Intel's Future Chips: News, Rumours & Reviews

Page 133 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Let me add some perspective to what too the last post I made. https://semiengineering.com/whats-after-finfets/

In theory, finFETs are expected to scale to 5nm as defined by Intel. (A fully-scaled 5nm process is roughly equivalent to 3nm from the foundries).
Click to expand...
Looking at the chart below we can see the CPP at Intel's 7nm and other foundries 3nm.
Transistor gate pitch is also referred to as CPP (contacted poly pitch) and interconnect pitch is also referred to as MMP (minimum metal pitch). ... TSMC reported their 10 nm process as having a 64 nm transistor gate pitch and 42 nm interconnect pitch.
Click to expand...
21826d1529099950-slide6.jpg

This shows at Intel's 7nm it will be at 37nm, which is below what IMEC has consider the tipping point for FinFet.

Here is another view of the different devices going smaller.
Fig.-1.png
 
I remember Intel was researching possible replacements to silicon years ago and most of their research is started 5-10 years before consumers ever hear of or see of it.

They also were planning to move from silicon from 7nm and on, such as using Indium Gallium Arsenide.

Graphene was also considered but has down sides, such as fragility, or Silicene or Germanene both which degrade within a few minutes or hours.

The there is QWFETs which have not gotten much information for a few years so who knows where that is.

The nanosheets sounds interesting but I want to see it in practice. A lot of breakthroughs always show the process with SRAM which is fine but in no way as complex as a CPU.
 
This Dell G7 15 Gaming laptop with i7-8750H and GTX 1060 6GB graphics is only $879.99
By Steven Parker @RealSteveParker · Jun 26, 2018 11:34 EDT
15.6-inch FHD (1920 x 1080) IPS Anti-Glare LED-Backlit Display
Intel i7-8750H Processor (6-Core, 9MB Cache, up to 4.1GHz w/ Turbo Boost)
8GB (2x4GB, DDR4, 2666MHz)
256GB Solid State Drive
NVIDIA GeForce GTX 1060 with NVIDIA Max Q Design technology, 6GB GDDR5
802.11ac + Bluetooth 5.0, Dual Band 2.4&5 GHz, MU-MIMO/160Mhz, 2x2
Connectivity: 6 x USB 3.1 2 x USB 2.0, 1x RJ-45 Network Ethernet 10/100/1000
Ports: Headphone/Mic, 2x USB 3.1 Gen 1 Type-A, 1x Type-C Thunderbolt 3 with support for 40 Gbps Thunderbolt and DisplayPort 1x HDMI 2.0 1x Gigabit Ethernet RJ-45
2-in-1 SD / MicroMedia Card slot
56 Whr, 4-Cell Battery (Integrated)
Click to expand...
https://www.neowin.net/news/this-dell-g7-15-gaming-laptop-with-i7-8750h-and-gtx-1060-6gb-graphics-is-only-87999

Link to this laptop @Dell
https://www.dell.com/en-us/shop/dell-laptops/new-dell-g7-15-gaming/spd/g-series-15-7588-laptop/dncwfsb632s?prg=1&VEN1=12578053-8289238-be9ff910-1926-4fa5-aa65-554a1e0c4840&AID=8289238&dgc=CJ&DGSeg=DHS&cid=198375&lid=45846&acd=12309198375458460&VEN3=812704463116434378

This is an outstanding deal with Intel's new 6 core and Nvidia's 1060 6GB for under $900!
 
It's a neat package, but I have to say I don't like "integrated" batteries in laptops. Also, it seems that it's a MaxQ; I wonder if you can attach an HDD to it to have some storage, at least as an option.

As a really good "portable gaming" device, it's really good and hard to beat. You will need to carry around an external SSD and use the port to have some room though.

Cheers!
 


If it's just a plugged in "integrated battery" it might not be too bad.
Inspiron 15 7000 Series Service Manual
GUID-360EF90A-1E3E-4C91-B6EA-5E2BCF306DE8-low.jpg

https://www.dell.com/support/manuals/ca/en/cabsdt1/inspiron-15-7559-laptop/inspiron-15-7559-laptop-v1/procedure?guid=guid-737a9126-d397-408b-8dcf-07aaa1b2a3d7&lang=en-us

Plus, you can probably upgrade the RAM and SSD manually if you have the know how.
 
