Intel Set to Fall Further Behind TSMC

R

rgMekanic

[H]ard|News
Joined
May 13, 2013
Messages
6,943
An article from The Motley Fool today outlines how Intel's troubled 10nm manufacturing may end up falling even further behind TSMC. Intel had originally planned to release it's 10nm chips in 2016, which was pushed to 2017, then pushed again to 2018, then a few weeks ago 10nm was delayed yet again. Today the article from the Motley Fool is saying that if all goes as planned, the Cannon Lake 10nm Intel chips should go in mass production, sometime in 2019. These delays are putting Intel about a year behind rival TSMC, who is slated to be mass producing 5nm chips in the second half of 2020.

This could honestly be a massive problem for Intel going forward. In my opinion they sat on their hands too long with the 3-5% performance increases, while their competitors snuck up on them in a big way.

If things don't go well for Intel, then we could see the first 10nm+ products in the second half of 2019, with the 10nm++ products coming in the second half of 2020. Such a schedule would likely mean the first 7nm products wouldn't arrive until the second half of 2021, or potentially even early 2022.
 
  • Like
Reactions: N4CR
like this
not really a problem, best arch will win at equal node, unless intel is truly lacking tallent in which case this is a huge problem for them.
 
alamox said:
not really a problem, best arch will win at equal node, unless intel is truly lacking tallent in which case this is a huge problem for them.
Click to expand...
I doubt they lacking talent. They just weren't pushing them. They were sitting tight thinking AMD was never getting their shit together.
 
One would think that Intel would already be working on the other Xnm sizes. So the problems they are having with the 10nm process hopefully will not affect them much with the smaller sizes down the road. ...

They just may not be able to recoup the $$$ spent on the the 10nm.
 
vegeta535 said:
I doubt they lacking talent. They just weren't pushing them. They were sitting tight thinking AMD was never getting their shit together.
Click to expand...

I agree. Intel sells 36 thread CPUs and they run up to $9000.00. Not exactly a simple thing to make. They just sat back and milked the enterprise while keeping gamers and prosumers from cannibalizing their Xeon sales by not offering cheap many core CPUs. Then came Zen and Threadripper and I'm sure they are freaking out.
 
My body is ready for Zen 2 on 7nm. I hope vega will be more impressive at 7nm but it probably wont be.
 
Is TSMC really a competitor to Intel though?

Sure, they both make chips, but in whatmarket do they compete?
 
In various Chips production. Intel produce not only their Cpu but also chips for Altera,Achronix,Tabula,Netronome,Microsemi and Panasonic,plus Nand flash and other type of chips.Losing the edge don't mean they are gonna lose any of those contract,but could affect the price intel is selling their technology at.
 
KINgGh0sT said:
Source? Cause this sounds very made up.
Click to expand...

Every chipmaker measures in their own way, and Intel is, oddly enough, the least of the bullshitters. TSMC's 7nm is about equal to Intel's supposed 10nm. Samsung is also not equal. So while TSMC is pulling away in actual production, it's not like their process is actually much smaller (and capable of the same complexity, IIRC).

Simple google search will tell you all you need to know.
 
KINgGh0sT said:
Source? Cause this sounds very made up.
Click to expand...

He's actually very correct. The accuracy of the nm figure has strayed *real* far from reality for tsmc/glofo/samsung. If you compare fin pitch, cell pitch, gate width, hell anything related to sizes and dimensions after 28nm people started fudging the figures in a major way. Intel 14nm is still the benchmark process. Anything called 16nm/14nm/12nm/10nm from anyone else can't hold a candle to what Intel have nowand the coming 7nm tsmc isn't anything remotely what Intel would call 7nm. In short the naming scheme has totally broken down for contract manufacturers.

Side note: I have zero horses in this race, building ryzen rig later in the year.

https://techreport.com/review/31660/intel-defends-its-process-technology-leadership-at-14nm-and-10nm
 
Everyone here is thinking this is terrible for Intel because AMD and Zen.

