/sci/ - Science & Math

>>14789127
Just YouTube how processors are made and how they work.

>>14789180
>step 1: take this silicon wafer
>step 2: inscribe these microscopic magic symbols on it
>step 3: finished
lmao

>>14789186
Electricity is channel through them, then given a certain of sequences, will spit the same thing out. Why do you think your OS boots up the same way unless you corrupt certain files (just representations of 0s and 1s on the harddrive)?

>>14789127
It all started with a steam engine and some punched cards, but then autists with nothing better to do with their time worked on it day and night for a couple centuries and here we are today.
>>14789180
Does YouTube have videos that reveal the trade secrets of TSMC? I had no idea.

We are the aliens. We will fuck other alien races, just 2 more millenia.

>>14789253
>Does YouTube have videos that reveal the trade secrets of TSMC? I had no idea.
Well, no. But I'm trying to say to him there is a logic to how processors work. Magic, if I'm not mistaken, implies some kind of force not fully understood in say another dimension or plane is being summoned to get something done. Processors and other parts of computers use electricity.

>>14789127
consider how many repeated elements there are in a city, how many molecules per wall, and how many elements of logic apply to its physics

>>14789274
nature proved that it's not necessary to consider those things in order to pragmatically simulate itself

>>14789299
Define consideration

>>14789299
No it hasn't retard that's why we need quantum computers to perform molecular dynamic simulations.
The church Turing thesis is not real

A universal quantum computer cam simulation any finite physical system completely by finite means (with polynomial blowup)
Classical computation is not amendable to this; classical computers can not simulate molecular dynamics without exponential blow up
This means that quantum computers can take exponential time algorithms down into polynomial time i.e. they are inherently superior to classical computation and are required to simulate physics. The universe can not be simulated on classical computation and the classical church Turing thesis is not correct. Nature has shown the exact opposite of what you claimed in that post

>>14789127
hows the view from up there?

>>14789127
i remember making a 8bit cpu with an 8bit gpu in minecraft at least 6-7 years ago. Its not magic, but its not easy to understand either

https://youtu.be/cNN_tTXABUA
First learn about transistors, logic gates and then construct adders, registers, decoders, encoders, mux, demux, counters etc using boolean algebra and logic gates. Then learn about address bus, control bus, data bus and how data moves around in general using these buses and ends up in registers and how decoders are used for memory organization etc. Then study the detailed operation of 8085/86. By this point you'll understand how this shit is theoretically possible. You can do this in a week.
After that just skim through how photolithography and cmos technology works.
I think this is enough for laymen

>>14789489
clockwork elves

>>14789581
>I think this is enough for laymen
Nigger that's literally half the syllabus of electronics engineering

>>14789601
Eh it's not really complicated or anything. You can easily study all of that within a week.

>>14789258
Just wait until you actually see what alien technology is like and you'll look back at your post in embarrassement in thinking this shit was remotely alien

>>14789581
Lmao

>>14789601
this is the basics. it only starts getting complicated once you want to go fast. instruction dependency graph, out of order pipelines, memory ordering, branch prediction, data locality prediction, cache coherency

>>14789601
>engineering
It's half the syllabus of trade school for electronics technicians. Much of it is taught in fucking high school. A lot of times though a lot of tradies love to slap "engineer" onto their titles to make themselves seem smarter than they really are.

What blows me away about the processor is the fact that it doesn't even do anything complex to simulate that virtual city. It basically performs one or two very simple operations over and over again. What makes it able to simulate a city is that it does those operations at a mind boggling speed.

>>14789127
The difference between a processor and your brain is how many times you folded the katana.

>>14789682
It isn't able to simulate a city. Simcity is not a real simulation

>>14789127
you’re complicating it. processors are just a network of switches. imagine a wall covered in hundreds of light switches, now imagine a language that’s conveyed solely with the switches, and how many/which ones are on or off at a given time. that’s a processor. except now instead of hundreds of switches, there are billions

>>14789127
They're not though, they're mostly just a massively scaled up (in terms of number of transistors) versions of the same basic logic that the first computing machines had.

>>14789601
You don't have to learn enough of it to be able to replicate it, just to understand how it functions on a fundamental level. Same way you can learn how an engine works enough so that you know your car isn't powered by "alien magic", but not know enough to be able to fix it.

But seriously, just understanding the concept of logic gates and the power of if->then statements is enough to demystify the basic process.

>>14789127
Calculator economics is cheap

>>14789127
Its not magic, but it might as well be to 99.9999% of people, and thats ok.

