[AT] Samsung Announces 14nm FinFET for Exynos 7

[witeken]

http://www.eetimes.com/document.asp?doc_id=1324643&page_number=2

Pretty strange that Intel only advertised its low density SRAM (at IDF, IM, etc.).

 

[bronxzv]

AMD is still up in the air. Nobody knows what they're doing; heck, even they don't know what they're doing.

AMD will use the Samsung 14nm process licensed to GF isn't it ?
http://www.globalfoundries.com/news...ering-of-14nm-finfet-semiconductor-technology

 

[Fjodor2001]

I wonder when GF will start releasing chips based on Samsung's 14 nm process that they have licensed. If it's not until AMD Zen, that will be more than a year later than the Samsung chips mentioned in the OP...

 

[geoxile]

They already are.

Last time I checked those were just rumors

 

[krumme]

I wonder when GF will start releasing chips based on Samsung's 14 nm process that they have licensed. If it's not until AMD Zen, that will be more than a year later than the Samsung chips mentioned in the OP...

A year at the earliest. Didnt amd ceo say they would only use mature nodes?
But amd is not that interesting because they lack all financial muscles to back it up anyway. We have qcom and samsung own arch comming.
Look at the market cap of the companies and the tremendous amount and acceleration of r&d for samsung.
Samsung next own gen it worth looking for. So is their 10nm and 10nm midnode euv. That can shake things up imo. Not zen.

 

[Arachnotronic]

Last time I checked those were just rumors

No, Samsung straight up said at their analyst day that they're doing a custom CPU core.

 

[AtenRa]

Samsung is an HSA founder, they already designing their own ARM cpus.

 

[ShintaiDK]

Samsung is an HSA founder, they already designing their own ARM cpus.

Unless you can document they specificly design HSA capable SoCs. Then the HSA comment is completely irrelevant.

 

[stuff_me_good]

What is all this fuzz about FINFETs? What are the real benefits if you would for example compare 14nm vs 14nm FINFET?

While reading these forums about finfets, it fells like it's the second coming of jesus.

 

[IntelUser2000]

What is all this fuzz about FINFETs? What are the real benefits if you would for example compare 14nm vs 14nm FINFET?

While marketing wants you to think otherwise, the reality is that technologies like FinFETs are merely enablers for technology development to continue at previous pace.

 

[III-V]

While marketing wants you to think otherwise, the reality is that technologies like FinFETs are merely enablers for technology development to continue at previous pace.

Hardly. While it's certainly an enabler, there are plenty of technical benefits unique to FinFETs and other "next-generation" transistor technologies.

Perhaps most importantly, FinFETs have a much better sub-threshold slope than traditional planar FETs. This is a huge benefit to low power devices. Minimum clock speeds at a fixed, low voltage can be raised drastically, or more likely, minimum voltage can be be lowered drastically.

Idle and low-load battery life goes through the roof. Very, very important for smartphones and the like, and very good for other embedded applications. It opens up a lot of new avenues -- tunnel FETs will be even better on this front -- although they're pretty bad for high performance, you can basically make self-powered devices a reality.

FinFET's superiority can also be applied to the high performance side of the equation -- FinFETs can achieve significantly improved Idsats at the same voltage, but unfortunately Intel and everyone else is seemingly uninterested in that side of the spectrum.

 

[witeken]

While marketing wants you to think otherwise, the reality is that technologies like FinFETs are merely enablers for technology development to continue at previous pace.

Well put. FinFET enables the continuation of Dennard-like scaling for power consumption. The sub-threshold slope that III-V's talking about is a benefit you'll only get once, and not particularly a spectacular one (it doesn't blow traditional Dennard scaling out of the water, but it certainly is a nice benefit that Intel should have capitalized more, earlier). And if we're lucky, we might even see a slight resurrection of performance scaling (but again, that's just enabling technology to move forward at current pace). Or to put another way: it's no surprise that we've only seen transistor innovations after Dennard scaling ended. If Dennard's Law was still alive we still wouldn't even have got strained silicon, in all likeliness. So your using of the word enabling seems quite correct to me.

Below: Edit too slow : (

 

[III-V]

Well put. FinFET enables the continuation of Dennard-like scaling for power consumption.

Surprised you're agreeing with him. FinFETs do much more than that.

 

[ThatBuzzkiller]

They should have called it 20FF. Its even worse than TSMC 16FF.

Will be interesting to see what volumes they can supply and when. Lets hope it wont be limited to the Korean market.

I wonder how much of their finfet process they stole from TSMC.

That's a lie ...

 

[podspi]

Surprised you're agreeing with him. FinFETs do much more than that.

I am far from an expert, but my impression is that FinFETs are very impressive from a technical point of view, but they are just the latest in an industry of "revolutions". We're still talking lotsa (technical term) transistors switching at very high speeds that we've seen years before, they now are just smaller and use less power.

(For some definitions of 'just' of course )

I feel bad because in a lot of ways I think this undersells the wonders of process tech (which honestly is magic to the average person, and even not-so-average) but from an end-product point of view the advancement IS iterative.