And yes Zen is going to eat into their consumer level profits, no doubt.

The real thing giving Intel management sleepless nights. (the ones that don't have lovely golden parachutes anyway)

Is Marvell and Cavium are now one... and getting ready to make some big noise in the data center / super computing markets.
https://www.cavium.com/news/cavium-announces-thunderx2-general-availability Yes a lot of big players in the Data center level machines like HP have been building Thunderx2 machines... as is Cray. I have heard these are fabbed at 14nm but haven't seen anything confirming that.

As for the not quite long term future.... it sounds like the rumors of Apple giving Intel the boot may not be as far fetched as some think...
https://wccftech.com/apple-a12-soc-...4-30-performance-improvement-over-a11-bionic/
Apple is using 7nm TMSC for A12... and the early reports are its FAST, and that it draws more power then the leakers thought. Which could make perfect sense if Apple is intending to bump the power planning to use them in Mac books.

Long term Intel better get their 10nm and beyond in order fast before companies start shipping ARM chips, that push the software world that way faster then Intel would like.
 
ChadD said:
Everyone here is thinking this is terrible for Intel because AMD and Zen.

And yes Zen is going to eat into their consumer level profits, no doubt.

The real thing giving Intel management sleepless nights. (the ones that don't have lovely golden parachutes anyway)

Is Marvell and Cavium are now one... and getting ready to make some big noise in the data center / super computing markets.
https://www.cavium.com/news/cavium-announces-thunderx2-general-availability Yes a lot of big players in the Data center level machines like HP have been building Thunderx2 machines... as is Cray. I have heard these are fabbed at 14nm but haven't seen anything confirming that.

As for the not quite long term future.... it sounds like the rumors of Apple giving Intel the boot may not be as far fetched as some think...
https://wccftech.com/apple-a12-soc-...4-30-performance-improvement-over-a11-bionic/
Apple is using 7nm TMSC for A12... and the early reports are its FAST, and that it draws more power then the leakers thought. Which could make perfect sense if Apple is intending to bump the power planning to use them in Mac books.

Long term Intel better get their 10nm and beyond in order fast before companies start shipping ARM chips, that push the software world that way faster then Intel would like.
Click to expand...


I would say that Apple having made an arm core thats knocking on the door of modern x86 cores is a massive omen. They are so clearly so far ahead of all the other arm cores. Will all that development effort give them the idea they could ditch intel in the future? Surely its piqued the interest of one or more big cheeses there.
 
Those that think intel was slacking are clueless. The PTD engineers have the hardest life in the company. They've been flogging the employees for years trying to get 10nm on track. It is just incredibly difficult and that is compounded by high turnover and little communication between teams that worked on successive generations.
 
R_Type said:
I would say that Apple having made an arm core thats knocking on the door of modern x86 cores is a massive omen. They are so clearly so far ahead of all the other arm cores. Will all that development effort give them the idea they could ditch intel in the future? Surely its piqued the interest of one or more big cheeses there.
Click to expand...

Cavium has the be concerning them as well. Marvell paying 6 billion for them at the end of last year surprised a few people I think. If Intel isn't sweating bullets over that they should be.
 
I don't see how intel would be concerned about anybody's fab process. Their 14 nm tech is better than anything out there. there is simply no need to rush anything as far as intel is concerned.
At this ppint everyone is playing catch up and it will most likely continue for quiet some time.
 
My quantum/photon computer scoffs at your ludicrously large 10nm process (I wish).

The smaller the process the bigger the e-peen.
 
spintroniX said:
Those that think intel was slacking are clueless. The PTD engineers have the hardest life in the company. They've been flogging the employees for years trying to get 10nm on track. It is just incredibly difficult and that is compounded by high turnover and little communication between teams that worked on successive generations.
Click to expand...

Agreed. I was getting a laugh, too. You don't want to be in any part of process engineering at Intel. I have plenty of friends in Hillsboro. They're all basically owned by their respective tools.

