Is Intel's "Process Lead" Somewhat a Sham?

[Arachnotronic]

Serious question for debate. With TSMC claiming that its 16nm FinFET+ will have similar performance to Intel's 14nm transistors, and with Samsung claiming 14nm high volume production this year, just how much of a process lead does Intel actually have? Why does Intel keep claiming that it's *extending* its lead here?

Seems to me that there is one but that the gap is shrinking...at least if these PRs from TSMC/GloFo are to be believed (and that's the real debate here...whether they are to be believed).

Discuss.

 

[frozentundra123456]

Well, that is all they are now is PR. I would wait to see if they come out on time and see how they perform. In any case none of these will be x86 so Intel's lead is safe there. Unfortunately where they need the process lead most is in tablets/phones where arm is more competitive.

 

[blackened23]

Speaking in general terms, the only thing that can be said is that there is a huge disconnect between what engineers do and what marketing claims their engineers are doing. Marketing will take the work of engineers and make a claim that is partially based on truth. In the end you can just look at what's happening on the market and make a decision for yourself. As far as I can tell, intel's 14nm is about to hit the market, while TSMC's 20nm. Appears to be very late this year to 2015. I dunno. We'll see though.

Time and time again marketing will take the good work of engineers and twist it into some really stupid spin. I personally feel that TSMC falls under this category, although no company is entirely free of this. I also think AMD is a major offender. AMD marketing slides are something of legend in terms of half truths and spin. But, again, no company is completely free of this. Everyone is guilty to varying degrees. Just different shades of grey. Just ask the question as to whether intel has met prior claims with regard to nodes, and has TSMC? I really don't know, don't care to scour the web for it.

 

[dave_the_nerd]

As for the future, any given company is going to either minimize their disadvantages, BS their disadvantage into a strength, or overstate their advantage. That's marketing.

Until the rubber meets the road (so to speak) we don't know.

What we do know is that if X-Brand chips of a given generation are performing better within a given power envelope than Y-Brand, then either they're better designed or the process tech is better.

Since it's unlikely that a single company would be able to permanently hire every engineer on earth that can design a chip worth a darn, I think it's safe to say that design will tend to be a wash, with certain chips doing certain things better depending on the priorities of the designers.

Which leaves process tech as the reason Brand-X > Brand-Y. Probably. (Bad design happens.)

 

[videogames101]

haha 14nm volume from samsung this year, intel has consumer products with 14nm in 2014

completely different, huge lead for intel

 

[blackened23]

Correct. The job of marketing is to be lying scumbags. Or at least BS your way through half truths to make the work of engineers seem completely overstated into positive spin.

In the end, you have to look at past precedents. Has intel met prior marketing claims? Have they delivered close to on time? Have they BS'ed in the past? Then take those same questions and apply the BS metric to TSMC. Which corporation is BS'ing more through marketing? Applying the BS metric to both intel and TSMC to me seems to heavily favor intel in terms of intel having a massive process lead. TSMC most certainly has BS'ed even with their 28nm node. Although i'm sure some will differ on that opinion.

All you must do is look at actions in prior years and the market as it to find the answer. With nodes specifically, as I mentioned earlier, intel's 14nm is in production. Not sure where TSMC's 20nm is.

 

[PPB]

Intel's problem is that their process lead is functional to the volumes they ship to customers and partners. If the demand declines they have to clean inventory of their former products first, streching their product cycles and the need for smaller and better manuf. Processes. TSMC, or GF for that matter, dont have that problem as they are purely a fabbing business. You cant compare both business models as they are totally different.

Delivering better products than their competitors, in the latter case, relies exclusively on having a better process than the competition and having it earlier. On Intel's case, to compete in the business that actually makes them money, selling CPUs, having a better process node is only one of many.

In the mobile space the process node became a determining factor to gauge your competitiveness. This and the fact that the mobile market is huge can explain the decision behind the subsidised price in shipping bay trail products. It lets them penetrate the market while speeding up the the pacing of their fab development cycle.

 

[mrmt]

Serious question for debate. With TSMC claiming that its 16nm FinFET+ will have similar performance to Intel's 14nm transistors, and with Samsung claiming 14nm high volume production this year, just how much of a process lead does Intel actually have? Why does Intel keep claiming that it's *extending* its lead here?

What TSMC calls 14nm is just 20nm with finfets, so yes, Intel does have a process lead. That TSMC calls its process 14nm and Intel also calls its process 14nm doesn't mean too much.

As for Globalfoundries... I think we can write off that one. They are just a Samsung licensee. It's Samsung we should watch, Globalfoundries is just an appendix. Just like we had to watch IBM to know where the alliance was heading, we now have to watch Samsung.

 

[erunion]

Serious question for debate. With TSMC claiming that its 16nm FinFET+ will have similar performance to Intel's 14nm transistors, and with Samsung claiming 14nm high volume production this year, just how much of a process lead does Intel actually have?.

