Intel falls further behind as ~0nm transistors created

[asus killer]

4 minutes ago, i_build_nanosuits said:

This news has nothing to do with intel so why is intel even in the title?

it's not ok to kick someone when they are down. Poor Intel 

[Master Disaster]

1 hour ago, Phentos said:

For now.

No, forever. Quantum computers are great at very specific workloads but break down entirely under others.

 

Even the top experts all agree that's it's incredibly unlikely a QC will ever replace a silicon computer. Think of it like an oven and a microwave, the 2 things do very similar workloads but they do it in very different ways and one doesn't cancel out the other. You wouldn't try to roast a chicken in a microwave and you won't be gaming on a QC.

 

[TacoSenpai]

3 minutes ago, Master Disaster said:

 Think of it like an oven and a microwave, the 2 things do very similar workloads but they do it in very different ways and one doesn't cancel out the other. You wouldn't try to roast a chicken in a microwave and you won't be gaming on a QC.

 

Case in point:

 

 

[Taf the Ghost]

1 hour ago, Trixanity said:

Lab results don't count all that much. We have graphene and super batteries in labs all around the world and have had that for a good number of years in one form or another; yet here we are.

Carbon Nanotubes have been "the next big thing" since before Duke Nukem Forever was vaporware. It's still into the realm of Fusion Power.

 

As for CPUs, I thing to keep in mind is that there is something like 5 or 6 different generations already well in the works. The actual structure of the transistor itself can shrink, while the spacing won't necessarily. There's also room for different ISAs in the future, and, potentially, multi-state but not Quantum computers. There's still a long way to go on the CPU engineering train.

[Jtalk4456]

58 minutes ago, VegetableStu said:

• is it drop Linus proof?

I corrected your statement

[rcmaehl]

3 minutes ago, LordOTaco said:

Case in point:

 

 

This reminds me I'm behind on an episode of S;G 0 Be back in like 45 minutes.

[Sniperfox47]

1 hour ago, Sauron said:

Might as well use picometres... 0.2nm is 200pm.

200pm small transistors doesn't sound nearly as impressive as 0nm small transistors though. 200 is a bigger number than 10/7, didn't you know? xP

[aezakmi]

 

[AngryBeaver]

I agree with many others here. This is a good POC type thing proving you can create a single atom transistor, but the problem becomes when you try to pack them on the chip. Even if they are low energy being that close together arcing and electrical leakage in that scenario are going to be big limiters. I mean what is the point of a .2nm transistor if you need 3nm of space to keep them from arcing. I mean we already see issues with this now at 14 and 12 nm. This is one of the main reasons that ln2 is so effective in high OC's btw. It increases the amount of voltage you can apply before these things happen. So at .2nm the amount of voltage would need to be incredibly low... which I just don't see being feesible at room temp.

[Nineshadow]

2 hours ago, Trixanity said:

Lab results don't count all that much. We have graphene and super batteries in labs all around the world and have had that for a good number of years in one form or another; yet here we are.

They do count. Just not immediately. 

[Trixanity]

38 minutes ago, Nineshadow said:

They do count. Just not immediately. 

Sometimes they do, sometimes they don't.

 

And as I said many things don't leave the lab so it's fairly pointless to discuss it as marketable until there's concrete evidence it is. That doesn't mean it isn't news but this has little relevance for Intel or any other foundry for that matter. It's no threat.

[i_build_nanosuits]

1 hour ago, asus killer said:

it's not ok to kick someone when they are down. Poor Intel 

i don't get why people say stuff like that...intel still produce the fastest and strongest CPU's on the planet at all levels...?

[rcmaehl]

1 minute ago, i_build_nanosuits said:

i don't get why people say stuff like that...intel still produce the fastest and strongest CPU's on the planet at all levels...?

 

No so fast bucko

[asus killer]

4 minutes ago, i_build_nanosuits said:

i don't get why people say stuff like that...intel still produce the fastest and strongest CPU's on the planet at all levels...?

i guess there isn't a CPI to rule them all, but for many cases you can now have a better and cheaper alternative in AMD depending on your needs. And with them not being able to go past 14nm anytime soon it may get worst for them.

[cj09beira]

15 minutes ago, i_build_nanosuits said:

i don't get why people say stuff like that...intel still produce the fastest and strongest CPU's on the planet at all levels...?

pretty sure amd wins with their 32 core beast against intel's 28 core at least in some scenarios 

[i_build_nanosuits]

15 minutes ago, rcmaehl said:

 

No so fast bucko

NANOSUITS Says it's next processors will outperfom ARM laptop chips by 259% (you get my point)

11 minutes ago, asus killer said:

i guess there isn't a CPI to rule them all, but for many cases you can now have a better and cheaper alternative in AMD depending on your needs. And with them not being able to go past 14nm anytime soon it may get worst for them.

they are cheaper, maybe better performance per dollar...but for absolute performance, top-end performance, ''if you can spend a bit more and not get AMD'' performance, they are still better, top dog...

''may get worst for them'' intel has unlimited budget, they have fuck you money, they will get out of that hole in no time flat if need be.

[i_build_nanosuits]

Just now, cj09beira said:

pretty sure amd wins with their 32 core beast against intel's 28 core at least in some scenarios 

28 intel cores > 32 AMD cores...in all scenarios...heck 6 intel cores is better than 8 AMD cores in nearly all scenarios already...

You guys lost the plot somewhere

[cj09beira]

5 minutes ago, i_build_nanosuits said:

28 intel cores > 32 AMD cores...in all scenarios...heck 6 intel cores is better than 8 AMD cores in nearly all scenarios already...

You guys lost the plot somewhere

https://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade/16

you lost your plot somewhere along the way

[Blademaster91]

6 minutes ago, cj09beira said:

https://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade/16

you lost your plot somewhere along the way

A synthetic benchmark says a cpu with more threads is faster, wow no way?

For real world server workload Intel is still faster, https://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade/18

Not sure why this thread turned into another AMD vs. Intel war, other than the clickbait headline that has nothing to do with the source articles.

[cj09beira]

2 minutes ago, Blademaster91 said:

A synthetic benchmark says a cpu with more threads is faster, wow no way?

For real world server workload Intel is still faster, https://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade/18

Not sure why this thread turned into another AMD vs. Intel war, other than the clickbait headline that has nothing to do with the source articles.

jumping straight to the only test where epyc does badly, biased much see 17 19 both show great results 

[FloRolf]

How do you even make a thing from one single atom? How do you access it? Where does stuff go in and out? Where do you even store a single atom and still have it in your view at all times knowing what exactly it does? 

 

Se Germans... 

[S w a t s o n]

I'm calling bullshit here. This is how they explain it works:

Quote

In Advanced Materials, the KIT researchers present the transistor that reaches the limits of miniaturization. The scientists produced two minute metallic contacts. Between them, there is a gap as wide as a single metal atom.

 

“By an electric control pulse, we position a single silver atom into this gap and close the circuit,” Professor Thomas Schimmel explains. “When the silver atom is removed again, the circuit is interrupted.”

Clearly the picture used in the article is not this because it's fuck hueg compared to 1 atom. I guess so they could show it? They could have at least provided electron scanning microscope pictures or something.

I'm still in for light based cpus in the long term

[Sauron]

3 hours ago, Sniperfox47 said:

200pm small transistors doesn't sound nearly as impressive as 0nm small transistors though. 200 is a bigger number than 10/7, didn't you know? xP

I heard that bigger is better 

[mr moose]

This is not the first time scientists have played around with single atoms:

 

 

 

[Arika]

The click-baitest of titles.