Intel Announces 10nm Is Shipping

[Guest]

Intel's senior vice president and general manager of the Client Computing Group at Intel Corporation, Gregory Bryant, announced at an early event at the company's booth that it is shipping 10nm products.

Intel Announces 10nm Is Shipping : Read more

 

[Cryio]

Maybe Intel will switch 10 nm.

AMD will ship awesome 14+ nm Ryzen+ CPUs in April and we'll get 7 nm Ryzen 2 CPUs next year.

 

[InvalidError]

Since Intel has Meltdown and Spectre to address before its next full-scale CPU launch, perhaps it has decided to mostly scrap the already two years late Cannon Lake (which may already be too far along to make any major changes for addressing the two exploits) and skip to Icelake.

 

[Giroro]

So, out of nowhere, Intel made an outrageous and possibly incorrect "announcement" without any further information, just because it would look good in a headline?

I don't like that. If Intel wants to try and one-up AMD's upcoming 12nm Ryzen, then they need to back that up with things like facts... which shouldn't actually be that hard if it were true.
Although I would expect a 2.0 Ghz dual core low-power chip to be more at home in a mobile or IoT device over a PC.

 

[Malik 722]

what??? intel began shipping cannon lake last year,is this a joke i can't even find coffee lake and they say they are shipping cannon lake.

 

[InvalidError]

what??? intel began shipping cannon lake last year,is this a joke i can't even find coffee lake and they say they are shipping cannon lake.

The Cannon Lake version that is shipping is a dual-core without IGP for embedded applications, not something you are going to see in anything resembling a conventional tablet, laptop or PC. You'll find those in display-less devices, devices with custom displays (arbitrary shape LCDs) or addressable self-refreshing displays, the sort that you often see used in Arduino projects.

 

[berezini]

you are not the OEM those get chips first.

 

[Darkbreeze]

Even if they WERE shipping desktop variants, which they're not, I'm not sure it would be a very promising product at this point anyhow. I don't see anybody wanting to pay new prices for a new product that gets Kaby Lake type performance after the hits on CPU and storage performance from the various patches and firmware are in place.

I agree with InvalidError, they'd be a lot better off just moving on to something they might actually stand a chance of correcting the issue with at the hardware level rather than trying to sell us ANOTHER cpu they knew had problems.

 

[Zaporro]

What, in previous article CEO Brian Krzanich says "no update on 10nm" and now there is "Intell ships 10nm Cannonlake"?

 

[Isokolon]

@darkbreeze

I think that's not viable. The way I understand it the meltdown & spectre issue has to be addressed pretty early in design stage. And since it has been around for so long, they'll need to completely redesigned this process.

To my limited understanding this would mean that Intel couldn't release new CPUs before Q3 2019?
Which would mean a massive breakdown in market share and a huge company loss.

For server & enthusiast/X-CPUs that might be the proper thing to do. For desktop CPUs however.... Little people know and care about this issue, there are fixes that won't impact performance for most users. It would be foolish by Intel to scrap a whole new line of consumer grade CPUs that cost millions to develop for an issue most buyers don't even understand

 

[bluepheonix1]

Intel saying they began shipping the 10nm chip could mean very little. If Intel made and shipped one single chip (say to a reviewer like Toms), then technically, Intel is not lying and technically, they "began shipping in 2017".

I like many of you are in the market for a new chip and have been waiting for the release of a true 10nm. Pushback after pushback and delay after delay, they still are not here.

Hopefully someone from Toms (or anyone really) at CES (or elsewhere) can ask the important question and press the issue for a real answer. Not "when did Intel begin shipping 10nm" but when will people actually be able to purchase desktop variants of the 10nm chip?

 

[InvalidError]

Intel saying they began shipping the 10nm chip could mean very little. If Intel made and shipped one single chip (say to a reviewer like Toms), then technically, Intel is not lying and technically, they "began shipping in 2017".

The Cannon Lake CPUs that Intel has begun shipping are dual-cores without IGP for embedded systems, not CPUs for laptops or PCs. Intel could ship millions of these and you still won't know where they're ending up because they get buried in devices where the CPUs running the show are largely irrelevant and not advertised. Very little talked-about Intel CPUs and micro-controllers are embedded in billions of modern devices.

 

[dullridge]

is anyone gonna buy from intel now

 

[Darkbreeze]

That's SERIOUSLY irresponsible though, since they already KNOW these devices are vulnerable. Just because they don't THINK your refrigerator or smart ashtray (Yes, I'm being a wise ass, kind of) won't get the prerequisite malware on it to make it vulnerable to Meltdown or Spectre doesn't amount to much when it's already been shown that these devices ARE in fact vulnerable to other kinds of security attacks.

