GlobalFoundries licenses Samsung process tech, grants AMD access to FinFETs

[TheRealDeal]

http://techreport.com/news/26336/gl...ung-process-tech-grants-amd-access-to-finfets

http://www.brightsideofnews.com/2014/04/17/samsung-and-globalfoundries-collaborate-on-14nm-finfet/

Interesting stuff. Could give amd a boost in Gpu's and Cpu's. Sounds like this kind of process tec could close the gap that Intel currently enjoy.

Perhaps the biggest beneficiary of this licensing deal will be AMD, who gains access to a capable 14-nm FinFET tech for the production of its future CPUs and SOCs. AMD's Kaveri APUs actually ran slower than their predecessors after transitioning from GloFo's 32-nm SOI process to 28-nm bulk silicon. Meanwhile, AMD competes most directly with the process tech leader, Intel.

The tech being licensed is the next major step forward, a 14-nm process that uses a new transistor structure known as FinFETs or, as Intel calls them, tri-gate transistors. Intel made the transition to FinFETs at its 22-nm process node and saw some fairly dramatic benefits in terms of switching speed and power efficiency (which are often two sides of the same coin in process tech discussions). Other firms in the industry have struggled to reap the usual power and speed benefits when moving below 28-nm process geometries without FinFETs. Most of these firms have scheduled FinFETs for the 16- or 14-nm nodes, with several delays. Meanwhile, even Intel has delayed its 14-nm process due to technical difficulties.

Samsung appears to have succeeded in developing a capable 14-nm FinFET process. Here's how the press release describes the technology being licensed:
Developed by Samsung and licensed to GLOBALFOUNDRIES, the 14nm FinFET process is based on a technology platform that has already gained traction as the leading choice for high-volume, power-efficient system-on-chip (SoC) designs. The platform taps the benefits of three-dimensional, fully depleted FinFET transistors to overcome the limitations of planar transistor technology, enabling up to 20 percent higher speed, 35 percent less power and 15 percent area scaling over industry 20nm planar technology.

"This unprecedented collaboration will result in a global capacity footprint for 14nm FinFET technology that provides AMD with enhanced capabilities to bring our innovative IP into silicon on leading-edge technologies," said Lisa Su, senior vice president and general manager of Global Business Units at AMD. "The work that GLOBALFOUNDRIES and Samsung are doing together will help AMD deliver our next generation of groundbreaking products with new levels of processing and graphics capabilities to devices ranging from low-power mobile devices, to next-generation dense servers to high-performance embedded solutions."

 

[LtMatt]

http://techreport.com/news/26336/gl...ung-process-tech-grants-amd-access-to-finfets

Interesting stuff. Could give amd a boost in Gpu's and Cpu's.

Any chance of a TL : DR?

If this means more AMD gpu's get Samsung gpu memory then thats great. Gives the Nvidia cards a nice performance advantage over Elpida and to a lesser degree Hynix. If it doesn't mean this, then ill take that TL : DR now plox.

EDIT

Another link here RD if you want to add to the op.

http://www.brightsideofnews.com/2014/04/17/samsung-and-globalfoundries-collaborate-on-14nm-finfet/

 

[TheRealDeal]

Any chance of a TL : DR?

If this means more AMD gpu's get Samsung gpu memory then thats great. Gives the Nvidia cards a nice performance advantage over Elpida and to a lesser degree Hynix. If it doesn't mean this, then ill take that TL : DR now plox.

EDIT

Another link here RD if you want to add to the op.

http://www.brightsideofnews.com/2014/04/17/samsung-and-globalfoundries-collaborate-on-14nm-finfet/

Done took out a few bits that i found interesting.

 

[Gregster]

Should be in the CPU section.

 

[TheRealDeal]

Should be in the CPU section.

You could be right but seen some graphics in there so not sure if it will just be apu's and cpu's.

 

[Gregster]

You could be right but seen some graphics in there so not sure if it will just be apu's and cpu's.

Yer it is. It is in response to Intel. Still good info and good to see Better feedback in the CPU section I reckon though

 

[TheRealDeal]

Yer it is. It is in response to Intel. Still good info and good to see Better feedback in the CPU section I reckon though

Unfortunately it's already been posted in the cpu section so no point moving it there.

Apu's still have graphics chips lol and by the time we get to 14nm they could get pretty powerful in there own right.

 

[drunkenmaster]

Bulk was not the only option at 28nm but the best option for the gpu as SOI + gpu wasn't the best combination. The biggest issue to date for APUs is that GPUs and CPUs generally want different types of processes. Kaveri's CPU would run faster and use less power at say 4Ghz if it was on SOI or a more tuned for higher performance process, but then the GPU would struggle on that process.

With 20/14nm we should see interposers start be used pretty soon which means when more process options are available you can make the gpu part on say bulk, the cpu part on potentially 14nm SOI then stick them together on an interposer. It's like the old dual/quad cores that were just two chips taped together, but stuck together in a far smarter effectively same die connection speed. Then you can get the cpu speed/leakage performance you want and the gpu characteristics you want, and make smaller dies with higher yields and then just stick them together. It also lets you do the same with stacked memory, make it all separate and then stick them together.

So realistically APU's improvement in the next few years will come from making them smarter, not better. Making individual parts on the best process available for that part then sticking them together will result in a better chip than just dropping the process node and using an in between process that isn't optimised for any part of the chip.

We'll likely see from using targeted processes for each part, CPU clock speeds raise again with power actually dropping at the same time. We'll see significantly denser gpu's meaning a more powerful gpu in the same die space or using a smaller die and getting better power usage. Then sticking memory on package will mean improved bandwidth and performance.

 

[humbug]

Any chance of a TL : DR?

If this means more AMD gpu's get Samsung gpu memory then thats great. Gives the Nvidia cards a nice performance advantage over Elpida and to a lesser degree Hynix. If it doesn't mean this, then ill take that TL : DR now plox.

EDIT

Another link here RD if you want to add to the op.

http://www.brightsideofnews.com/2014/04/17/samsung-and-globalfoundries-collaborate-on-14nm-finfet/

Who cares about that? AMD will get 14nm GPUs by early 2015

 

[LtMatt]

Who cares about that? AMD will get 14nm GPUs by early 2015

So ive heard. Does this mean AMD will skip 20mn and go straight to 14mn?

 

[humbug]

So ive heard. Does this mean AMD will skip 20mn and go straight to 14mn?

If they end up on Discrete GPUs, its all speculation, you know how AMD in competition with Nvidia don't like to show their hand...

There is no point in making 20nm GPU when GloFo are up and running with 14nm in less than a year.

AMD have just signed a $1.2bn contract with GloFo, thats a big commitment, something is going on.

 

[Calin Banc]

In other news - http://www.dsogaming.com/news/cryte...-that-real-time-graphics-will-hit-cg-quality/

Probably 2 of the new r390x would be capable of something like that.

 

[humbug]

In other news - http://www.dsogaming.com/news/cryte...-that-real-time-graphics-will-hit-cg-quality/

Probably 2 of the new r390x would be capable of something like that.

Mantle is available right now on existing GCN GPU's

Your in the wrong thread, http://forums.overclockers.co.uk/showthread.php?t=18589166&highlight=AMD+Mantle

 

[Calin Banc]

I was just pointing out that Mantle will help getting there, but still requires some power and most likely won't happen on current gen hardware, hence this thread. AMD will have quite the launch demo in that case.