>>14789127
Watch Sam Zeloof's work on ICs.

>>14789889
>It isn't able to simulate a city.
Uh hu.
https://www.youtube.com/watch?v=NYYpisXsiRU

>>14789682
Imagine how long it takes you to add or subtract binary numbers with a pencil and paper. This isn't hard at all, and after a bit of practice you can carry out the calculations on paper and display the final answer pretty quickly on the paper in a few seconds at most. That's with markings made by macroscopic pencils, scratches with about 5x10^20 atoms of graphite per symbol (maybe more depending on the size of your handwriting) and inscribed on a piece of paper with 10^23 atoms per A4-size sheet. Ordinarily, then, you could fit about (10^23)/(5x10^20) = 200 symbols on one piece of paper with which to do binary arithmetic. There's obviously a lot of assumptions going into this back-of-the-envelope calculation here, but the salient point is that at the atomic level there's SO MUCH unused space with which to do calculations, it's laughable. Imagine being able to fit one 1 or 0 per atom on a piece of paper. Not only would we be able to fit more and larger numbers into memory while we compute, we'd also be able to compute them quicker because atoms are much closer together and the information involved with each simple operation doesn't have to travel farther than a couple nanometers between intermediate computations (as opposed to several hundred symbols and centimeters between symbols, respectively, on a piece of paper). There's lots of room at the bottom, both in sheer number of "workers" capable of storing and transferring information and in the space in between them necessary to transfer intermediate computations being smaller so it takes far less time. It's mind-blowing because human brains don't deal with or experience things at the atomic scale in everyday life, but the sheer scale of the number of computing components available in a microprocessor makes the phenomenon far more graspable than mystical.

>>14789889
4jynp0

>>14789127
Yeah, it's a magic tablet inscribed with symbols and then we send lightning through it to do things.

>>14789889
There's no such thing as a real simulation, because even just perfectly replicating the movement of air molecules in a single cubic meter would take all the processing power of the world combined. All simulations are approximations. And the fact is that real cities are not understood, and of course a game is programmed using simple rules, so we couldn't simulate them anyway.

>>14789604
>its easy, just do it lol
>99.9% literally can't even do it
This is such a dumb thing to say. You might as well tell a short person "just be 6 feet."

>>14790146
shhhh...don't let the subhuman masses know they'll never be human. I need their false hope for entertainment.

>>14789380
that's not what the dunning-krueger effect actually says.

>>14789990
>>14790022
Not simulations of cities
>>14790057
>There's no such thing as a real simulation, because even just perfectly replicating the movement of air molecules in a single cubic meter would take all the processing power of the world combined
EXACTLY

>>14790292
Simulation, just not perfect and accounting for every variable (in the case of UE5, just simulating the look of the thing, which it accomplishes nicely). You can simulate macro phenomenon while only fudging the details and still end up with something that reflects the behavior of the real thing to the degree you can use it for prediction and analysis.

>>14789589
factorio autism

>>14789601
>half of the syllabus
aside from photolithography and CMOS, that's literally just digital logic + computer architecture pair of classes that you see standard in any CE/CS curriculum...
At most there's an extra microarchitecture or OS level class.

>>14790292
>hurr guys we can't have meaningful models with different levels of abstraction
hope you're cursing newton's lack of specificity in his mechanical theory once i beat you to death

>>14789186
Yeah bro a bunch of switches are literally magic.

>>14789589
>clockwork elves
It's crazy that they can make so many of these unfathomably precise machines.

>>14790987
>just inscribe 9 billion switches on a tiny stone. It's pretty simple bro
Processor development was not gradual. We suddenly switched from processors that were as big as big rooms to fitting millions of them on tiny chips. Something is fishy.

>>14791061
>Processor development was not gradual.
Yea, I heard it was quite a process.

>>14789604
>>14789647
>Much of it is taught in fucking high school.
No it isn't you fucking retard.

Read Code by Charlez Petzold then do Nand2Tetris.

>>14789267
>Magic, if I'm not mistaken, implies some kind of force not fully understood in say another dimension or plane is being summoned to get something done
what a peculiar opinion to have where did you find that hot n spicy opinion

>>14789375
there is no such thing as a quantum computer besides an overly wordy described convoluted proprietary machine that makes retards pretending to be smart (academia) shill out the dough for their own and it's predominantly just a new type of computer with bit registers greater than binary giving the illusion of complexity but not the benefit of anything you kooks could actually define since it still mechanically can only hold a single known value per bit.

snake oil salesmen

it's a series of tubes

>>14791610
this quantum speedup is not possible. google is coping and seething

>>14791061
we switched from computers using purely mechanical switches to computers using electromechanical switches to vacuum tubes and then finally transistors, which turned out to be easy to miniaturize

I recall a vid a arab prince made in the kabba (the black cube/former site of the home of their religious figure) the inside is now a sort of museum of rare old artifacts one of which looked like a crude circuit board.

any sufficiently advanced technology is indistinguishable from magic

if alien magic is real then where the fuck is my android gf huh?