I realize they are trying to save themselves from being out millions or billions of dollars on investments, man hours, material costs and testing, not to mention marketing and stock values, but to ship out products for the SECOND time that you already know are baseline vulnerable, no matter what they are to be used in, is gross negligence IMO.

 

[Darkbreeze]

@darkbreeze

I think that's not viable. The way I understand it the meltdown & spectre issue has to be addressed pretty early in design stage. And since it has been around for so long, they'll need to completely redesigned this process.

To my limited understanding this would mean that Intel couldn't release new CPUs before Q3 2019?
Which would mean a massive breakdown in market share and a huge company loss.

For server & enthusiast/X-CPUs that might be the proper thing to do. For desktop CPUs however.... Little people know and care about this issue, there are fixes that won't impact performance for most users. It would be foolish by Intel to scrap a whole new line of consumer grade CPUs that cost millions to develop for an issue most buyers don't even understand

So you are ok with one contractor who has access to department of defense, or some similar agencies network, getting their system compromised and then allowing that vulnerability to gain access to critical systems, potentially, like the power grids, DOD, NASA, nuclear arsenals, hospital networks, the stock market, etc.?

Those places don't need to be directly compromised if some employee or independent contractor working from home gets compromised and opens the door for them. And acting like it can't happen is foolish. It HAS happened before and WILL happen again, this just potentially makes it either worse or increases the potential for it happening. They don't even need to attach those specific systems using either of the current exploits if they've already exploited a system that has unrestricted access to these others.

 

[rgd1101]

MERGED QUESTION
Question from bluepheonix1 : "Intel - 10nm - At CES Intel claims they began shipping 10nm chips at end of 2017"

I read a few articles within the past 24 hours that said Intel began shipping the 10nm chip at the end of 2017. Yet, I do not know of a single person or organization who has received one.

This statement from Intel could mean very little. If Intel made and shipped one single chip (say to a reviewer like Toms), then technically, Intel is not lying and technically, they "began shipping in 2017".

I like many of you are in the market for a new chip and have been waiting for the release of a true 10nm. Pushback after pushback and delay after delay, they still are not here.

Hopefully someone from Toms (or anyone really) at CES (or elsewhere) can ask the important question and press the issue for a real answer. Not "when did Intel begin shipping 10nm" but when will people actually be able to purchase desktop variants of the 10nm chip?

Can someone here reading this forum (either Tom's staff) or anyone, please ask Intel this question and try to reasonably press for a definitive answer?

 

[InvalidError]

So you are ok with one contractor who has access to department of defense, or some similar agencies network, getting their system compromised and then allowing that vulnerability to gain access to critical systems, potentially, like the power grids, DOD, NASA, nuclear arsenals, hospital networks, the stock market, etc.?

To exploit Meltdown and Spectre, you need local access for all but the browser variant, which means that your system has already been compromised by unauthorized local or remote access to run the exploit code in the first place.

These exploits are mostly a concern for virtualized hosts where you have no idea and no control over what the other instances running on the same machine are doing. Run your secure stuff on a dedicated server and your data is as secure as your OS, server software and employees with access to the server are.

 

[Isokolon]

@darkbreeze

I think that's not viable. The way I understand it the meltdown & spectre issue has to be addressed pretty early in design stage. And since it has been around for so long, they'll need to completely redesigned this process.

To my limited understanding this would mean that Intel couldn't release new CPUs before Q3 2019?
Which would mean a massive breakdown in market share and a huge company loss.

For server & enthusiast/X-CPUs that might be the proper thing to do. For desktop CPUs however.... Little people know and care about this issue, there are fixes that won't impact performance for most users. It would be foolish by Intel to scrap a whole new line of consumer grade CPUs that cost millions to develop for an issue most buyers don't even understand

So you are ok with one contractor who has access to department of defense, or some similar agencies network, getting their system compromised and then allowing that vulnerability to gain access to critical systems, potentially, like the power grids, DOD, NASA, nuclear arsenals, hospital networks, the stock market, etc.?

