Researchers Build a CPU Without Silicon Using Carbon Nanotubes

[notb]

An order of magnitude? My CPU is multiple orders faster.

Your Ryzen is also many orders faster than other CPUs made today. Computers aren't just the x86 machines we think about most of the time.
If you're around 40 or more, your PC CPU is multiple orders faster than another PC CPU you've likely owned in the past.

Anyway, this is all irrelevant.
Even if these carbon CPUs can only do very simple stuff today, they'll catch up at some point. Maybe in 5, 10 or 50 years - we don't know that. But we know their theoretical limits are far beyond what silicon offers.

 

[FinneousPJ]

Your Ryzen is also many orders faster than other CPUs made today. Computers aren't just the x86 machines we think about most of the time.
If you're around 40 or more, your PC CPU is multiple orders faster than another PC CPU you've likely owned in the past.

Anyway, this is all irrelevant.
Even if these carbon CPUs can only do very simple stuff today, they'll catch up at some point. Maybe in 5, 10 or 50 years - we don't know that. But we know their theoretical limits are far beyond what silicon offers.

Well yes, I hope new technologies will catch up to and surpass current ones. I'm pointing out an error in the text, current CPUs are not AN order of magnitude faster.

 

[Vayra86]

I'm picking Quantum CPUs will be out for the consumer before this ever makes it to market

The logical way forward is more specialization. I don't think silicon is out the door anytime soon, despite all the progress made in that area. It will take several decades (!) for silicon to run out of juice for at least the consumer market. And even then it will remain relevant for a large part of that market for another decade or more. There is also a point where 'enough is enough' and its likely that sort of CPU is still easily made on silicon. We've already actually reached that point, as we notice our old CPUs do a pretty good job even today.

Within that time I do believe quantum computing and perhaps solutions such as CNT will slowly start doing 'things' but not for us lowly consumer peasants. But cryptography, for example, and perhaps for CNT, applications that rely on extremely low power use - what about space? - and aren't computationally intensive, or less time critical.

You have to consider the fact that we've become pretty good at building efficient CPUs by now, so any new material is going to have to combat a highly finetuned set of best practices and architectures. But our CPUs also rely on several factors these new materials won't rely on, or in a different way - temperature, power, packaging etc.

 

[Wavetrex]

Imagine the people that made the first computer with vacuum tubes, which was order of magnitude faster and much quieter than previous computer using mechanical relays:
"THIS IS THE FUTURE"

It was... for about 15 years.

It's absolutely amazing how come we've come in less than the lifetime of a single person.

 

[Prima.Vera]

I'm picking Quantum CPUs will be out for the consumer before this ever makes it to market

This. Time to put de x86 dinosaur out of its misery tbh. It has been along for too many years, time for a new architecture and design based on quantum tech.

 

[medi01]

Gotta love how it only took 6 comments for someone to bash Intel on an article having nothing to do with them.

OP asked for it, ignoring how inadequate does word "thrash" describe cited post.

 

[notb]

This. Time to put de x86 dinosaur out of its misery tbh. It has been along for too many years, time for a new architecture and design based on quantum tech.

No, quantum CPUs will not replace x86. They don't support the instruction set, they can't be run in room temperature. And it makes no sense anyway.

 

[DeathtoGnomes]

No, quantum CPUs will not replace x86. They don't support the instruction set, they can't be run in room temperature. And it makes no sense anyway.

have you heard about the magnetic cooling ???

Researchers Develop Permanent Magnetic Cooling System for Quantum Computers

Researchers from the German Technical University of Munich developed a permanent magnetic cooling system for quantum computers, making them more practical.

www.tomshardware.com

 

[notb]

have you heard about the magnetic cooling ???

Researchers Develop Permanent Magnetic Cooling System for Quantum Computers

Researchers from the German Technical University of Munich developed a permanent magnetic cooling system for quantum computers, making them more practical.

www.tomshardware.com

Yeah, I've heard about multiple ways of cooling an object to near 0K. That's a fairly common problem in physics.
Any way you do this, you need a lot of energy (and time!) to cool the chip to near 0K - not to mention making it stay there under load.

But that's hardly relevant in this discussion. Because even if you're able to buy and run a quantum computer, it's still not a solution for general computing. Quantum processors are being developed for particular problems at which they'll be vastly superior to deterministic chips.
They will in fact work as "hardware accelerators" for particular tasks (e.g. tensor cores, hardware RNG), not standalone machines.

