Intel, Samsung, Toshiba Team For 10nm Chips

[Marcus Yam]

These three big guns are working on tiny chips.

Intel, Samsung, Toshiba Team For 10nm Chips : Read more

 

[ALANMAN]

Awesome, can't wait!

 

[halodude23]

If you can provide smaller chips, why not lower the prices of the 25 nm hard drives and 45 nm e.t.c?

 

[Haserath]

We've gotten to the point that companies have to start working together to make the transistors smaller. That must mean we're at the point that it is getting close to the end for the transistor to get smaller, but maybe they'll come up with a breakthrough that will keep us going for some years to come.

 

[lauxenburg]

H0LY BALLS. That's small. That's like half the size of a virus.

 

[lauxenburg]

[citation][nom]Haserath[/nom]We've gotten to the point that companies have to start working together to make the transistors smaller. That must mean we're at the point that it is getting close to the end for the transistor to get smaller, but maybe they'll come up with a breakthrough that will keep us going for some years to come.[/citation]

Honestly we felt the same way like 10 years ago. There were a lot of boundaries we thought we'd run into but we got past them. However, it's true that every few years we can't just expect a 10-15nm change. Advancing to a new manufacturing tech will take much longer unless we figure out some other way to do things, like you said.

 

[marraco]

With each new generation, fab costs increase exponentially.

Not so far in the future, the world will afford only one fab company, and AMD, Intel and all the other would be forced to share his costs.

 

[Albyint]

It would be interesting to see a breakthrough that goes past transistor size, something akin plastic magnets and such. But 10nm is a good start lol.

 

[IMAC]

One thing is for sure AMD is out of their list. Well hope AMD forms it´s own alliance and both race for the 10nm goal.

 

[g00ey]

Well, I can't wait till we get picometer circuits...

 

[the_krasno]

That is close to the barrier where quantum physics screw with how a transistor works.

Which means that at this point it will be more productive to work on a viable quantum chip.

Just saying...

 

[bison88]

Aren't they going to have to start looking for another way? I mean they've been talking about it for years. The escalating costs cannot be manageable in the future. Eventually there is going to be a tap out of microprocessors and chip shrinkage unless those big guys are holding onto some secret advancements that are game changing to try and milk what exists now. I have my doubt personally, but I fear another equivalence of a Gigahertz dustbowl is imminent.

 

[lauxenburg]

[citation][nom]g00ey[/nom]Well, I can't wait till we get picometer circuits...[/citation]

Highly unlikely. There will be different technologies aside from simply making the transistors smaller.

 

[TomTomz]

[citation][nom]IMAC[/nom]One thing is for sure AMD is out of their list. Well hope AMD forms it´s own alliance and both race for the 10nm goal.[/citation]
Then one of them will build GlaDOS to speed up production, and you know where they leads.

 

[jimmysmitty]

[citation][nom]IMAC[/nom]One thing is for sure AMD is out of their list. Well hope AMD forms it´s own alliance and both race for the 10nm goal.[/citation]

Thats is probably because AMD no longer has any FABs nor do they put any funds towards R&D of process tech. All their R&D goes to the arch and they have GF who will probably work closley with IBM for process tech.

As for the first there, it will no doubt be Intel and its consortium. Intel already has working 22nm so now they will just work on enhancing the process tech and ramping it out. So R&D will focus on 10nm.

GF and IBM are still pushing on 32nm.

Still it will be interesting to see if they find something new to push past the 10nm barrier or maybe a new material that will drop silicon all together.

 

[drutort]

the transistor size to me is the same as the Ghz race... I think pretty soon companies will have to start to be more conservative with there designs and design new things in a smart approach and not hoping that oh well we can just thrown in x amount of transistors and it will work, now they will have to work harder and smarter.

Remember the Ghz race? same thing with the multi cores... you can just keep throwing more and more cores at the cpu's pretty soon new real tech is going to be the major roll player in making things faster and less power.

going smaller is only one way of saving power

 

[beckstrom12]

[citation][nom]jimmysmitty[/nom]Thats is probably because AMD no longer has any FABs nor do they put any funds towards R&D of process tech. All their R&D goes to the arch and they have GF who will probably work closley with IBM for process tech.As for the first there, it will no doubt be Intel and its consortium. Intel already has working 22nm so now they will just work on enhancing the process tech and ramping it out. So R&D will focus on 10nm. GF and IBM are still pushing on 32nm.Still it will be interesting to see if they find something new to push past the 10nm barrier or maybe a new material that will drop silicon all together.[/citation]

They do its called graphene look it up

 

[ikefu]

At 22nm a transistor is only about 200 atoms wide. Therefore 10nm would be only about 91 atoms across.