Those places don't need to be directly compromised if some employee or independent contractor working from home gets compromised and opens the door for them. And acting like it can't happen is foolish. It HAS happened before and WILL happen again, this just potentially makes it either worse or increases the potential for it happening. They don't even need to attach those specific systems using either of the current exploits if they've already exploited a system that has unrestricted access to these others.

no, I'm not "okay" with it.

but it's illusionary to think Intel will put everything on hold, scrap all the millions they've invested into this chip because there are some concerns.
furthermore these exploits are especially dangerous for servers,
and no, most people aren't in fact contracted to the ministry of defense.
and no, most Intel chips are in systems noone gives a damn about.
and no, most users won't notice any difference in performance after the patches in windows
and lastly, no, Intel probably won't sell significantly less CPUs in 2018 because most people don't even know what meltdown is.

 

[InvalidError]

and lastly, no, Intel probably won't sell significantly less CPUs in 2018 because most people don't even know what meltdown is.

Since both Spectre and Meltdown are primarily about information leaks, the only systems where they are particularly relevant are shared systems handling private information on a sufficiently regular basis for the exploits to have a chance of catching a transaction in progress. For the typical home user, they are largely irrelevant.

Also, since the exploit code needs to run locally, simply not executing code of questionable origin would keep 99.9% of normal people safe. If you can't do that, then you have much worse security issues to worry about.

 

[Darkbreeze]

no, I'm not "okay" with it.

but it's illusionary to think Intel will put everything on hold, scrap all the millions they've invested into this chip because there are some concerns.
furthermore these exploits are especially dangerous for servers,
and no, most people aren't in fact contracted to the ministry of defense.
and no, most Intel chips are in systems noone gives a damn about.
and no, most users won't notice any difference in performance after the patches in windows
and lastly, no, Intel probably won't sell significantly less CPUs in 2018 because most people don't even know what meltdown is.

First of all, there are NO systems that nobody gives a damn about, because nearly all systems either ARE or can be connected to all other systems. So that's a senseless comment. My neighbors refrigerator can talk to a server in California which can talk to an entire network of systems, and so on, but you know this and are just being snarky. So EVERY system is a potential liability if it's vulnerable, especially if it's accessing your home network, which generally, it is.

And as I said before, yes, it's fully understood that there needs to be some other form of malware or exploit already in place on the target machine for any of these particular strains to be dangerous, but the problem is that as we've discovered there are PLENTY of systems out there, refrigerators, home computers, phones, ovens, whatever, that have already been shown to be particularly vulnerable to the kinds of infections that need to be present for this to happen, so it's really not a stretch to say that this can and likely will happen at some point.

How many people do you know of that are going to patch the software on the refrigerator, or update it's firmware. Practically none, so if the hardware is vulnerable and you are selling it that way knowingly, that can't be acceptable to anybody with a half a brain.

I agree that Intel is "unlikely" to scrap the whole gen, but I also think it's incredibly stupid and short sighted not to. They're already in hot water, have multiple class action lawsuits filed against them because they released Coffee lake knowing full well it was vulnerable, and now we're talking about releasing another round of vulnerable, or hamstrung, if patched by way of OS and firmware, processors AFTER the fact? Seems incredibly risky from a PR standpoint if not a legal one.

And as far as the performance issues are concerned, unless you are willing to stick your head in the sand it's kind of hard to ignore the reality which is not as you've described. We've seen TONS of data already that refutes what AMD and Intel have been piling on about there being minimal decreases in performance. When architectures seem to barely gain more than ten percent per generation, then taking a ten percent hit while gaming or up to a thirty pecent hit depending on workload for PCIe and SATA SSD performance, is pretty damn noticeable in my book.

I don't think I need to post links to the data on that, since it's plastered all over this and many other forums where the ACTUAL data from testers is getting dumped, so if you want to believe the sugar coated crap the damage control people are slinging to the front pages and the media, be my guest, but much as with the ostrich whose head is in the sand, you're going to get either run down or eaten whether you can see the danger coming or not.

Maybe you're ok with a ten percent hit in some types of performance and a thirty percent hit in others, for the record, I'm really not ok with that.

 

[Darkbreeze]

since the exploit code needs to run locally, simply not executing code of questionable origin would keep 99.9% of normal people safe. If you can't do that, then you have much worse security issues to worry about.

The problem with this is, we DO know that most people DON'T have secure systems. They have very f^$#%* unsecure systems, even MOST businesses networks and workstations are NOT secure in any meaningful sense. That means those people are vulnerable to this as well. And since THOSE unsecured systems have access to other systems, secured or not, it makes for an incredible potential for risk.