TSMC and Samsung still haven't matched the 22nm process that Intel launched 2 years ago. But we're supposed to believe that they're able to completely skip that node and match future Intel nodes? Not likely.

Intel's lead can't change until someone launches a new product. We have to keep waiting....

 

[R0H1T]

Intel's problem is that their process lead is functional to the volumes they ship to customers and partners. If the demand declines they have to clean inventory of their former products first, streching their product cycles and the need for smaller and better manuf. Processes. TSMC, or GF for that matter, dont have that problem as they are purely a fabbing business. You cant compare both business models as they are totally different.

Delivering better products than their competitors, in the latter case, relies exclusively on having a better process than the competition and having it earlier. On Intel's case, to compete in the business that actually makes them money, selling CPUs, having a better process node is only one of many.

In the mobile space the process node became a determining factor to gauge your competitiveness. This and the fact that the mobile market is huge can explain the decision behind the subsidised price in shipping bay trail products. It lets them penetrate the market while speeding up the the pacing of their fab development cycle.

Agree with everything you said except this, display is still IMO the number one factor driving sales of handheld devices(including tablets) & then comes internal storage, both more important than SoC atm. The chips powering these devices are now good enough for 90% of the general populace, though GPU power still needs to go up as we'll see 1600p screens on smartphones soon enough, so a(ny) node advantage is minimal at best because you're saving very little in terms of battery power. Add to this the fact that Intel is set to loose over $ 3B on their mobile division this year making their transition to 10nm more bumpy, this is what some of us said in the other thread that it's not sustainable & that Intel will either have to permanently settle for low margins (wafer thin) as a norm in this sector &/or leave it eventually because the only (major) players making money already have their own SoC which will not be replaced by Intel inside anytime soon.

 

[TechFan1]

I would think 3D chip stacking may start being a factor in the next couple years along with process node.

 

[Arachnotronic]

TSMC and Samsung still haven't matched the 22nm process that Intel launched 2 years ago. But we're supposed to believe that they're able to completely skip that node and match future Intel nodes? Not likely.

Intel's lead can't change until someone launches a new product. We have to keep waiting....

TSMC and Samsung have made some pretty bold claims and it won't be too long before they are tested.

I find it puzzling that both of these players have been able to 'catch up' to Intel given the lead Intel had as well as the massive R&D resources Intel puts into this, so something may be amiss.

 

[Roland00Address]

Quick question is samsung/global foundry 14nm a true 14nm, or is it old node with finfets added like tmsc? I ask since samsung only claims 15% increased densisty over industry standard 20nm.

 

[Ajay]

Quick question is samsung/global foundry 14nm a true 14nm, or is it old node with finfets added like tmsc? I ask since samsung only claims 15% increased densisty over industry standard 20nm.

Which would make it like TSMC's process. First metal layer is pretty much the same as in 20nm. But it becomes a somewhat odd distinction when GFL/Samsung skip 20nm.

 

[FwFred]

...

 

[VirtualLarry]

TSMC and Samsung still haven't matched the 22nm process that Intel launched 2 years ago.

What about the data that suggests GF's 28nm (Kabini AM1 CPU) has lower idle power than Intel's 22nm (Bay Trail-D), when used with equal power supplies?

Edit: Data here:
http://forums.anandtech.com/showpost.php?p=36265992&postcount=181

 

[Idontcare]

Is Intel's "Process Lead" Somewhat a Sham?

I think the question posed is an unintentional red herring.

The process lead, real or imagined, waxing or waning, growing or shrinking, takes secondary importance to the more pressing question of "does Intel's process lead matter?"

To x86 CPU customers clearly the process lead matters, and to AMD's employees, shareholders, and enthusiastic customers it also matters.

But does it really matter (has it ever mattered?) to Samsung, TSMC, UMC, etc that Intel has a <insert preferred adjective to describe the breadth of the gap> lead over them?

Until Intel leverages its process lead in areas that fabless companies (AMD being the only exception here) live and die in, does it really matter to those fabless companies or their foundries that Intel may or may not have a process tech lead?

It did not matter to Nvidia on the eve of Larrabee's launch. And it doesn't seem to have Apple, Qualcomm, or Samsung running to hills either.

To whatever extent Intel has a process lead, it appears to be equally adept at making its lead mostly irrelevant to all other business segments outside x86 processors in traditional form factors.

Of course that is about to change, but only if you read Intel's PR...which goes back to the OP's point (regarding reading PR and extrapolating reality from it)

 

[FwFred]

I think the question posed is an unintentional red herring.

The process lead, real or imagined, waxing or waning, growing or shrinking, takes secondary importance to the more pressing question of "does Intel's process lead matter?"

How would a process lead (density, power, yield, etc) not be beneficial (i.e. matter)? Of course many other factors are at play when it comes to winning sockets. Perhaps you meant the ranking of factors? I could see process technology ranking lower on the scale in highly integrated phone SoCs compared to dense servers.