^That is the majority of the newer ones I have worked on. It just makes them more compact than the old switch release battery types and since the motherboards have gotten so small they have been able to increase the area for the battery thus increasing the total battery size and duration.
 


I can't imagine Intel has seriously been looking at a new uArch for more then a year; that's when their 10nm problems and security problems would have first gotten some serious attention.

And yes, we'd be far better off with Itanium, if only we accepted the short term x86 performance loss during the transition period. We're stuck with x86-64 now though, given Windows 10 is going to be around for years, if not decades, due to Windows now effectively being a service. We're stuck with x86 based CPUs now, for better or worse.



Call me when costs are low enough for mass production.
 
Intel’s firing of CEO Brian Krzanich is a cover for deeper problems
Opinion: Past behaviors indicate the official excuse is just that
Jun 29, 2018 by Charlie Demerjian
Now for the real problem, the public reasonings for the firings are quite simply a pathetic excuse, so what are they covering up? There are three things that SemiAccurate can think of which would lead to this firing, the 10nm failure being the key one. Intel has known how bad things are for quite literally years and hasn’t been telling the truth. SemiAccurate has known about the yield problems since a few months after Cannonlake silicon came back. Since then the things our sources have been telling us have been quite simply diametrically opposed to the official word from the company.

A good example of this is the claim that 10nm is shipped in 2017. Technically Intel is correct, it did ship in the last remaining seconds of 2017 in volumes you could keep in one of your pants pockets. And we can go on but we already covered this in detail back in May.

Yields of fully working chips rounds to zero. Intel’s 10nm process flat out doesn’t work and SemiAccurate’s sources are telling us about the continuing struggle for Intel. Contrast that with the Q1/2018 analyst call and you have an interesting dichotomy. Intel officially said that they knew what the problems were, had a fix, and were going to ramp in 2019. Sounds good right?
Click to expand...
Messy Messaging:

It does unless you know what is really happening. First the 2019 date for the volume ramp was internally set at Q4/2019 or about 7 quarters away from that statement. Worse yet Intel internally had almost zero confidence they would hit that date, but this was not messaged at all to the analysts. If you understand semiconductor fabrication, you know it doesn’t take 18-21 months to implement a known fix. Technically speaking the 2019 date was on the roadmaps and Intel is undoubtedly implementing things to fix the process, so we can’t say they lied. Intent is a much grayer area that you can decide on for yourself.

It is SemiAccurate’s informed technical opinion that Intel’s 10nm process will never work at a financially viable level. According to the company it is already shipping, already doing well, and volume is just around the corner, see why we say zero credibility? In and of itself this is reason to get rid of a CEO, but to come up with a flimsy excuse instead of coming clean is a bigger problem for Intel.
Click to expand...
To Thine Own Self Be True:

Why? Because there is no fix. We will again not go out on a limb and say Intel will not have a competitive server part on the market until 2021 best case, but we are doubtful of that given what we know now. Why? Cooper Lake. And Cooper Lake. And what Cooper Lake intones. And what Intel is telling OEMs and ODMs when they asked business relevant and pertinent questions that should have an easy answer. Intel’s core business is in deep and abiding trouble and the board is doing everything possible to cover it up.

We understand why, but again credibility is easily lost and hard to regain. If an organization wants to fix it’s image, they need to come clean and do so in a credible way. The firing of CEO Brian Krzanich does neither, in fact it moves things in the wrong direction entirely. Unfortunately it also backs up the messaging malaise Intel has of late, strongly intoning there is no will to fix the problems at the level of management where change has to start. This whole ‘affair’ was not a handling of a problem, it was completely avoiding one.
Click to expand...
But the worst thing to come out of this firing was the clear and unambiguous message that Intel’s board does not intend to come clean about the problems at the company, and that should really worry you.
Click to expand...
https://semiaccurate.com/2018/06/29/intels-firing-of-ceo-brian-krzanich-is-a-cover-for-deeper-problems/
 
https://www.semiaccurate.com/2018/07/02/intel-custom-foundrys-10nm-meltdown-is-crushing-a-20b-market-cap-tech-giant/

And that is really big news... Bad news at that, unfortunately.

