What does this breakthrough mean for the future of computing? Help me understand the new petahertz transistor.

[Space Lynx]

University of Arizona scientists unveil breakthrough petahertz-speed transistor

A team of scientists has unveiled a breakthrough that could one day propel computers to operate at speeds millions of times faster than today's most advanced processors.

www.techspot.com

Just finished reading this article, and I don't fully contemplate it. Does this mean like a future supercomputer might use this technology and basically leave every other supercomputer in the dust? Or will it lead to the development of actual new CPU's that use lasers and nano-carbon tubes with light being on and off counting as 1's and 0's as seen here: Light Based CPU

Looks like graphene is what led to the breakthrough, graphene seems to be doing quite a lot for us in a lot of technological advancements. If I am not mistaken, this is why we no longer need thermal paste as well, with the Kryosheet cooling.

 

[ty_ger]

"If successful, this research work could mark the dawn of a new era in computing, where the speed of light, rather than electricity, sets the pace for innovation."

¿Que?

The propogation speed of electricity is very near speed of light. Makes me instantly wonder what I am supposed to be reading.

 

[Space Lynx]

"If successful, this research work could mark the dawn of a new era in computing, where the speed of light, rather than electricity, sets the pace for innovation."

¿Que?

The propogation speed of electricity is very near speed of light. Makes me instantly wonder what I am supposed to be reading.

Well, those devices are limited my metal based silicon, and what I am understanding here is that this will be graphene based with lasers, so much much faster, because no limit by the metal?

I don't know either, that's why I made the thread, to try to gain some clarity.

 

[Solaris17]

Does this mean like a future supercomputer might use this technology and basically leave every other supercomputer in the dust?

Every other being what exactly? Its not like a machine with a PHz cpu is going to come out of no where. Graphene in this case has taken decades to even get to this prototype phase. I think a popular misconception is that things like this are new. They were messing with gallium, graphene shit several metal types to dope silicon back when socket A machines rocked TPU.

Something like this is not only lightyears away it will be cultivated by several places as the research papers get read. Its like how quantum is still currently the size of several rooms but several institutions have them.

Speeds like this are great in theory, IBM? I think it was came out with terahertz several years ago and it was some like small amount of transistors.

It should also be noted that switching transistors this fast isnt even a break through in the sense of new. Here is Darpa doing it in 2014.

https://www.extremetech.com/defense/193343-darpa-creates-first-1thz-computer-chip-earns-guinness-world-record

If it came to light for consumers not only will I likely be dead but it will happen in a certain order for sure.

1: Research
2: Small test chips
3: Military
4: Space
5: Big data (Enterprise)
6: Consumer

Roughly. I got to live in the time line where we transitioned from vac tubes to microchips and its wild to even fathom I once met a grey screen with green text and was like "wtf is this???". It moves fast for sure, but things like this are reported in there infancy; This is as close to processing x86 as I am to being a saturation diver. I will likely be in the dirt when this trickles down to consumers, IF it ever does.

Plenty of projects like this died in there infancy or somewhere on the chain above because of the advent of something better coming to market faster. Ask GE how they felt when the microchip came out and the vac tube industry fell from under them.

Its still sick and ita awesome to read things like this, but the reality I think is that it will be human generations, not GPU generations of time before we are wearing our qubit powered gshocks.

 

[Space Lynx]

Every other being what exactly? Its not like a machine with a PHz cpu is going to come out of no where. Graphene in this case has taken decades to even get to this prototype phase. I think a popular misconception is that things like this are new. They were messing with gallium, graphene shit several metal types to dope silicon back when socket A machines rocked TPU.

Something like this is not only lightyears away it will be cultivated by several places as the research papers get read. Its like how quantum is still currently the size of several rooms but several institutions have them.

Speeds like this are great in theory, IBM? I think it was came out with terahertz several years ago and it was some like small amount of transistors.

It should also be noted that switching transistors this fast isnt even a break through in the sense of new. Here is Darpa doing it in 2014.

https://www.extremetech.com/defense/193343-darpa-creates-first-1thz-computer-chip-earns-guinness-world-record

If it came to light for consumers not only will I likely be dead but it will happen in a certain order for sure.

1: Research
2: Small test chips
3: Military
4: Space
5: Big data (Enterprise)
6: Consumer

Roughly. I got to live in the time line where we transitioned from vac tubes to microchips and its wild to even fathom I once met a grey screen with green text and was like "wtf is this???". It moves fast for sure, but things like this are reported in there infancy; This is as close to processing x86 as I am to being a saturation diver. I will likely be in the dirt when this trickles down to consumers, IF it ever does.