Much smaller and you start to lose structural integrity and run into other quantum road blocks. I think the next major push (which is already starting) is efficiency. How many instructions per clock cycle can you push with the same die size?

 

[Pyroflea]

And soon, we'll approach the pm scale

 

[ares1214]

This is big, actually quite small, but would be a massive accomplishment. Just getting 3 major companies in the same room is a pretty sizable achievement in its own right! :lol: But this is about as small as silicon will let them go. Innovation time!

 

[saturnus]

What has been said above about 10nm being close to the limit is true. The actual limit is around 8nm whereafter quantum physics makes it impossible to shrink current technology any further. So it really comes as no surprise that it takes basically all the industry giant to get to 10nm. I wouldn't be surprised if Intel even asked AMD to join because there's some AMD/IBM patents that they need for it to succeed at all.

Luckily, we don't have to shrink current technology any further, we just have to shift technology to memristors and layered metal diodes instead. All emergenging technologies coming out in the next couple of years.

Going to 10nm is in my opinion just as a big of a goal as Intels previous 10 GHz goal. It's just not economically viable or physically possible in the foreseeable future (the next 10-25 years).

 

[bombat1994]

the one thing that i must state that the one thing holding back the cpu industry is not, transistor sizes or cpu speed. its heat. cpu performance could be instantly doubled if heat was not a question. look at the amd phenoms under liquid nitrogen, 7.0GHZ. if a mechanically viable material that was much more capable at high temperatures or if it didnt care for heat at all, we would see great improvements in everything.


Look at the car industry, toyota has started to really push there ceramic engine. the ceramic engine has a much higher heat compacity than steel nearly double. car performance and efficiency would be greatly improved, because not only does the engine less energy through heat transfer, but the need for a cooling system is all but eliminated. so a radiator is not required.

if this happened in a computer the thing could be near silent. personally i thing that fans, and hardrives are the loudest component of the computer. so a laptop would be silent if it had a ssd then the noise problem is solved.


but it is still interesting how close they can get to the quantom limitations

 

[iam2thecrowe]

I heard a while back HP working on "memristors" that can hold multiple states (more than just high or low voltage). I want to see more work on something like that rather than flogging a dead horse (transistors). Also lasers! Lasers are cool.

 

[alidan]

[citation][nom]saturnus[/nom]What has been said above about 10nm being close to the limit is true. The actual limit is around 8nm whereafter quantum physics makes it impossible to shrink current technology any further. So it really comes as no surprise that it takes basically all the industry giant to get to 10nm. I wouldn't be surprised if Intel even asked AMD to join because there's some AMD/IBM patents that they need for it to succeed at all.Luckily, we don't have to shrink current technology any further, we just have to shift technology to memristors and layered metal diodes instead. All emergenging technologies coming out in the next couple of years.Going to 10nm is in my opinion just as a big of a goal as Intels previous 10 GHz goal. It's just not economically viable or physically possible in the foreseeable future (the next 10-25 years).[/citation]

i thought the real limit was 6nm, o well.

also, silicon is about dead, at least as close to dead as it can get.
it heats up to fast, gets to hot and aside from being small doesn't offer anything significant in return.

graphene (sp) will be the future, now correct me if im wrong, but isnt that crap possible to go into the few hundreds of ghz with air cooling?

i mean even if they went back to 60 or 90 nm process to use it, the ghz gain would be significant enough to offset the size difference.

 

[JamesSneed]

This is a huge challange. I would have expected a big push for a newer fab technique like 3D / EUV lithography or the sorts. Or maybee thats where this is going they just are not making mention of it. I know im ready for 1x nm 3D lithography to take hold for NAND chips so I can finally aford a decent sized SSD. I have to assume that this 10nm push is for CPU production not NAND flash but it could be looking at the team of companies.