Intel's Future Chips: News, Rumours & Reviews

Page 142 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
"10nm is dead! Long live 10nm!"

Marketing can call it whatever they want. Intel has to prove itself now more than ever before; if they're lying about 10nm being *technically* alive (and not marketing-wise), then they're just messing with us. Well, not that is any different from now.

EDIT: too cheery :p

Cheers!
 

aldaia

Distinguished
Oct 22, 2010
533
18
18,995


Intel denying it is not prof the rumor is not true. September 2017 there was a rumor that 10nm was delayed to at least late 2018. Intel denied the rumor. We all know what happened since then.

So until we see no proof, both Semiaccurate article and Intel`s tweet are plausible.

By the way Semiacurate article has been updated after Intel's tweet.
Update October 22, 2018@3:30pm: Intel has denied ending 10nm on Twitter. The full tweet is, “Media reports published today that Intel is ending work on the 10nm process are untrue. We are making good progress on 10nm. Yields are improving consistent with the timeline we shared during our last earnings report.” SemiAccurate stands by its reporting.


 

aldaia

Distinguished
Oct 22, 2010
533
18
18,995

This is exactly what many knowledgeable people is wondering
See david schor for instance
https://twitter.com/david_schor/status/1054419167267508224

Ian Cutress even asked intel about it.
See Intel response :ouch:
https://twitter.com/IanCutress/status/1054420072960270338

 


It's the obvious question to ask in all fairness. Specially after how badly they lost face to the whole 10nm thing.

There's also some interesting things to keep on talking about 14nm with the i9 9900K. Looks like they still cheaped out on soldering and they screwed up with it. der8auer has an interesting video about it.

Cheers!
 


The fact that they denied it puts Charlies claims into the fake news category. Until either Intel states 10nm is dead or we have some hard proof its just him guessing.



Really? This review shows the solder working better than with Hydronaut:

https://www.gamersnexus.net/hwreviews/3378-intel-9900k-cpu-review-solder-vs-paste-delid-gaming-benchmarks-vs-2700x/page-2

intel-9900k-soldered-thermals.png


However not everyone will be happy. Unless Intel magically pulls a 10GHz 64 core CPU out of their behinds as I guess that matters. Oh and it runs at 32c on passive cooling.

Jokes aside I am sure it is fine as a normal CPU. People dissect everything and some people find more wrong than others. I am at the point where I wonder if overclocking is even necessary anymore. Most CPUs max out at 5GHz (Intel, AMD looks to be 4.3 normally). The boost levels are actually decent for most CPUs and performance gains from making a CPU run at xGHz always doesn't seem to be much. I guess until we get to a point where we can OK beyond what most CPUs can boost to I am having a hard time seeing any benefit to doing it. And I love OCing chips. Bought my Q6600 for just that reason, it was a great OCing part. However that was a CPU that with good air cooling could go from 2.4 to 3.6GHz on air (a 50% OC) which did benefit as the CPUs didn't boost at all back then.
 

goldstone77

Distinguished
Aug 22, 2012
2,245
14
19,965
Solder is a reversion back to the Sandy Bridge era, and one that has been needed to get an extra couple hundred MHz out of the existing process. This is still an improvement over thermal paste, but it’s not perfect. The reason it isn’t perfect is largely because of how thick the solder layer is, whereas liquid metal can run a much thinner interface to the IHS.
https://www.gamersnexus.net/hwreviews/3378-intel-9900k-cpu-review-solder-vs-paste-delid-gaming-benchmarks-vs-2700x/page-2

This corresponds to the video by der8auer. Liquid Metal performs better than the solder, because of the thickness of the solder/gap between the IHS and silicon.

Edit: The significance of this: Ryzen wasn't able to clock higher with liquid metal, so the inference is that the STIM was not properly applied by Intel.
 

aldaia

Distinguished
Oct 22, 2010
533
18
18,995

The title of this thread is Intel's Future Chips: News, Rumours & Reviews We should differentiate rumors from fake news.

A rumor is information passed on by someone who doesn’t know or bother to check if it’s true. No effort has been made to check it out before passing it along. It might be true. It might be partially true. It might be false.

Fake news, on the other hand, is a deliberate lie presented as actual news. Its purpose is to deceive as many people as possible.