>>14791697
Most technology is magic to most people in the world. People have no idea how planes, computers, phones, or cars work. Magic is the default position, people just pretend otherwise.

>>14789267
And what is electricity? Is it fully understood? Where does it come from? Hypotheses non fingo.
Science and magic have more in common than you would like to admit. The only difference is that science usually works, while magic does not. Science itself ultimately evolved from Renaissance natural magic and alchemy.
OP is clearly a retard but it is also true that in the eyes of a retard there is no difference between magic and advanced technology, and there is a reason for that.

>>14791610
>>14791612
Quantum computers exist, but even if we assume they don't and that you're correct, well then it would just mean that it's not possible to simulate nature at all on any machine.

>>14791774
you're a homosexual pedophile who jacks off the children's cartoons. your desire for a robot gf is your acknowledgment that your own personality is so odious and repulsive that even you recognize that purchasing a robotic slave is the only way you'll ever get a willing human-like companion of any sort.
you could work on becoming a less repulsive person, but instead you choose to wallow in your ridiculous escapist fantasies and lies

>>14792287
Quantum computers exist exactly as defined and you are retarded
>there is no such thing as a quantum computer besides an overly wordy described convoluted proprietary machine that makes retards pretending to be smart (academia) shill out the dough for their own and it's predominantly just a new type of computer with bit registers greater than binary giving the illusion of complexity but not the benefit of anything you kooks could actually define since it still mechanically can only hold a single known value per bit.

>2022 and the only evidence of quantum computers is still just CGI / After Effects animations and/or just people explaining how they imagine such technology might work
>no direct evidence suggesting otherwise

>>14792554
Quantum computing machines exist and the principle of physical superposition is used to perform calculations that can not be performed on classical machines.

>>14792558
no, you are just mentally retarded and believe in any old retard selling you lies.

your imagination is not real bruh

>>14792566
wtf are you talking about
Classical machines are not capable of computing a shitload of functions that nature computes in constant time.

what the fuck are you talking about
nothing anyone knows about is capable of doing such computations

get out of your fucking imagination and get a real job

>>14792596
There is an entire field of physics concerned with this.
If what you are saying is true, why can't classical computers simulate even the most basic molecular dynamics?

What's magic about it? We can control lasers pretty precisely. Just look at bluray players and how they can put 250MB of data onto a blu disk with a laser in your home with a burner that costs 100$. Just imagine a couple billion dollar device can do the size of a room.

>>14789127
Have you seen a EUV machine before, it's the same shit. Knowledge so far from my comprehension thats its basically alien.

>>14791061
>Processor development was not gradual
Yeah we just went directly from vacuum tubes to the modern 5nm process. It's not like the original integrated circuits that had transistors big enough to be seen with the naked eye or anything.

>>14789127
It all makes sense when you consider the basis of circuitry are shrunken-down logic gates, smushed into as compressed a space as humanly possible, which makes it difficult to recognize from a glance.
What beats me is having to track any of it with precision beyond what's needed for it to do its job. Guys like me make lousy programmers.

>>14789127
any sufficiently complex system will look like that there. Check out this portion of my unfinished base in factorio, for example.

>>14789127

Non computer engineer here.
How do you actually get those little transistors to do what you want them to do with software. Like how is this process mediated. I'm at a loss.

It's IP built over 70 years of practice. Look at the early microprocessors like the Intel 4004, those have thousands of transistors which is a lot but still simple enough to be designed by hand.

>>14789127
Obviously more of a thread for >>>/x/, but there is a conspiracy theory that spreads the idea that cellphones and fiber optics are alien technologies deciphered from Roswell. (Also always amazes me how all these folks who say technological progress has come to an end, forget how smartphones changed the world in their own lifetimes.)