When the lazy bastard manager of some third rate grocery chain doesn't give a rats ass about whether the store system is secure or not, at least not in any meaningful way, and that system gets infected, and then goes on to infect the main system which all of the stores systems connect to which then goes on to maybe infect the databases or other systems of the financial institutions THEY connect to, seems like that could get really bad, really fast, especially since there won't actually be any kind of "definition" to guard against since it won't actually be a specific worm or trojan that's infiltrating the system anyhow.

I'm sure you're far more advanced in this area than I am, it just seem like a really unpredictable, potentially catastrophic scenario looking in on it. I mean, we've seen nearly every major company and half the governement organizations that exist get hacked in the last two years. Google, Target, Wal Mart, Amazon, DOD, the Democratic national party headquarters, list goes on an on. This just seems like stacking cans of gasoline next to the already burning apartment building, to go ahead and sell hardware you fully well know is at risk.

 

[InvalidError]

And since THOSE unsecured systems have access to other systems, secured or not, it makes for an incredible potential for risk.

The risks for an average PC user are grossly over-stated. First, in order to exploit Meltdown or Spectre, you already need to get malware which uses Meltdown and Spectre on the machine. At this point, you already have far worse security issues than Meltdown and Spectre to worry about in the form of software with remote code execution flaws or someone running questionable binaries on your machines. Regardless of Meltdown and Spectre, you need to address your remote execution software and irresponsible users first. Prevent malware from getting on your PCs in the first place and you'll never need to worry about Spectre and Meltdown.

Second, unless your computer runs large amounts of security/privacy-critical transactions, the chances of Meltdown or Spectre successfully managing to extract useful data from other processes is low to nonexistent. Side-channel attacks are inherently slow due to the limited scope of what they can observe, so unless the side-channel attack is running on a system handling high volume secure transactions that share a limited subset of keys, the chances of the exploit code running at the same time as the critical parts of processes they want to eavesdrop on and the right CPU core to be able to observe said process' behavior often enough to infer data is very low.

Even for datacenters, although the exploits have been demonstrated under controlled conditions, it doesn't necessarily make them practical in a real-world environment. These patches are mainly about plugging possible leaks before someone finds a way to reduce the effort/luck required for a successful exploit or gets lucky.

 

[Darkbreeze]

I get it, really. I understand and obviously there's very little reason to disagree. Totally sensible explanation. I still think though that I'd want to adhere to the exploding gas tank policy. Just because most vehicles with a recall won't get hit in exactly the right spot to cause it to instantly explode is no reason to press your luck by continuing to drive one around that has that vulnerability. My luck, it would happen twice. LOL.

 

[Isokolon]

.

First of all, there are NO systems that nobody gives a damn about, because nearly all systems either ARE or can be connected to all other systems. So that's a senseless comment. My neighbors refrigerator can talk to a server in California which can talk to an entire network of systems, and so on, but you know this and are just being snarky. So EVERY system is a potential liability if it's vulnerable, especially if it's accessing your home network, which generally, it is.

And as I said before, yes, it's fully understood that there needs to be some other form of malware or exploit already in place on the target machine for any of these particular strains to be dangerous, but the problem is that as we've discovered there are PLENTY of systems out there, refrigerators, home computers, phones, ovens, whatever, that have already been shown to be particularly vulnerable to the kinds of infections that need to be present for this to happen, so it's really not a stretch to say that this can and likely will happen at some point.

How many people do you know of that are going to patch the software on the refrigerator, or update it's firmware. Practically none, so if the hardware is vulnerable and you are selling it that way knowingly, that can't be acceptable to anybody with a half a brain.

I agree that Intel is "unlikely" to scrap the whole gen, but I also think it's incredibly stupid and short sighted not to. They're already in hot water, have multiple class action lawsuits filed against them because they released Coffee lake knowing full well it was vulnerable, and now we're talking about releasing another round of vulnerable, or hamstrung, if patched by way of OS and firmware, processors AFTER the fact? Seems incredibly risky from a PR standpoint if not a legal one.

And as far as the performance issues are concerned, unless you are willing to stick your head in the sand it's kind of hard to ignore the reality which is not as you've described. We've seen TONS of data already that refutes what AMD and Intel have been piling on about there being minimal decreases in performance. When architectures seem to barely gain more than ten percent per generation, then taking a ten percent hit while gaming or up to a thirty pecent hit depending on workload for PCIe and SATA SSD performance, is pretty damn noticeable in my book.

I don't think I need to post links to the data on that, since it's plastered all over this and many other forums where the ACTUAL data from testers is getting dumped, so if you want to believe the sugar coated crap the damage control people are slinging to the front pages and the media, be my guest, but much as with the ostrich whose head is in the sand, you're going to get either run down or eaten whether you can see the danger coming or not.