I would imagine it's LG Electronics, but I have a slight idea it might actually be Qualcomm.

Cheers!
 


Intel Custom Foundry’s 10nm meltdown is crushing a $20+B market cap tech giant
Exclusive: This is the beginning of a far reaching story
Jul 2, 2018 by Charlie Demerjian
After extensive research, SemiAccurate can say this company cannot survive Intel’s lack of delivery intact. The consequences of this are far reaching for both Intel and their customer.

We would like to start this out on a personal note, two actually. First and most importantly is this story makes us sick, it is going to mean the loss of thousands of jobs and the destruction of a once proud and innovative company, all because of executive stupidity. While we are just reporting the issue, it is still pretty horrible to watch and think about, our condolences to those affected but not to those involved at a high level.

Secondly this is by far the biggest story that SemiAccurate has yet to cover, and we have had some pretty big ones lately. It is pure business, bad management decisions, and some scary potential fallout. Technology only comes into it as far as Intel and their utter inability to produce a functional 10nm process after years of delay.
Vague Hints:

That said the current issue is the result of weeks of research, over a dozen talks with sources inside the involved companies, at competitors, suppliers, and in upstream and downstream industries. All pointed to the exact same problem and result, there is no ambiguity on this one among our sources. A massive $20+ billion market cap tech giant bet everything on Intel’s 10nm process to get a leg up on the industry during an upcoming generational change. Intel has not announced this company as a Custom Foundry client either, and would not provide a list of current customers when SemiAccurate requested it last week.

As you know Intel’s 10nm process is now years delayed, is not economically or technically viable, and is unlikely to ever work financially speaking based on what SemiAccurate understands of the problems Intel is still trying to fix. The customer in question put their entire upcoming line of chips at Intel on 10nm, and Intel failed. There was no Plan B, no out, and according to multiple sources, the customer in question can not survive. This is mainly due to a major industry transition that is going on now, the company in question will not have a product to sell into it.

This isn’t a mom and pop company with a trivial product line, it is a foundational technology giant. They have been in business for decades and much of a sector depends on them. And as far as we can see they are dead. Without a product line for a major industry transition, they can not survive and will be acquired or simply die. There is no way out at this point.
Click to expand...

The is a big customer that is going to fall, because of the 10nm debacle. Jobs are going to be lost, but who is it?
This isn’t a mom and pop company with a trivial product line, it is a foundational technology giant.
Click to expand...

It's not LG!
Nope. Article says this customer was not announced as ICF customer. Also, LG phones do not rely on LG SoCs, so ICF being dumb won’t affect that.
Click to expand...
Qualcom... Hmmm....
 
To be fair Charlie has never been an Intel fan and has always been harder on the news stories he covered on them than AMD. Same with nVidia and ATI/AMD.

In 2013 he also spelled out doom and gloom for nVidia. Yet currently AMD is not very competitive with nVidia. Maybe with Navi but nVidia controls the consumer GPU market pretty well right now.
 


Yeah, Charlie has been predicting Intel and NVIDIA failing for almost a decade now. He takes any sign of hardship and runs with it. And since Intel and NVIDIA have so many people who dislike them for one reason or another, he's able to turn those articles into $$$.

But hey, predict the end of the world enough, you'll eventually be correct.
 
While I don't disagree, his articles are as good as the information he manages to gather. In this particular article though, I see people nodding at him from the distance, so the % of it being accurate is quite high, sadly.

As usual, attack the message and not the messenger.

Cheers!
 


There is another member of the tech press who claims they know which company Charlie is talking about. We simply have to wait and see which company, 20+ billion dollar market cap, fits the description:
This isn’t a mom and pop company with a trivial product line, it is a foundational technology giant. They have been in business for decades and much of a sector depends on them. And as far as we can see they are dead. Without a product line for a major industry transition, they can not survive and will be acquired or simply die. There is no way out at this point.
Click to expand...