Charlie statement should be taken as a rumor.
On the other hand the second sentence (Intel continues to lead in 10nm technology.) of this tweet is a good example of fake news.
https://twitter.com/intelnews/status/872939120334479361
At that time Intel was well aware of the issues with 10 nm yet deliberately published a lie with the purpose of deceiving people.
 

aldaia

Distinguished
Oct 22, 2010
533
18
18,995
Continuing with rumors, pressumed Ice Lake Client geek bench results.

ICL U vs m3-7Y30, both running Ubuntu. Average difference: ST: 15% MT: 15.5%

DqNQYqJWoAEXsw_.jpg

DqNQZrAWwAALZu4.jpg


Mos of the improvement comes from the use of AVX-512
 


No one else has 10nm yet and the 7nm that people are pushing out is closer to Intels 10nm so I can't say thats fake news as again no one else is even trying to develop a 10nm process.

And a rumor has some basis in facts unless you are WCCFTech and then you just post every rumor available. However posting something this big that can tank a stock price, especially if no one knew about it, and then not offering hard evidence doesn't make it a rumor. Rumors also tend to come from more than one source. For example we had multiple leaks from multiple sources about the i7-9900K. Not just a single. Yet Charlie is the only one with this information.

Either way I never trust anything he puts out "rumor" or not.
 


As I noted before, Charlie is the type of person who takes information that is valid, then tries to extrapolate it in such as way as to create a story he can sell that's based on "analysis" of said facts. Sometimes he ends up being correct, but he's wrong even more (see: NVIDIA/Intel doom and gloom for the past decade). But he trumpets the times he is correct in order to continue selling premium content.

Charlie may end up being "correct"; several people here including myself have openly wondered if Intel should kill 10nm and focus on 7nm. The difference is I pose it as a question, where Charlie poses it as undeniable fact.

Basically: He's shooting in the breeze, and sometimes ends up being correct.
 

Isaiah4110

Distinguished
Jan 12, 2012
603
0
19,010
The bottom line here is that you HAVE to give more weight to Intel's "Press Release", even though it came from their Twitter account. A Tweet from Intel's official Twitter account intended to mislead investors is a punishable offense. This was even noted in a Tom's Hardware article:

Intel's unusually quick response is telling. The company certainly doesn't need any lingering questions surrounding its progress on its 10nm node, particularly as it nears its earnings call later this week. As an official dispatch from an Intel twitter account, the statement does hold all of the legal weight of any official Intel statement, and misleading statements to investors is a punishable offense.
 

That hasn't stopped them before though :D

Cheers!
 


I guess with newer people who have not been around to see the history its hard to define for them. I really need to get a list of all the articles he made that were just plain wrong but I am just too damn lazy to care.
 

goldstone77

Distinguished
Aug 22, 2012
2,245
14
19,965
I don't understand how people can forget that 10nm was set by Intel to launch late 2015 early 2016. They have been flat lying about it ever since! Just this year there was reports of single digits yield, and they have moved the goal post back to end of 2019 holiday season. There no question Intel has been flat out lying about 10nm, and doing it for multiple years!

Edit: They knew damn well in 2015 that it wasn't going to be ready to ship!
2nd Edit: Just for clarity:
Intel: We know how to make 10nm chips
Intel researchers believe they have a workable manufacturing method for 10nm processors, paving the way for future chips that consume less power and have greater performance than their forebears.
By Jack Clark | September 12, 2012 -- 20:40 GMT (13:40 PDT) | Topic: Processors

Intel has surprised the IT world by changing its plans. Let's break down what's happened.

At IDF in 2013, Chipzilla boasted that it could get processors fabricated using a groundbreaking 10-nanometer process onto the market by 2015.
Intel's research group are also exploring technologies for 7nm and 5nm solutions, though these are a very long way off as 10nm is not expected to go into production qualification until 2015.
https://www.zdnet.com/article/intel-we-know-how-to-make-10nm-chips/
 

goldstone77

Distinguished
Aug 22, 2012
2,245
14
19,965


No one else has 10nm yet and the 7nm that people are pushing out is closer to Intels 10nm so I can't say thats fake news as again no one else is even trying to develop a 10nm process.

10nm is a name, and other foundries have made 10nm, and have improved on that process to make 8nm. TSMC and Samsungs 10nm were more dense than Intel's own 14nm. 14nm++ is actually less dense the the first products.

