Intel Makes 22nm 3-D Tri-Gate Tech for Ivy Bridge

[vickybat]

Intel Corporation today announced transistors using a three-dimensional structure will be put into high-volume manufacturing. Intel introduced a revolutionary 3-D transistor design called Tri-Gate, first disclosed by Intel in 2002, into high-volume manufacturing at the 22-nanometer node in Ivy Bridge.

Source

 

[Gaurav Bhattacharjee]

Yep. Intel will mass produce 3D transistors, allowing processors down to 10nm by 2015.
2D transistors were only going to allow a shrink down to 16nm.

Intel says we'll see the new technology first in its 22nm Ivy Bridge CPUs, going into mass production in the second half of the year, and it's planning 14nm chips in 2013 and 10nm chips in 2015.

Source

 

[abhidev]

Thats awesome!!!!!!!!!! Also heard about the news that intel will be entering the mobile and tablet market soon and also has decided to partner with both google and microsoft......thats a win-win situation for Intel!!!!

 

[bhushan2k]

^^yup...that's good..cortex, ARM and other will be worried now...

 

[saswat23]

First 65nm
then 45nm
then 32nm
then 22nm
then 15nm
then 10nm
..
..
..
..
Very soon i hope there will 0nm chips. I mean zero-nm chips.....

 

[abhidev]

the unit then might go to 0.50nm or so....

 

[saswat23]

means that wont be visible. So how are we gonna place the proccy in the socket..???

 

[coderunknown]

^^yup...that's good..cortex, ARM and other will be worried now...

i don't think so. with Intel's high price & also cause it doesn't have as much experience as Qualcomm, Samsung & Texas Instruments. it'll be a hard fight.

also cortex is nothing but ARM name for one of the cores it developed. ARM just designs the processor. further enhancement & manufacturing is done by others.

First 65nm
then 45nm
then 32nm
then 22nm
then 15nm
then 10nm
..
..
..
..
Very soon i hope there will 0nm chips. I mean zero-nm chips.....

after 22 it should be 16 & 11nm. well, it'll be interesting as how far can these shrinks in manufacturing process can go.

means that wont be visible. So how are we gonna place the proccy in the socket..???

each processor comes with only 1 transistor?

 

[vickybat]

i don't think so. with Intel's high price & also cause it doesn't have as much experience as Qualcomm, Samsung & Texas Instruments. it'll be a hard fight.

Buddy don't underestimate intel. You have no idea about the potential of this company. Its R&D centre is like area51 i.e totally confidential and who knows what these guys constantly plan there.

They even have research centre in siberia (total icelands). If intel makes a move , it will be strong. Who knows, their quadcore soc offerings might be efficient and architecturally superior to their arm counterparts.

 

[saswat23]

sam,
what i meant is

the unit then might go to 0.50nm or so....

So, if the proccy is 0.5nm

means that wont be visible. So how are we gonna place the proccy in the socket..???

What i actually meant was that if the proccys will be 0.5nm then how would we put them in the socket. They would be invisible to naked eyes.

So, very soon we would be using microscopes to put these proccys in their sockets.

 

[coderunknown]

Buddy don't underestimate intel.

i know Intel's potential but if they price a normal SOC 2-3 times higher than its competitors (as they are doing now), clients will surely look for the other source. yes, tablets with Windows7/8 will run mainly on Intel processors as even if ARM is supported, writing new software just for tablets, its madness.

You have no idea about the potential of this company. Its R&D centre is like area51 i.e totally confidential and who knows what these guys constantly plan there.

well thats kind of true. they don't leak info like AMD Nvidia do with their slides.

They even have research centre in siberia (total icelands). If intel makes a move , it will be strong. Who knows, their quadcore soc offerings might be efficient and architecturally superior to their arm counterparts.

but do remember its not a 1 vs 1 fight. its like jumping into a pool full of sharks. Intel maybe no.1 player in desktop/laptop/server but others are the master of tiny processing. anyway, future computing is going to be damn interesting

sam,
what i meant is

So, if the proccy is 0.5nm

What i actually meant was that if the proccys will be 0.5nm then how would we put them in the socket. They would be invisible to naked eyes.

So, very soon we would be using microscopes to put these proccys in their sockets.

see, currently the processor we use. their actual core is really small. around 1/5th the total size of the processor. but they are manufactured using 32nm (not visible to naked eye). but as they pack close to or over 1B transistors, they acquire a size of 75-80sq mm. so they packed inside the processor & visible. and so we can handle them easily.

so think 2015. 11nm. 50-60B transistors. so basically size should be same as the die of now & i feel, we'll have the same processors until they are integrated into the PCB permanently, making it more like a SOC.

hope i answered it properly.

 

[bhushan2k]

Buddy don't underestimate intel. You have no idea about the potential of this company. Its R&D centre is like area51 i.e totally confidential and who knows what these guys constantly plan there.

They even have research centre in siberia (total icelands). If intel makes a move , it will be strong. Who knows, their quadcore soc offerings might be efficient and architecturally superior to their arm counterparts.

Exactly. Completely agree..

@sam,

If all are casual sharks then intel will be the blue whale.. but to be honest, intel is true giant..and they also have experience of mobile proccy i think..xscale for blackberry i guess..so if they take a serious note for mobile computing then no matter how many rivals are there, intel will let them on one side..

