtTab[i+1]-ISqrtTab[i],f);
if(!(e&1) )
x=Mul24(x+1,11863283);

return (x>>(15-e/2));
}

this is a square root. somehow.
its like the fisr algorythm, only all their code is this sort of black magic

>>106750259
i dont get it
>>
File: blaming-the-user.png (624 KB, 1024x1344)
624 KB
624 KB PNG
>>106750296
>>106750528 cont
anyhoo
ima catch some sleep. its 11 pm where i live and i woke up early today
thanks for the brain cookies

see you around, anon
>>
File: rzP3EYr.jpg (70 KB, 500x560)
70 KB
70 KB JPG
>>
File: nYJVYMD.png (232 KB, 1080x917)
232 KB
232 KB PNG
>>
File: IMG-9011.jpg (108 KB, 895x800)
108 KB
108 KB JPG
>>
File: IMG-8714.jpg (159 KB, 871x800)
159 KB
159 KB JPG
>>
>>106750951
NIGGUH JUST GOON

ayo boss when do we begin
>>
File: 1739555423956680.jpg (60 KB, 1080x1080)
60 KB
60 KB JPG
>>
>>106750528
AVGN and Nostalgia Critic intros combined.
>>
File: p0Sn8uC.png (336 KB, 735x778)
336 KB
336 KB PNG
>>
File: R4PlFRn.png (226 KB, 863x1096)
226 KB
226 KB PNG
>>
File: LrMFBzg.jpg (201 KB, 1076x1055)
201 KB
201 KB JPG
>>
File: AxBLZY3.png (481 KB, 498x630)
481 KB
481 KB PNG
>>
File: XWXBD4P.jpg (120 KB, 1242x1122)
120 KB
120 KB JPG
>>
>>106736154
the coffee and the walkoff is 100% accurate
>>
File: iGrYaEN.png (1.01 MB, 1080x1245)
1.01 MB
1.01 MB PNG
>>
File: ndORZAH.png (1.35 MB, 1165x1738)
1.35 MB
1.35 MB PNG
>>
>>106751080
The author is either 15 or 35.

not Tolkien you retard
>>
File: file.png (6 KB, 325x104)
6 KB
6 KB PNG
>>106750528
>hex as boundary in loops
That's a very common in even high level code to be honest.

Thinking in powers of 2 comes naturally once you start doing things.

The first 4 lines in there(I'm extrapolating meaning from the function and array names):
>Extract the number of leading zeroes, managing sign (which seems slow desu. It's something that will likely be done iteratively for most ISAs, but maybe I'm missing something) and puts it in E
>Discards All Leading Zeroes + Sign bit and puts it in X
>Extracts the Single Precision Float exponent into I (Atleast that's what I think, but it'll actually extract the first byte from the first enabled bit in the float, This is only the exponent when the exponent is positive in this case (Floating point exponents are biased with 127 so, an exponent value of 0x80 = 128 corresponds to the 2^1.). And Extracting the exponent whole is definitely much easier than counting the leading zeroes and doing the rest of this fuckery. Still, let's just take this assumption for now.
>F is loaded with the rightmost 24 bits of x ((23-(e) bits of mantissa and e+1 trailing zeroes)). For positive exponents, that's
basically mantissa left shifted by 1 (functionally, this means doubling the fractional part of the significand (but since the mantissa is only 23 bits, one bit at the top overflows into the exponent range).


The next part I don't understand either. Admittedly it's over 3am here and my brain is just barely still functioning
The next calculation pulls the Inverse square root(Presumably?) of the exponent in I from a table (Presumably a mapping like exponent 1 being mapped to -2, what you'd usually expect from a inverse square root operation on an exponent of 1 on base 2, but then the next part makes no sense because the answer would be just -2) then it's added with the 24 bit product (higher bits discarded, I'm guessing) of the diff of those table pulls and F.

I'll maybe look at it again after sleeping for a bit.
>>
>>106751119
there's a jigaboo in my skid meme
>>
Im going insane there is this image that always got posted in these threads of a peak accurate paint image of a dude with a holier than thou smugface , and the poster is asking for someone to post the original image
Does anyone have it
I tried drawing the guy from memory but i only draw anime girls sorry



[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.