The Race to Build a Perfect Computer Chip

Thank you

thx

Thank you brother

Thank you!

You are welcome Mr.Singhania.

they said something without saying anything

Epyc.

It’s genuinely great

@Brendon Pitcher Anyone concerned about what we're doing to the Earth, absolutely yes.

@Wallon from all of this, thats what you take away?

@Alan hat , yes.

R. Daneel Olivaw had a positronic brain in the 1940s...

@Mr Fake thanks for the info

@Joshua Thomas sure,whatever satisfies your comfort zone

Look into metamaterials. First company to try and scale advanced nano materials

@Nayeb 222 you have faith in fiction

@a a I do have faith in religion, so it won't be an option for me

Silicon is common

@Amos Batto I would just slightly expand on your thoughts. In the case of desktops there have been great increase in efficiency over the last 20 years, through the increased density and lowering of voltage on CPUs. Hard drives were replaced with SSD drives. However, when it comes to GPUs, it's the complete opposite which is why the overall power usage has not declined significantly. It should also be pointed out that in most cases, this increase in GPU use of power has been for gaming. So the real question is how much of the lack of power efficiency in a desktop is due to the demand for better gaming and is this truly the way to power efficiency nirvana?

"once vendors see that they can do more with less, they'll just do more and so the cycle continues" thats called progression.

@Michael Edwards Rustacians always come out of their woodworks in these type of tech videos Rust is not the modern and better version of C until they sort out and upgrade memory management better than C but they will never sort it out like GNOME developers I have yet to see real actual upgrade to old pure C

@Michael Edwards If a piece of data is in a file, it will take a lot longer to read than if it's in RAM, yes. Everybody already knows that, you don't need to tell me. You already can change the microcode for some x86 CPUs (AMD K8, K10). But your custom instructions won't run any faster than regular code because the microcode sequencer can't generate any more operations than the regular non-microcode instruction decoder (it's not designed for speed, it's there for security and 16-bit compatibility). You're gaining nothing. Micro-op quantity is not tricky. Your operations on ALL out-of-order CPUs are: Load, Store, Arithmetic operations, and Jump. Everything else is just a combination of these including fancy stuff like Micro-op fusion and Macro-op fusion.

Computer chips work in digital domain as it is easier for us too handle digital data in hardware. Analog design is really hard compared to digital design. I myself being a digital design engineer know how things get easier after analog to digital conversion. Real time analog data is easier to understand in binary.

Well ye, but more than that because the tech that works well enough is the one already receiving most of the investment. If silicon had become untenable 10 years ago we would have been forced to switch faster to something radically different. The same thing is true with NAND flash memory - it's truly terrible for power consumption, and the wear rate of each memory cell is anything but great for long term storage. But because the investment was so deep even the most promising advanced replacement tech has constantly been left to rot even as flash becomes an ever greater power drain on our mobile devices.

Nonsense. We're using all of our existing technology and pouring hundreds of billions of dollars per year into researching new methods.

yup the consistency is the death of developement

Epyc.

Data's neural network was photonic.

@Owen Hand yea

@Owen Hand yup

@Owen Hand A single company has a technological "moat" for making this UV equipment. Their production costs are increasing as the technological difficulty increases exponentially. Currently, tin droplets are precisely dropped into a focal point and vaporized to an excited plasma in order to emit the characteristic UV light. This tin element has to be changed to another one that emits even shorter wavelength. As wavelength is reduced, the masks and photoresist will have to be made compatible with higher energies into deep UV or even X-ray. Also, as the energy level is increased, this will lead to increasing ionizing radiation damage to equipment and semiconductor. Edge diffraction is an issue. Consistently reproducible X/Y physical positioning of wafer substrates and optics to well within 3 micron *for each chip position* over an entire wafer diameter is *already very difficult.* It will get harder as the demanded precision gets even tighter. This slewing, settling into position, and flash exposure must also occur more often as the number of chips on a wafer increases, which results in slower production.

@Amren Miller The Tesla Dojo tiles are each using 40 optical fiber interconnections to the adjacent [array] tiles. I presume most supercomputers are now using optical fiber connections between server cabinets, perhaps even between blades.