Plenty of projects like this died in there infancy or somewhere on the chain above because of the advent of something better coming to market faster. Ask GE how they felt when the microchip came out and the vac tube industry fell from under them.

Its still sick and ita awesome to read things like this, but the reality I think is that it will be human generations, not GPU generations of time before we are wearing our qubit powered gshocks.

You are probably right, but when IP theft is so rampant in the modern consumer world + combined with advanced AI that none of us have access to, I wouldn't be surprised if the timeline is much sooner than people realize. You are correct though, it will be military/space stuff first and foremost, and we will be long dead before we see a gaming PC using these speeds, probably anyway.

 

[eidairaman1]

peta
P
10^15
quadrillion Hz

1000000000000000 Hz

 

[Solaris17]

You are probably right, but when IP theft is so rampant in the modern consumer world + combined with advanced AI that none of us have access to, I wouldn't be surprised if the timeline is much sooner than people realize. You are correct though, it will be military/space stuff first and foremost, and we will be long dead before we see a gaming PC using these speeds, probably anyway.

Its not even IP theft and im not going to let this forum get polluted by any more bullshit like that in this thread. These are just research papers, researches and students can access this. As a normal person you can buy them. Its all over, always has been, you can read about this and tons of other tech related shit from most prominent institutions around the world to see how they got something to work. Didnt you teach? You should know this.

 

[Space Lynx]

Its not even IP theft and im not going to let this forum get polluted by any more bullshit like that in this thread. These are just research papers, researches and students can access this. As a normal person you can buy them. Its all over, always has been, you can read about this and tons of other tech related shit from most prominent institutions around the world to see how they got something to work. Didnt you teach? You should know this.

Huh? I think you misread my comment. What I am saying is that other nation states may take information from researchers like this and use their advanced AI to make it happen way sooner than any of us realize. That's all I was trying to say, is that I think the timeline is actually going to be quicker than people realize, but I was just agreeing with you that it won't probably ever trickle down to us as it will be for space/military stuff first. They did say in the article they are filing for a patent on the technology, but if the information is already out there, no doubt in mind other players are already working on it.

peta
P
10^15
quadrillion Hz

1000000000000000 Hz

I need to understand this in the context of a terahertz though, 1000 megahertz is 1 gigahertz, is 1000 terahertz 1 petahertz?

 

[eidairaman1]

Huh? I think you misread my comment. What I am saying is that other nation states may take information from researchers like this and use their advanced AI to make it happen way sooner than any of us realize. That's all I was trying to say, is that I think the timeline is actually going to be quicker than people realize, but I was just agreeing with you that it won't probably ever trickle down to us as it will be for space/military stuff first. They did say in the article they are filing for a patent on the technology, but if the information is already out there, no doubt in mind other players are already working on it.



I need to understand this in the context of a terahertz though, 1000 megahertz is 1 gigahertz, is 1000 terahertz 1 petahertz?

Yes thats correct

 

[Space Lynx]

Yes thats correct

holy crap, this really is groundbreaking.

 

[Shrek]

I need to understand this in the context of a terahertz though, 1000 megahertz is 1 gigahertz, is 1000 terahertz 1 petahertz?

• Mega 10^6
• Giga 10^9
• Tera 10^12
• Peta 10^15

 

[lexluthermiester]

peta
P
10^15
quadrillion Hz

1000000000000000 Hz

So in reality, it's not petahertz it's quadrahertz.

 

[johnspack]

I'll just chuck this in here... been around for years:

Optical computing - Wikipedia

en.wikipedia.org

 

[Space Lynx]

I'll just chuck this in here... been around for years:

Optical computing - Wikipedia

en.wikipedia.org

Why would this be big news then?

 

[Athlonite]

f it came to light for consumers not only will I likely be dead but it will happen in a certain order for sure.

1: Research 10~15 years
2: Small test chips 15~17 years
3: Military 20~22 years
4: Space 25~30 years
5: Big data (Enterprise) 35 years
6: Consumer 40 ~45 years

I'd be atleast 95~101 years old before this makes it to consumer PC's by that time we'd probably be using our own brains as compute devices and our own eyes as monitors

 

[Apocalypsee]

I remember Intel TeraHertz transistor back in 2001 but as you can read here it's not being used.

Intel TeraHertz - Wikipedia

en.wikipedia.org

 

[TumbleGeorge]

Why would this be big news then?

1.Team of developers advertising themselves.
2. Medias try to get more readings and money for ads.
3. You is too excited from everything that is presented with some suspense and/or inspiration language.