Maybe you're ok with a ten percent hit in some types of performance and a thirty percent hit in others, for the record, I'm really not ok with that.

I feel touched that you care about my grandmother's computer which she solely uses to write a letter in OpenOffice and mail me the document to print it out

Joke's aside: yes, unpatched systems are an issue. But as described by others, they are an issue anyway, no matter if Meltdown happened or not. Yes, Meltdown and Spectre are really dangerous. But by the time you can exploit it, any system is already compromised to a degree that Meltdown & Spectre are the tip of the iceberg.

I do reckon that with servers and cloud services, this is a real threat and it further encourage my scepticism of cloud services that aren't local (meaning not connected to the internet, on a separated network). But that's not what I was talking about. I was talking about consumer grade CPUs. those i3-9100s, i5s and i7s. Yes, I've read the test concluding that database work is heavily slowed down by this. and it really sucks. But - and I may stand corrected, but I'm pretty sure tbh - the majority of people are still using their computer for light office tasks, correspondence (whether it's emails, messengers or social media), Netflix (and maybe gaming) and if your machine wasn't already struggling with these tasks I still haven't seen data indicating you'll have to fear a severe performance decrease. I happily revert my point of view if there's evidence in wrong, so far I haven't seen any -- maybe you know more than I do?

As for being fine with that - I guess there are different breeds of people. I personally don't care if the eSATA port on my mainboard is broken since I don't use it and it hasn't been the feature I bought my board for.
Same with my car - I don't care whether it's able to go 190 or 220 per hour, it's well above the speed limits and I won't be able to drive that fast anyway. As long as the acceleration, breaks & comfort is still what I'm expecting, I'm fine with that. I do reckon there are people who have a different opinion on this, but after years in customer support I can say, I hate people who are hysterically complaining about things that don't concern and apply to them.

Also this issue is exploitable since...I know all of the i-core CPUs got this issue but I think it was the same with CPUs since the early 2000s? Late 90s? And the world is still alive. Don't get me wrong, it is a major concern but the hysteria going on is a bit unwarranted. There's people worried that they can't upgrade their FX-6300 to an i5-8600k for their gaming rig bcs of meltdown, not because of security but because of performance issues. And that's just bloody stupid. I blame that on the hype-hysteria with which

As for refrigerators and the IoT, I must say that I know too little about it. I don't see a reason to connect my refrigerator to the internet in the first place. I don't see any benefit in doing so. It's one of these "because I can" things. Sure, I can also only eat at McDonald's, but what's the benefit of it?

As for Intel fixing this and skipping lines -- and as a leftist I really hate myself saying this -- I hope they don't. Addressing this flaw will probably take major resources. That costs a lot of money.
Last news I've heard of cannonlake is that the 10nm process (do you say process? English isn't my native language)is giving them major headaches, the CPU 'otherwise' in terms of features, design, architecture is finished.
Scrapping this line completely would be foolish, a major loss with 3 possible outcomes:
- Intel goes out of business (which would be pretty bad considering their know-how and the potential lack of competition arising. Without Ryzen I'm certain we wouldn't have seen Intel mainstream hexacores.
- Intel getting vulnerable to hostile takeovers (by companies who may not share the same goals and are more interested in their parents than their consumer/server CPUs)
- Intel having to downsize and therefore not being able to dedicate as much resources to fixing this as we'd like them to, resulting in insufficient R&D, such as badly designed new architecture with flaws on their own, major bugs or a general hold in CPU performance increases

So from a totally egotistic point of view I hope Intel sells a shitload of CPUs while losing some stock market value, just enough so they start to work thoroughly again.
And of course people patching their systems. (While I personally find win10's forced updates annoying I do understand why it was implemented like that..)

 

[Darkbreeze]

Well, all I can say is you must be an asshat if you don't drive over 100mph.


Ok, that is a lie. Heh.

However, consumer processors is a very minor portion of Intel's overall business model. They could throw the entire desktop and laptop product lines in the trash and they'd barely notice it, at all. So while I DO fully see your point, and I agree that as far as THAT segment is concerned it IS or rather COULD be a big problem to do something like that, the biggest harm from it would be in public relations, not in cash infusion. Nobody is going to be taking over Intel. It's pretty doubtful even that any of the lawsuits against them will be allowed to go very far. I just don't see any of that happening.

Intel has yearly gross income comparable to several small countries, so a few million or even a billion in R&D might sting, but it definitely wouldn't hurt them in any impactful way.