Wirth's Law. We don't necessarily need better computers. We need software to be more efficient. Nowadays it's normal for computer programs to occasionally crash due to memory leaks or bug in the code. I work at a datacenter and I have to use this app on my phone to do daily routine inspections and the app crashes when open for too long... It's crazy how tolerant we became of unstable software.

That is true, but back when that happened the worldwide use of computers was just a tiny fraction of what it is now. The increase in use means we need to push the hardware efficiency ever further to keep up.

Epyc.

I think this comment is built on a false premise, that code has become less efficient over time. Nearly all code is run way more efficiently now than it was 20 years ago. It’s just that as processing power increase, programmers are able to create applications that do more. And in the arms race for consumer attention, you then have to do more in order to compete. So the issue is that more functionality is being added, not that the functionality is less efficient.

@Meleeman mate, i dont want to learn anymore than necessary when it comes to computers. They are supposedly made to adapt to us not us to a binary code. Yes i have done unix and linux before but i gave up, i dont wish to reinvent the wheel, i will let other nerds like you do that. Yes you can, yes i can do this and that but i chose not to. I can think of better things to do with my life.

@Meleeman Using c++ server side can have a reduction in energy usage

the quickest thing you could do is learn how to use linux and a terminal, and you would already be using less power the a majority of people, and use a window manager like i3wm. and use more terminal applications including on your phone like termux. its not as conveniant but you can do quite a bit with a terminal. so much i'm convinced that real work is done in a terminal.

nope. because in order to make money code needs to be shipped fast. there are better. what we can do is encode more in less. so instead of binary computers we use ternary, or even quarternary computation. and that could increase the amount of possible calculations. the reason developers are less concerned with memory constraints is because its expensive to write efficient code, as it takes longer and you need to understand more math and how computers work in order to write efficient code. its also more prone to bugs and errors. what you need is something simple enough to write but provides enough control for the task at hand. and most people don't even know until the product is shipped, optimizations happen after the product is built. a real solution would be using analog computers and a whole bunch of them to do specific calculations, and then translate them into binary. this in principle is how and why asic mining exists because instead of abusing sand and making it think, we simply just let it read the electrical charge outputs from several mechanical computers and let it process those inputs via conventional silicon which will need less power to operate since it need only read from its mechanical computer counterparts and maybe do a few calculations here and there.

Stacking is an effective path to higher density IF the generated heat can still be conducted away rapidly enough. For example, memory circuits typically are lower power than processing circuits, so memory can easily be stacked. At the moment, memory chips are diced apart, tested, and bonded into stacked silicon sets that pass buss signals vertically. Four layers are already common, and thicker stacks are coming. Typically, the validated memory stacks, processors, capacitors, and interface chips are bonded to a multilayer metalized silicon substrate that passes power, bus, and clock signals among the chips.

It will be mayhem

Energy consumption would increase either way

Any technology that can build a >1Ghz 32bit exact adder...

really like for example

Me to .....1 cubic cm. of DNA is capable to store all the information of all datacenters in the world 🌎 ( I was told )

@mynameisjoejeans Right but mobile devices are also incredibly efficient. Point is that I don't think media consumption time has increased significantly or kept up with the vast decrease in energy consumption afforded by digitization/streaming. Counter point: 4k 60fps and scrolling through Tik Tok streams. Still think it's handily in favour of energy reduction on a per capita advanced economy basis.

@djayjp hours of tv watched is an unrepresentative metric, as digitisation has led to many more content forms which are no longer classed as TV. Streaming, social media etc are all new categories that when combined constitute far greater consumption than was possible via analogue technology.

@mynameisjoejeans Right but I bet, *all else being equal*, energy consumption is still down in that regard (media consumption). One can look at the number of hours of TV watched for example and I don't think it has changed much over the past few decades.

@djayjp The Jevons Paradox is fairly well documented in a variety of contexts. Energy and resource consumption due to convenience is certainly one of the most prominent examples.