 

[DeathtoGnomes]

Yeah, I've heard about multiple ways of cooling an object to near 0K. That's a fairly common problem in physics.
Any way you do this, you need a lot of energy (and time!) to cool the chip to near 0K - not to mention making it stay there under load.

But that's hardly relevant in this discussion. Because even if you're able to buy and run a quantum computer, it's still not a solution for general computing. Quantum processors are being developed for particular problems at which they'll be vastly superior to deterministic chips.
They will in fact work as "hardware accelerators" for particular tasks (e.g. tensor cores, hardware RNG), not standalone machines.

Relevant, really? The size of the cooling solution shown in the article matters, quantum computers currently require a huge elaborate system. The article doesnt say how much exactly power is required, and I'm sure you know more than those on the research team, and judging from the photo it can fit your bedroom closet, so it must need 440 atleast.

If you're buying a quantum computer your not buying it to use for "general computing". LOL

 

[Athlonite]

If you're buying a quantum computer your not buying it to use for "general computing". LOL

No I'd be buying it to play Crisis LOL

 

[kapone32]

No I'd be buying it to play Crisis LOL

 

[notb]

Relevant, really? The size of the cooling solution shown in the article matters, quantum computers currently require a huge elaborate system. The article doesnt say how much exactly power is required, and I'm sure you know more than those on the research team, and judging from the photo it can fit your bedroom closet, so it must need 440 atleast.

It takes around 20 seconds to go from the article *you suggested* to the corporate website of this team:
https://kiutra.com
The photo in that article shows their current product - a research cryostat.
It's way too weak for a quantum processor, so we should treat its requirements as a lower limit (likely by a significant margin):
Power consumption: 6.6-7.2 kW
Cooling water: 6-9 l/min, 5-25°C

The cryostat IBM actually uses isn't much larger, but surely is a lot more advanced and expensive.

If you're buying a quantum computer your not buying it to use for "general computing". LOL

The idea that quantum computers will replace what we have today was suggested in earlier post - you've followed it.

 

[DeathtoGnomes]

It takes around 20 seconds to go from the article *you suggested* to the corporate website of this team:
https://kiutra.com
The photo in that article shows their current product - a research cryostat.
It's way too weak for a quantum processor, so we should treat its requirements as a lower limit (likely by a significant margin):
Power consumption: 6.6-7.2 kW
Cooling water: 6-9 l/min, 5-25°C

The cryostat IBM actually uses isn't much larger, but surely is a lot more advanced and expensive.
View attachment 131022


The idea that quantum computers will replace what we have today was suggested in earlier post - you've followed it.

yea maybe you should have spent that 20 seconds before you replied to me, afterall, I was replying to you not what everyone else said. Misunderstandings happen when speaking generically.

 

[notb]

yea maybe you should have spent that 20 seconds before you replied to me, afterall, I was replying to you not what everyone else said. Misunderstandings happen when speaking generically.

Replying to Prima.Vera, I said quantum computers won't replace x86 because - among other things - they need very low temperatures.
You said, quoting this particular comment, that there's something called magnetic cooling. In the following post you've also mentioned the size being somehow compact.
Is all that true? Or maybe I misunderstood something?

Let's move to the subject you started: the cooling itself. I can talk about that. Do you at least acknowledge that it's extremely hard to run these computers?
What's your opinion? What was your goal when you decided to challenge me on cryostats?
And now you're backing off in such a rough way...

 

[DeathtoGnomes]

Replying to Prima.Vera, I said quantum computers won't replace x86 because - among other things - they need very low temperatures.
You said, quoting this particular comment, that there's something called magnetic cooling. In the following post you've also mentioned the size being somehow compact.
Is all that true? Or maybe I misunderstood something?

I was referring to whats in the article not anything more. The image you posted is misleading. The image from the article clearly shows how small it is and after

Let's move to the subject you started: the cooling itself. I can talk about that. Do you at least acknowledge that it's extremely hard to run these computers?
What's your opinion? What was your goal when you decided to challenge me on cryostats?
And now you're backing off in such a rough way...

You just picked a random product to use for power reference? which happens to be the largest unit they have. I'm surprised more people didnt call you out.

as for backing away, yea real life calling....

 

[TomBattista]

Haha probably not you have to hit -1,000,000 Kelvin to get Quantum CPUs to even do what they are supposed to do. We will need an Alexa sized cooler to be able to do that at home and that is many years away.

Negative 1,000,000 Kelvin? How do you get a negative value on a scale of natural numbers?