As far as commentary about the health of 10nm. Simple ask yourself when was it supposed to be released? ~2016... How late is it, and Intel has claimed it's coming tomorrow for years, now. Charlie has pointed out "citing industry sources" that 10nm was in trouble for about ~2 years now. BK the ex CEO, which was likely fired over 10nm, also claims in the last Intel earnings report that "they" know what the problems are, but are unsure when they will be fixed claiming 1H 2019 or 2H of 2019. That doesn't sound like a time table to fix something they are sure they know how to fix. Also, BK's statements about multi-patterning being the majority of the problem, because 10nm is a process before EUV has been developed. If that is true and the claim BK said about Intel being too aggressive in the shrink with sub 40nm pitch, which no one else in the industry attempted, then he is basically admitting they would have to wait for EUV before they can go into volume production of 10nm. BK said in the earning report that they are shipping 10nm, then they have to release a 10nm chip or be sued for lying in the earnings report. They release in China, in very limited quantities, the dual core i3-8121U, which has worse specs than 14nm++, no functional iGPU, and the same power draw as 14nm++. They were shipped with AMD 550 graphics. This again points to serious issued with yield if they can not produce a fully functionally dual core laptop CPU ~2 years after original release date, but "they" claim to know what the problems are and know how to fix them. You don't have to take Charlies word for it, because the writing is already on the wall.

Intel makes its first 10nm Cannon Lake chips official
Intel said it was making 10nm processors; now we know what processor and which customer.
PETER BRIGHT - 5/16/2018, 8:40 PM
Intel's transition to building processors on a 10nm manufacturing process has been delayed repeatedly. Once upon a time, the company said that it'd go into mass production at the end of 2015; with its most recent financial results, the company pushed that back, again, to 2019. But Intel has also said that, although the yields aren't good enough for large-scale production, it has been shipping 10nm processors, codenamed Cannon Lake, to an unspecified customer.
Click to expand...
https://arstechnica.com/gadgets/2018/05/intel-makes-its-first-10nm-cannon-lake-chips-official/

They are not proud of that 10nm chip, which was release because BK opened his mouth in an earnings report saying it was shipping 10nm. They had to release something or be sued by share holders for lying.
 
For a comparison after the recent symposium, VLSI 2018.
"Intel's 22 nm process (2012) had 16.5 MTr/mm², 14 nm process (2014) had 44.67 MTr/mm², and 14++ nm process had 37.22 MTr/mm²." 14nm++ is Intel's current volume production process."Intel's 10nm has a density of 100.76 MTr/mm²"
TSMC is in volume production of 7nm right now, and Samsung showed off 8nm uHD, which is it's lowest density volume production chip.
8nm uHD cell has a transistor density of 61.2 MTr/mm² – Samsung’s highest none-EUV density cell.
21824d1529099818-slide4.jpg


Leading process technology's in volume production today:
Intel:14++ nm process had 37.22 MTr/mm²
Samsung: 8nm uHD cell has a transistor density of 61.2 MTr/mm²
TSMC: 7nm 96.49 MTr/mm²
 
I can see why Intel would have issues with their process with the density they are pushing for.

Guess its a wait and see though. Nothing more we can do really.

And while Charlie can quote sources forever I still hold everything he states with the largest grain of salt in the world.
 


I will say his INFORMATION is usually correct; it's his conclusions which are all wrong. But the fact his info is mostly correct is what draws people to him even though he's constantly wrong with his conclusions.

Here's what I think:

1: Intel has to seriously start to do a cost/benefit analysis of continuing to develop fabs sub-14nm; they're behind, and honestly, Intel might just be better off externally sourcing it's CPUs. There's also the question of Quantum CPUs, which might be viable on older process nodes, making farther development of late to market nodes questionable.

2: In the short term, Intels best options are to increase clocks (via soldering) or increase core count, both on it's existing process.

Oh wait, that's exactly what they're rumered to be doing.
 


Qualcomm is on Samsung's 10nm process, per a recent announcement about a month ago.

I am wondering here...transitional period, no new product, foundational company.

Could it be...Apple?

The partnership would make sense...they are going to get away from using Intel products supposedly for CPUs, but they could easily use Intel's foundry.

I mean, Mac has not seen much of an update in 5 years at this point...
 
I've already pointed this out, but these are the issues with 10nm that BK talked about in the first earnings call. This is why I'm paying close attention to the state of EUV readiness, since EUV could potentially fix most of Intel's yield issues.
Ross C. Seymore - Deutsche Bank Securities, Inc.