@Floris Maybe you guys are somehow under the impression that I'm against going digital. Not at all. I am extremely pro-digital, makes everything more efficient and convenient. I was thinking of the universal truths of how that extreme convenience always create infinitely more demand. One quite random example is the traffic problem. We attempted to solve it by increasing the number of lanes and finally just realised that it creates more traffic when we do that. Our models seem to show that even going autonomous may not solve traffic just because so many more people will be travelling. Elon had to come up with boring company to attempt to solve traffic.

@J X Hahaha You must be joking

@J X agreed China is leading compared to India

@Bhuvanesh s.k Compared to India. India has potential but it just isn't as big as many people think

@J X u mean compared to USA?

China's already much further ahead in all types of chips, whether carbon, photonic or RISC

They will eventually get cheap enough, the beauty of the capitalist system. 50 years ago owning a computer was impossible. Today the average American has two computers.

New challenges will create New opportunities. Maybe, not with commercial applications but, these technological breakthroughs will initiate their own journey with defence and space Applications !!!

Lets hope not

Arduino is an example of this. The programmer, if you can call him that, has all the clever libraries written by someone else and all he does is connect these together. He then believes he's written some code, with no understanding of anything at the hardware level.

@Michael O that should be a completely separate job from the developer , the developer only needs to develop , but the conversion itself is the optimization factor and that can be done separately by those who create the languages

@Zhinku Nakur The OP is correct, I write embedded code in C, yet I use 128Kilobyte of flash(rom) and 8Kilo byte of ram(actually 6K) not Mega bytes Not Giga bytes, yet its considered leading edge in the industry its put into.... yet PC/Android/Linux consume so much room... unlike my bare metal programming My power use without even trying to green it ... 2.4 Watts that max during motors driving. drops to under 1 watt running the rest of the time 99.9% time. Motor only run 8 seconds.

Interesting point. I would say compilers are fairly bloated (unnecessary libraries added), but in terms of programmers - it depends as some write their own code for each thing while others use libraries (which does add unnecessary overhead) and since companies like their programmers to be as efficient as possible, library based programmers will become more prominent. Personally, I think you can run a lot of real resource intensive programs if we multi-threaded our programs or programmed them in some lower level languages (eg. Assembly) but that would contribute to the shortage of programmers as Assembly creates a very steep learning curve. So mabye?

@imeaktsu7 That's the whole point, it should be taken into account. There is a nearly 2 decades power efficiency gap between C (or Rust) and Python. Also we should aim to more frugal softwares and websites, with less useless stuff.

“It could make chips to beat quantum computing”, based on what I know about quantum computing, I don’t even think that’s possible. Quantum chips calculate the probability of an objects state before its even measured.

Interesting idea, but I don't think pure carbon has any ferromagnetic properties.

Remember, it is these validation seeking individuals which are pushing scientists and engineers to innovate and come with fundamentally new solutions, ensuring that we progress as a species. If we didn't need to upgrade, no one would care to innovate, and we would still be happy with stone-age technology.

Not when they are copper.

@AlanTheBeast100 Makes sense.

You're correct about the second part. But unless you're using a space heater instead of an HVAC, it's not zero. Heat pumps which move thermal energy are actually more efficient than pure electric to heat converters.

@Eric MOULOT first of all his point number 1 about python and c is ridiculous. And 2ndly risc v or any risc architectures can't replace cisc. Most of data centre cpu run of cisc for a reason

Bruhhhh ur point number 1 doesn't make any sense. U r talking in the software domain. Plus there's a reason why everyone use python. It's faster to build and test

@Raju Aditya Sounds like a reasonable argument to me. Maybe you have some insights to share that'd change my mind and his?

@Raju Aditya we all are interested learning and yeah we normies may not understand key concepts

This is such a normie take that it is hilarious.

@User That is the problem. Can you say untested loops. Make an error like that and things fall apart. As long as humans have a hand in the process, there will be mistakes! Can you say self-driving Tesla auto?

You will have to emulate / translate older code and hope it works

That said technology has experienced major innovations in the last 20 years, but news always like to make it seem like we’ll have futuristic technology in the coming decade

Yea, it's an interesting puzzle