Likewise https://en.wikichip.org/wiki/WikiChip is great for process comparisons for us laity.
 
Last edited:
I don't see the issue? :S

There's always refinements, there's always yield maturity, there's always iterating and if recent history has proven anything is that AMD leaps notwithstanding, Intel is still the go-to for single thread performance; and reversely for AMD.
On top of that, Intel saying 10nm really means 10nm; other fabs, they tend to get a bit.. creative with their own numerology, because PR. What some think of as "14nm" today? Go do some reading :)

So excusing the PR, hype-related sensationalism and false impressions (which yes, will influence the market, but am not here to do market analysis)? Not much to write home about. Yet. Intel has a good leap, still, and is already ready for the next refinement.
I may favor AMD, but i cannot forget the reality of things. We're talking monolithic design here.
So as far as i'm concerned, barring those willingly jumping on the anti/hype train?

The only people that should legitimately worry are:
- Stock holders, obvious reasons
- Intel high level execs, obvious reasons

A reminder about my emphasis on yet.
 
I think we've all forgotten about the Spectre/Meltdown flaws Intel has been hiding for years now, that's finally caught them utterly unprepared for the fallout. The hell with process size: they NEED to fix these flaws ASAP, before they issue yet another compromised chip. I think they're scrambling like hell to do both, and the process shrink is suffering from the focus on fixing those flaws in future generations.
 
Big_Rig_Stig said:
I think we've all forgotten about the Spectre/Meltdown flaws Intel has been hiding for years now, that's finally caught them utterly unprepared for the fallout. The hell with process size: they NEED to fix these flaws ASAP, before they issue yet another compromised chip. I think they're scrambling like hell to do both, and the process shrink is suffering from the focus on fixing those flaws in future generations.
Click to expand...

Good point. I'm not buying a new CPU until I'm reasonably satisfied its immune to these attacks. If they're delaying to make that true, delay away Intel.
 
alamox said:
not really a problem, best arch will win at equal node, unless intel is truly lacking tallent in which case this is a huge problem for them.
Click to expand...

they just hired the miracle worker, i forget his name.

plus "Intel" is a name brand that sells on name recognition alone. it would take various generations of AMD being on top to sway public opinion
 
Not sure why many thinks that the difficulty in shrinking the process node is simply due to Intel being complacent,
as if this is something you can simply will it out, just need to want it enough :cautious:

We're pushing the limits of physics here, and gate size is not the only metric that will influence how efficient is the final product.
We've seen many times in the past, if your process have issues such as leakage, your end product is going to struggle.
 
When it rains it pours......just more bad news for intel. You get big enough and think you are untouchable until the floor falls out on you. Its a far way to fall too.
 
spintroniX said:
Those that think intel was slacking are clueless. The PTD engineers have the hardest life in the company. They've been flogging the employees for years trying to get 10nm on track. It is just incredibly difficult and that is compounded by high turnover and little communication between teams that worked on successive generations.
Click to expand...
But hey, at least they have the most diverse work force in the tech industry! ;)

More seriously, if you have to have that kind of work environment to get things done, no wonder things aren't getting done. Overworked and stressed out employees with low morale don't tend to do their best work.
 
All of these responses about Intel being "complacent" or "lazy" are shortsighted. If you think Intel is falling behind or has, you don't understand the technology or the industry. It's one thing to be a computer enthusiast, it's a completely different thing to understand the manufacturing process and the semiconductor industry as a whole.

You can really pick out people who don't have a clue by these comments. I don't know if it's them being fanboys (hint, after working for both companies there isn't a rivalry like people make it out to be) or just that clueless, but it shows.
 
I'm not an intel fan by any means but I think they are getting a bad rap. This is all marketing plain and simple. Everyone jumping on 7nm will market that fact. We are on 7 they are still on 14nm..we win! Marketing doesnt care that intels 14nm is more refined, fine tuned and is on par or better than their 7nm. All they care about is saying they are on a different # which makes them better.
 
You must log in or register to reply here.
Back
Top