The progress to microscopic transistors is pretty straight forward, even their explosion into the market took over a decade. Computers were, at first, a decidedly unintegrated technology. They were composed of vacuum tubes, resistors, capacitors, inductors and mercury delay lines, and as a result were huge, expensive and power hungry. The situation improved with developments in microelectronics based on solid-state germanium transistors and diodes, which began replacing vacuum tubes in radios and phonographs, and eventually led to the first commercial transistor-based computers in 1959, with more than ten times the processing power in the same space. That same year, the development of the planar process at Fairchild Semiconductor allowed tens of silicon transistors to be fabricated at the same time on the surface of a single-crystal silicon wafer (ten years later it would be possible to fabricate thousands of transistors). This inventionwas quickly followed by the commercialization of the first bipolar digital integrated circuits (ICs) in 1962, and from that point on, progress in semiconductor ICs became exponential, with the maximum number of components integrated in a silicon chip doubling every year, at least initially. But what allowed all the functions of a general-purpose computer to be integrated together was a monolithic central processing unit (CPU) — a microprocessor — and the first commercial microprocessor was born nine years later in 1971, the Intel 4004.

It was another 10 years before you really had home computers and the processing capability pretty much doubled consistently every few years until today.

>>14789604
Kek anon, understanding transistors and then moving on to how to construct logic gates with them easily takes a week (assuming you want surface knowledge). Stop larping, if any one believes this takes a week, they're trolling.

>>14789127
Earff iz a spaceship nigga

>>14794212
>understanding transistors
>takes a week
you clearly have never studied microelectronics
i bet you don't know the difference between a BJT and MOSFET, but can draw an AND gate real gud

>>14794045
More of her, please

>>14794233
>i bet you don't know the difference between a BJT and MOSFET, but can draw an AND gate real gud
not him but this describes me

>>14792605
computers can simulate a lot with enough intelligent programming. what do you believe video games are.
better answer: because not enough is known about molecular dynamics.

>>14795482
You don't need to know how transistors work if you're not an EE.

>>14794233
>Assuming you want surface knowledge
I was being very generous. I know you're right anon. Hell, we have a 6 month physics course on transistors in my uni.

>>14789127
I work with CPU design, I can answer questions you might have. We're designing a novel general purpose architecture (not mill)

>>14795491
They require quantum mechanics to simulate

>>14795688
what's the point? does it save a ton of power somehow?

>>14789127
>This shit is literal magic and you can't convince me otherwise.
Literally make one yourself and you won't be shocked anymore.
>How can this small fucking thing perform so many complex calculations at such a quick pace
It's a bit too complex to grasp all the levels of abstraction involved at once but any decent CE grad student could build one, given sufficient resources and time.

>>14795863
Yes, it tries to do as much of the decision making in the compiler, which means it's very different from regular RISC processors. This means we will have much larger code size, and you can forget about binary compatibility, however we believe that the datacenter market provides a viable point of entry. No amount of nice to have will trump doing 30% more compute per watt for data-center workloads.

We are very far away from this goal mind you, but we are starting to see good performance on smaller benchmark suites on our FPGA implementation of our chip.

>>14789186
Any sufficiently advanced magic is indistinguishable from technology.

>>14795705
You don't need quantum mechanics to simulate macro level phenomenon. You don't even need molecular or micro mechanics or chemistry, plain old top level Newtonian physics will suffice 99.9999% of the time to get you something useful.

An imperfect simulation is still a simulation.

>>14789581
In English doc?!

>>14789127
I assume you mean specifically the advance ones and not some microchip any chink can make on a wednesday

>>14797535
Eh, the most advanced ones are made by chinks in Taiwan. (And China is gearing up to make 4nm chips.)

>>14797576
Taiwanese are mega chinks I mean the chinks that sell sewage oil and put it on their dogs also make microchips on the side so it's not that hard

>>14797576
>Implying Taiwan isn't part of China
Careful, you're gonna summon the CIDF.

>>14789127
The fabs are more impressive than the processors themselves.

>>14789127
Microprocessors are the speartip of human progress im almost sure. Unfathomable amount of research and ideas and engineering go into their mass production and an unfathomable amount of opportunities are available because of them. Not to mention GPUs are the first step t computronium.

>>14795869
>It's a bit too complex to grasp all the levels of abstraction involved at once but any decent CE grad student could build one, given sufficient resources and time.
There's a massive difference between the different tiers of CPU's. That's like saying you could swim across the Ocean as long as there was no time limit.

>>14795688
>We're designing a novel general purpose architecture
Any details?
Over the last ~decade, I've submitted a bunch of different things to various Academic/Government/Military Agencies and Departments, but none of them seem interested, despite the fact that soon after I email them my projects, I start to see articles and research papers that are almost identical to the research proposals I submitted.