VLSI 2018: Samsung’s 8nm 8LPP, a 10nm extension
samsung-density-14nm-10nm-8nm.png

Using the MTr/mm² metric, the new 8nm uHD cell has a transistor density of 61.2 MTr/mm² – Samsung’s highest density non-EUV cell.
https://fuse.wikichip.org/news/1443/vlsi-2018-samsungs-8nm-8lpp-a-10nm-extension/
You can see Samsung's 10nm was 51.82 MTr/mm², which is more dense than Intel's 14nm process.
MTr/mm²
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.
https://en.wikichip.org/wiki/mtr-mm%C2%B2
 


Not sure how you consider them putting a date out then having to push back due to errors. They estimated 2015. Then came an issue so they have to push back.

But to each their own. I can understand a company not being able to deliver a product as complex as this, which BTW was estimated to have a density of 100.76MTr/mm² per the second link which would have made it almost twice as dense as Samsungs 10nm and more dense than Samsungs 8nm.

And again per your links Samsungs 10nm was closer to Intels 14nm (original) which was my statement.

How much do you want to be that Intels 10nm will be just as or slightly more dense than Samsungs 7nm?
 

goldstone77

Distinguished
Aug 22, 2012
2,245
14
19,965


You missed the point of my statements. They been lying about 10nm as far back as 2012. It's almost 2019, and they are not close mass production of 10nm. You would rather make excuses for them instead of looking at history. Intel makes a lot of claims, but has nothing but a broken slow 10nm part to show for it, Cannon Lake. The yields were, reportedly, in the single digits.
You don't see Intel reporting other foudries 10/8nm nodes surpassing their 14nm do you.
Intel_Bohr_MD_foundry_density.jpg

Look at this recent image Intel released. Do you notice some errors around the 2017-2018 time frame?
9110881-15271462873472602_origin.png

By 2019 we should have already had 10nm+ released! You think a company making billions of dollars a quarter could possible know what is happening with their process? You think this billions of dollar company doesn't know what it's doing? It's knows exactly what it's doing, and it's being doing it for a very long time, lying!
 

aldaia

Distinguished
Oct 22, 2010
533
18
18,995
Yesterday was Intel Earnings conference call.
Intel posted the best quarter in its 50-year history. They acknowledged a 14nm capacity shortage in Q3 and expect the shortage to become more severe in Q4.

W.r.t.10 nm Intel reiterated that its 10nm process is on track for release by the 2019 holiday season. However they also stated that some capacity from 10nm has been repurposed to deal with 14nm supply.

On the rumor about 10 nm being killed seems there is a second source stating that on practice 10 nm is dead. Killing 10 nm, doesn't prevent Intel to make another 10nm paper launch to save face like they already did in late 2017 or to develop a new less ambitious process and call it 10 nm.

jerFBgz.png


Of course it could be a fake, or Charlie in disguise :D or even could be Charlies source, making it a single source rumor, so take it with a grain of salt.

 


I am sure they know whats wrong and are working on it. I am sure they wanted to launch 10nm in 2015 as it would have been at the time and even till now nearly 2x as dense as anything out there, probably even most 7nm processes coming out.

I wont argue with you as its not worth it. I see it differently than you do, I don't see it as lying rather as unforseen issues, and that's all there is to it. I understand the complexity of this technology and know that what they were trying to achieve is not an easy feat. No other company was trying to push that high of a density.

For all we know they could have a viable working 7nm process they are going to ramp up quickly after they push their 10nm out the door.



A guy risking his job like that seems a bit sketchy to me. Who knows though. The guy would never post his actual credentials or proof he works for Intel. I am sure he needs the money to live.

Or he could be the janitor. Who knows.
 

goldstone77

Distinguished
Aug 22, 2012
2,245
14
19,965


Jimmy, correct me if I'm wrong, but it doesn't appear that you understand the lithography process, and the limitations of current tooling. It's impossible to accurately make the features for what Intel is calling 7nm at this time. So, for all we know, they absolutely do not have a viable 7nm working. And without EUV it's a extremely hard task to make a single 10nm chip at 100.76 MTr/mm² without it being riddled with defects. The tools just are not capable of measuring that small accurately. So, I'll go back to what I originally been trying to say, Intel has been lying about 10nm for a very very long time!