 

[Joker]

@saswat: currently u dont know how these chps are made. google/youtube to find out more.

intel's low power atom are as behind as amd brazos as much as amd phenom II is behind intel sandy bridge. x86 will never be able to compete with ARM as far as low power designs are concerned.

x86 has a long long long way to go before it competes with ARM processors. ARM has been designed from groundup for low power computing. x86 is powerful but not power efficient.

will intel license ARM designs?? NEVER. too stubborn for that. it is like accepting defeat.

currently there were rumours about AMD taking over ARM (which they refused but who knows). i cant see them taking over ARM, but i can definitely see them licensing ARM proc designs and pairing them with their gfx cores. ARM is going to conduct a keynote at amd fusion consortium.

no i dont want amd to take over ARM. that will be monoply as they will have everything at their disposal. ARM is a cheap company to buy btw. and i hate monoply like intel. for 8 years 1998-2006 amd had superior processors which didnt sell because of intel's monoply measures with OEMs.

 

[vickybat]

intel's low power atom are as behind as amd brazos as much as amd phenom II is behind intel sandy bridge. x86 will never be able to compete with ARM as far as low power designs are concerned.

x86 has a long long long way to go before it competes with ARM processors. ARM has been designed from groundup for low power computing. x86 is powerful but not power efficient.

will intel license ARM designs?? NEVER. too stubborn for that. it is like accepting defeat.

currently there were rumours about AMD taking over ARM (which they refused but who knows). i cant see them taking over ARM, but i can definitely see them licensing ARM proc designs and pairing them with their gfx cores. ARM is going to conduct a keynote at amd fusion consortium.

Don't go by history. X86 was never competitive in mobile computing does not mean that it won't pose a threat in the upcoming times. Intel doesn't have to license ARM designs. They have the potential to come up with something designed from groundup.

You must have realised by reading the article in the first post. That is something revolutionary. Besides we will be witnessing a 2nd generation atom successor. Who knows how power efficient and power house that will turn out to be?

 

[rajan1311]

my uncle does r&d work at intel...he working on an 80 core cpu due to launch n 2012...

who underestimated them? they have the $$, fabs and the talent

btw, AMD used to own a company called "qualcomm" (rather, its now known as qualcomm) , but they sold it ....so i dont think they gonna acquire any soon..

 

[Joker]

Don't go by history. X86 was never competitive in mobile computing does not mean that it won't pose a threat in the upcoming times. Intel doesn't have to license ARM designs. They have the potential to come up with something designed from groundup.

You must have realised by reading the article in the first post. That is something revolutionary. Besides we will be witnessing a 2nd generation atom successor. Who knows how power efficient and power house that will turn out to be?

x86 (atom and brazos) will never be as power efficient as ARM. why?

if u want to compete with ARM...design something power efficient from scratch rather than optimizing x86 as an aftermath. ARM is designed for performance & power efficiency from the beginning itself. unlike x86. x86 is designed for performance.

just like windows will never be as secure as linux. windows treats security as an aftermath unlike linux which is designed for security from the bottom.

 

[skeletor]

btw, AMD used to own a company called "qualcomm" (rather, its now known as qualcomm) , but they sold it ....so i dont think they gonna acquire any soon..

hmm not true. Qualcomm has always been a separate company. AMD had sold off their SoC/mobile division to Qualcomm in 2009 as it desperately needed funds.

 

[rajan1311]

^my bad...had read the thing some time back....anyways, i am sure they wish they had not done that, with the smartphone bizz booming, they could have cashed in...

 

[abhidev]

Also to give an headsup....IMacs soon will be available with Z68 chipsets enabling better performance due to ssd caching and with amd gpu......howzzatt!!!!!!

 

[coderunknown]

@sam,

If all are casual sharks then intel will be the blue whale.. but to be honest, intel is true giant..and they also have experience of mobile proccy i think..xscale for blackberry i guess..so if they take a serious note for mobile computing then no matter how many rivals are there, intel will let them on one side..

its a long time to go before Intel can put real pressure on other. yes, Intel got the talent, the fabs & the $$$ but you can't win the world just cause you got it all. in OEM, price plays a big role & if they don't price their mobile SOC accordingly, they'll hurt themselves.

about X-scale, this is my story:

*i56.tinypic.com/11kjtl2.jpg
chip is fast & can even be overclocked.

intel's low power atom are as behind as amd brazos as much as amd phenom II is behind intel sandy bridge. x86 will never be able to compete with ARM as far as low power designs are concerned.

but Intel makes really power efficient chips when the size is big (Atom>Brazos & Core i3/i5/i7 >>> Phenom2/Athlon2). will be interesting to see how can they compete when the size of chip is really really small.

x86 has a long long long way to go before it competes with ARM processors. ARM has been designed from groundup for low power computing. x86 is powerful but not power efficient.

true & its not like only Intel is the transistor boss. who knows what others are using & planning to welcome Intel. but taking Intel lightly, can be the last bad thing others can do.

will intel license ARM designs?? NEVER. too stubborn for that. it is like accepting defeat.

they done it in the past with the Xscale thing. thing wasn't a success & so sold it to Marvell.

currently there were rumours about AMD taking over ARM (which they refused but who knows). i cant see them taking over ARM

1st they need the money. for now fight Intel & gain back the market share then they can think of jumping into the mobile battlefield.

ARM is going to conduct a keynote at amd fusion consortium.

AMD can't license their X86 license to ARM as the license is given to AMD by Intel. but the opposite is highly likely, AMD getting hold of a Cortex A15 license.

no i dont want amd to take over ARM. that will be monoply as they will have everything at their disposal. ARM is a cheap company to buy btw.

if they try a move like that, one of the semiconductor giants may try to get hold of AMD itself. & if this was possible, why Qualcomm/Intel/Samsung already not tried it? infact Intel decided to make their own architecture wasting billions of $. doesn't makes sense.