 

[Space Lynx]

1.Team of developers advertising themselves.
2. Medias try to get more readings and money for ads.
3. You is too excited from everything that is presented with some suspense and/or inspiration language.

They are trying to get money? They work for a Uni, and would have had to sign contracts with that Uni before being employed by said Uni to do research, aka, most of anything they invent money wise will be going to the Uni... (usually these contracts have it built in that even if the researchers patent something part of that money will go to the Uni... usually a big chunk of it actually).

Care to explain your thought process on this? I am curious.

 

[TumbleGeorge]

It is natural to assume that it is for money and as much as possible. To do this, they need to become famous on the Internet and again, their goal is to be as famous as possible. Where they work does not negate greed. Even in this case, it probably allows them more fame, because if they worked for a company, they would be as limited in their statements as possible so as not to publish data that could be useful to competitors.

 

[leonavis]

University of Arizona scientists unveil breakthrough petahertz-speed transistor

A team of scientists has unveiled a breakthrough that could one day propel computers to operate at speeds millions of times faster than today's most advanced processors.

www.techspot.com

Just finished reading this article, and I don't fully contemplate it. Does this mean like a future supercomputer might use this technology and basically leave every other supercomputer in the dust? Or will it lead to the development of actual new CPU's that use lasers and nano-carbon tubes with light being on and off counting as 1's and 0's as seen here: Light Based CPU

Looks like graphene is what led to the breakthrough, graphene seems to be doing quite a lot for us in a lot of technological advancements. If I am not mistaken, this is why we no longer need thermal paste as well, with the Kryosheet cooling.

So basically this could switch on and off in 1"10^-15 s, or 1fs. That is pretty fast. If it proves viable, it also circumvents the problem of quantum tunneling with ever smaller gate length. Still on a loooong way to planck-time

 

[Shrek]

Heisenberg's

• ΔE Δt ~ h

works out to about 1 eV for 10^-15s so I would imagine this is a fundamental limit for electrons (the Planck time would imply much larger energies)

 

[Onasi]

Huh? I think you misread my comment. What I am saying is that other nation states may take information from researchers like this and use their advanced AI to make it happen way sooner than any of us realize.

That’s… not IP theft though. These research papers are available. Scientific research on these things is, for a lack of a better term, “open source”, that’s how it’s meant to be. To put into perspective from my own field, if some company or country takes the data from our research institute published paper on, say, more efficient ways to combat persistent organic pollutants in water and uses it to improve on their industrial filtration systems then, well, no IP theft is involved there. The system is actually working as intended, that’s what fundamental research IS DONE FOR.

 

[Frick]

It is natural to assume that it is for money and as much as possible. To do this, they need to become famous on the Internet and again, their goal is to be as famous as possible. Where they work does not negate greed. Even in this case, it probably allows them more fame, because if they worked for a company, they would be as limited in their statements as possible so as not to publish data that could be useful to competitors.

This is how developing stuff works. Have theories, develop prototypes to test it out practically, and many years down the line maybe something will shake out. It's not all bottomless greed.

Anyway this is the reason this might be a big deal:

"Unlike many scientific breakthroughs that require highly controlled laboratory environments, this new transistor functioned in everyday, ambient conditions. ... Their next goal is to develop a version of the transistor that operates using standard, commercially available lasers, making the technology more accessible to industry partners."

 

[leonavis]

Heisenberg's

• ΔE Δt ~ h

works out to about 1 eV for 10^-15s so I would imagine this is a fundamental limit for electrons (the Planck time would imply much larger energies)

hmm lets see. h is ~4.14feVs, so h/Δt would equal 4.14eV.

Usually that's ΔE Δt >= h/2pi though. But yeah, that would suggest that that's close to the ceiling. Taking the reduced Planck constant 1PHz would already be unstable...

They exposed it to a laser switching on and off at an astonishing rate of 638 attoseconds

from the article. I really wonder how they did that. Lucky for us, one can simply download the study.

So that'll be some hard reading...

Edit again: Found something from Wiki - there's a whole branch of physics dealing with attosecond movement of particles. So cool.

 

[Space Lynx]

Anyway this is the reason this might be a big deal:

"Unlike many scientific breakthroughs that require highly controlled laboratory environments, this new transistor functioned in everyday, ambient conditions. ... Their next goal is to develop a version of the transistor that operates using standard, commercially available lasers, making the technology more accessible to industry partners."

Yeah, another reason I made the article "help me understand" because this isn't hypothetical like a lot of the articles I read, so kind of neat indeed.