 

[erunion]

What about the data that suggests GF's 28nm (Kabini AM1 CPU) has lower idle power than Intel's 22nm (Bay Trail-D), when used with equal power supplies?

Edit: Data here:
http://forums.anandtech.com/showpost.php?p=36265992&postcount=181

That specific set of data? Erroneous.

But in all serious, that is a complete non sequitur. That is comparing idle power of several different computers.

 

[AtenRa]

I think the question posed is an unintentional red herring.

The process lead, real or imagined, waxing or waning, growing or shrinking, takes secondary importance to the more pressing question of "does Intel's process lead matter?"

To x86 CPU customers clearly the process lead matters, and to AMD's employees, shareholders, and enthusiastic customers it also matters.

But does it really matter (has it ever mattered?) to Samsung, TSMC, UMC, etc that Intel has a <insert preferred adjective to describe the breadth of the gap> lead over them?

Until Intel leverages its process lead in areas that fabless companies (AMD being the only exception here) live and die in, does it really matter to those fabless companies or their foundries that Intel may or may not have a process tech lead?

It did not matter to Nvidia on the eve of Larrabee's launch. And it doesn't seem to have Apple, Qualcomm, or Samsung running to hills either.

To whatever extent Intel has a process lead, it appears to be equally adept at making its lead mostly irrelevant to all other business segments outside x86 processors in traditional form factors.

Of course that is about to change, but only if you read Intel's PR...which goes back to the OP's point (regarding reading PR and extrapolating reality from it)

Excellent post as usual IDC,
I will only add that because of Intel's intentions to enter the mobile market that GloFo/Samsung/TSMC are now investing more and adjust their processes to match Intel's.
They also have to be more competitive now because Intel has started producing not only for itself but for others too. So they know very well that if they slip more behind they will soon loose customers to Intel.

I will also add that SamSung/GloFo 14nm FinFet and TSMCs 16nm FinFet are closer to Intel's 14nm in electrical/performance characteristics than density. From the little available data we have, it seams to me that Intel will have better density.

 

[Abwx]

That specific set of data? Erroneous.

Denial with total absence of arguments..

Tell us what is erroneous but you cant , obviously.

But in all serious, that is a complete non sequitur. That is comparing idle power of several different computers.

Theses are same systems with same features if we except the SoCs differences and the results are what they are, that is , they contradict Intel s claims of vastly superior process.

I m done with the bad faith we can read here but since Computerbase.de is not enough let s see the findings of PClab wich is not a site known to favour AMD.

First this graph, pay attention to the Intel J1800/1900 iddle power that is measured using an Antec with a 90W DC-DC converter while the Kabinis have a FSP 550W , theses are theses kind of set ups that wrongly make people think that the Bay Trayls where slightly more efficient :

But seems that PClab noticed the brouhaha that occurred at AT and elsewhere about their test conditions so they re measured the Kabini plateform with said 90W Antec case :

Here we are, but if not enough let s check other sites who used equal conditions with PSUs either very good or good enough :

Golem.de :

The dual core J1800 has a single watt less idle comsumption than the higher clocked quad core Kabini, i spare you the load power....


Hardware.info

Check in the 5350 threads what is measured by AT forum members and you ll see that thoses numbers are indeed accurate and if you dont agree then come with numbers not with hollow words to forever elude the technical debate that would be otherwise fruitfull since the past days reviews also allowed to have a hint about how BT chips works physically speaking.

This isn't the thread for arguing over 5350
-ViRGE

 

[theeedude]

Intel's process lead is most certainly not a sham, but it is now as big as it's ever likely to be in foreseeable future. They have FinFETs, which are a major discontinuity, out now, and the competitors don't. That is an undisputed win by knockout. Once TSMC and Samsung have FinFET, then whose fins are better or smaller or faster or cheaper will be a win on points.
So it's unlikely that Intel will in the future have a process lead that will give it some advantage that it doesn't already have. So whatever you though the process lead was going to do for Intel, if it hasn't done it for them by now, it probably won't later.

 

[Abwx]

Reading through this thread it is obvious that no one answered Intel17 question other than with vague and meaningless statements based on psychological perception and often on a strong bias that will render any technical debate outcomeless.

In this respect i find it hilarious that some people often dismiss let say AMD s slides as being pure PR but when discussing Intel s actual supposed process lead they will systematicaly rely on Intel s PR marketing ignoring real world numbers that give a few clues and indeed none of the posts have numbers, only speculations wrote out of a variable gullibleness of said PR material.

 

[ShintaiDK]

Intel continues to grow its process lead. I estimate them to be around 3½ years ahead. Node and transistor combined.

 

[Abwx]

Intel continues to grow its process lead. I estimate them to be around 3½ years ahead. Node and transistor combined.

What is the basis of such an affirmation ? Intel s slides? Hearsay , or more simply wishfull thoughts.?