Edit: Just to add to the tooling, the new tools are expected in 2021, and they are 2 stories tall! It is going to be a monumental feat into even integrate those tools to a clean room environment.
2nd Edit:https://twitter.com/lasserith/status/1042750214757646336
https://thechipcollective.com/posts/moshedolejsi/what-is-up-with-euv/
 


Again not going to argue. We don't know what Intel has or don't have.

Also, they have been working on FAB42 quite a bit recently, it is currently being retooled for 7nm, and they have multiple large cranes. I could take a few pictures as I pass by if you would like.

As for viable, they could have a viable 7nm process. I didn't say something they could ramp up right away Ias they do have tools that they would have used in the research phase for small scale testing.

Samsung as well announced that they have a EUV "7nm" process and are beginning to mass manufacture it. So there are tools just not the ones required for what Intel would probably want to push density wise. Still doesn't mean they might not have something similar to Samsungs that is viable.

PS while I don't know the innate details of lithography I do understand it. I know there are tools required to be developed. At the same time they have to do research to develop the tools and process for use with those tools. From my understanding they typically do small scale tests trying to get the results desired then retool the FABs to suit it. My assumption is that they did research in small scale and now are retooling FAB42 for 7nm, be it EUV or a different design, as now 5nm is in research.
 

goldstone77

Distinguished
Aug 22, 2012
2,245
14
19,965

1. "Viable" 7nm, as in working, they have not produced any evidence of that. They barely have a "viable" 10nm chip.
2. Samsung's 7nm is using a process that current tools are capable of producing in HVM. Unlike Intel's original 10nm, which has features too small to accurately make for HVM.

Sure it's possible to make features smaller than current tooling allows,but doing it accurately is the problem. They will be riddled with defects, and have very low yields. Similar to the single digit yields that has been reported.

TLDR: There has been a lot of work on EUV by many people and a lot of challenges have been overcome, but right now even trying to make 20 nm CD (40 nm pitch) features with the state of the art NXE3400b with a source power of 250W requires a dose of roughly 80 mJ/cm^2 which means you can only make 40 wafers per hour. It’s just not cost effective, and no one knows when it will be. In the meantime, we need to use density multiplication to extend 193 immersion. Stay tuned for an article on that.
https://thechipcollective.com/posts/moshedolejsi/what-is-up-with-euv/

40nm pitch is already hard to do, and Intel went below the 40nm pitch hard deck with 10nm(100.76 MTr/mm²), and that's why it's been impossible to produce HVM. That is why rumors of Intel relaxing the specs on 10nm made sense in an attempt to try and get a product viable for HVM.

Edit:
Intel's 10 nm process consists of a minimum metal pitch of 36nm with 8 diffusion lines for a cell height of 272 nanometers.
https://en.wikichip.org/wiki/mtr-mm%C2%B2

2nd Edit: It didn't take Intel 4 years to figure this out. The tooling to make these features for HVM won't exist until 2021.
 

aldaia

Distinguished
Oct 22, 2010
533
18
18,995
According to Digitimes, it has been confirmed that TSMC will produce several processors and chipsets for intel.
https://www.digitimes.com/news/a20181030PD205.html

For demand from entry-level PCs and Internet of Things (IoT) devices, Intel is planning to outsource its entry-level Atom processors and 14nm chipsets to outside makers and expects the shortages to be resolved by the first quarter of 2019, the sources said.

The sources pointed out that Intel and TSMC have been in talks about outsourcing production for the chip giant's CPU and chipset since mid-2018.
 

goldstone77

Distinguished
Aug 22, 2012
2,245
14
19,965


Just note:
As its processor supply continues to fall short of demand, Intel reportedly has begun planning to outsource production for its entry-level Atom processors and some of its chipsets while keeping its high-margin Xeon and Core CPU production in-house, according to sources from the upstream supply chain.
However, Intel declined to comment on "market rumors."
Intel had worked with the Taiwan-based foundry house including having TSMC manufacture SoCs using Atom as the basic architecture in 2009 and the production of Intel's SoFIA handset SoC in 2013. Currently, TSMC is the manufacturer of Intel's FPGA series products.

If true, a very interesting turn of events. There was speculation by several people that Intel should go fabless, so it can use the most competitive process available.
 
Status
Not open for further replies.