That's very helpful. For my follow-up question, for either of you, frankly, on the 10-nanometer pushout, do you believe that the competitive lead you have versus your competition is shrinking, or is this a challenge everybody is going to have? And then the gross margin side of that equation, Bob, you said it was going to be a headwind into the full-year guide. Any sort of linearity about when that starts to move from being a headwind to a tailwind would be great. Thanks.

Brian M. Krzanich - Intel Corp.

Sure, Ross. So let me start with we absolutely have product and process leadership. We're shipping 10-nanometer products today. So I did want to make sure that that was very clear to you, and those are the densest, highest performing products out there.

We're slowing the ramp down as we go and fix these yields, and we're able to do that. A), we understand the yield issues. They're really tied to this being the last technology tied to not having EUV and the amount of multi-patterning and the effects of that on defects. But also, the real strength of 14-nanometer, I mentioned in my prepared remarks that we've done 70% improvements in the performance of that technology over its current lifetime. And we believe it continues to have legs, that we can continue to make improvements, both within that process technology and architecturally. That's really giving us the breathing room to go and make these yield improvements.

So it's really balancing between delivering the world's best products. So we believe our roadmap for 2018 is as strong or stronger than it's ever been. And we have the ability to carry that into 2019, allowing us to get the yields where we want them to be. So the cost and the spending are really in line with what you as a shareholder expect from us.

We believe that if you take a look at others during this timeframe, if you looked at anybody else and said 70% improvement on a technology node, they may rename those nodes as we go through this. And we have always chosen to be really transparent and clean and just say it's improvements on the existing technology rather than renaming. So we believe we have that.

Now as we look out in time, we do see the density. If you just take that component, the density gap is narrowing a bit, but that's out in time. But again, performance is really a function of multiple parts of the process around power and performance and in architecture. And that's why we think our products continue to lead and be the world's standard.
Click to expand...
Stacy Aaron Rasgon - Sanford C. Bernstein & Co. LLC

Hi, guys. Thanks for taking my questions. I wanted to follow up on that 10-nanometer point. So as the volume production pushes out into 2019, given you understand the yield issue supposedly, is this a first half pushout, or does it push out into the second half? And when it actually does ramp, do you think it actually will be the current 10-nanometer process that's shipping, or will that be slipping out to 10-nanometer plus potentially?

Brian M. Krzanich - Intel Corp.

So I'm just going to correct you. You said that supposedly we have the solutions. We do understand these, and so we do have confidence that we can go and work these issues, Stacy. Right now, like I said, we are shipping. We're going to start that ramp as soon as we think the yields are in line. So I said 2019. We didn't say first or second half, but we'll do it as quickly as we can based on the yield.

The last part of your question about whether will it be a 10 or 10-plus-plus or 10-plus I think was your question, the yield improvements that we're making are just that, more focused on yield. So think of them as improvements to the various edge stuff, the lithography stuff, thin cleans (33:54) and things like that in order to really drive the multi-patterning and, in some cases, multi-multi-patterning, where you have four, five, six layers of patterning to produce a feature. It's really about that. They aren't necessarily around performance.

We do have plans on 10-nanometer already, similar to 14-nanometer, for 10-plus and 10-plus-plus. And so we think all of these technologies now have multiple years of performance improvements built into them as they come off the floor.
Click to expand...
https://seekingalpha.com/article/4166652-intel-intc-q1-2018-results-earnings-call-transcript?part=single
 
Intel CPU Roadmap 2018-2020: Intel Sticking With 14 nm, 10 nm Intel CPUs Pushed To 2020
https://segmentnext.com/2018/07/10/intel-cpu-roadmap-10-nm-2020/

The slide is hard to see, but it shows several interesting things:

  • - Canon lake has vanished from the roadmap.
    - Only 10nm parts shown in the slide are Icelake and, later on, Tigerlake
    - 10 nm server Icelake is expected in Q3 2020
    - 10 nm desktop Icelake expected in Q4 2019
    - 10 nm laptop (H series) Icelake expected Q2 2020
    - The earliest 10 nm parts are expected in Q2 2019 in the ultramobile segment. The same that canonlake targeted. I wonder if they will launch low volume too to act as pipecleaner for the 10 nm process.

I don't know if the slide is legit, but it's consistent with all rumors I heard.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom