Intel Process Technology Update: 10nm Server Products in 1H 2020, Accelerated 7nm in 2021 (anandtech.com) 40
Intel provided an update regarding its upcoming fabrication technologies at its 2019 Investor Meeting today. From a report: The company is on track to produce server-class products using its 10 nm manufacturing technology already in the first half of 2020. What is relatively surprising is that the company intends to start production of chips using its 7 nm process already in 2021. Intel's 7 nm production technology had been in development independently from the 10 nm process and by a different crew, so this one is closer than one might think. The node technology is set to use extreme ultraviolet lithography (EUVL) with laser wavelength of 13.5 nm for select layers, so it will not heavily rely on multipatterning, the source of problems with Intel's 10 nm process. In fact, the use of EUV will simplify development of products, make it easier to produce them, and will likely shorten production cycle times too.
Somebody is really late... (Score:3, Interesting)
Also, have they finally fixed Spectre and Meltdown? I find the silence _very_ suspicious. Are they hoping they can just continue as they are and not lose speed they obviously should never have gained...
Re: (Score:2, Informative)
There is no silence. You just don't know how to listen. Spectre and Meltdown mitigations are in Cascade Lake hardware. Which is already on the market.
Re: (Score:1, Interesting)
Mitigations, not fixes. New versions of these flaws are being exploited all the time, the design is shite. They tried to take a shortcut to beat AMD, they got it shorted alright. Intel shit the bed, face it kiddo.
They haven't fixed that yet.
Re: (Score:2)
Re: (Score:2)
There's nothing suspicious about silence, not about their shipping units not about the share price. The reality was clear from the start, is an overblown problem that doesn't affect most use cases. Intel's response to it now is appropriate.
Re: (Score:1)
Also, have they finally fixed Spectre and Meltdown?
Intel reported that it did (expect Spectre variant 1): https://www.anandtech.com/show... [anandtech.com]
AFAIK, Spectre variant 1 affects all processors with speculative execution, and there is no known fix for it without unacceptable high performance penalty.
Re:Intel's back, blaby! (Score:5, Funny)
Would someone tell me how this happened? We were the fucking vanguard of overpriced easily-hackable processing in this country. The Intel Pentium Pro was the FPU to own. Then the other guy came out with a "functional" FPU. Were we scared? Hell, no. Because we hit back with a little thing called Hyperthreading. That's three threads, pause, and an OPU strip. For data cloisture. But you know what happened next? Shut up, I'm telling you what happened—the bastards went to four threads. Now we're standing around with our clocks in our hands, selling three threads and a backdoor. Marketing or no, suddenly we're the chumps. Well, fuck it. We're going to five threads.
Sure, we could go to four threads next, like the competition. That seems like the logical, business savvy thing to do. After all, three worked out pretty well, and four is the next number after three. So let's play it safe. Let's make a thicker cache strip and call it the Pentium3SuperTurbo. Why innovate when we can follow? Oh, I know why: Because we're a failing business, that's why!
You think it's crazy? It is crazy. But I don't give a shit. From now on, we're the ones who have the edge in the multi-thread game. Are they the best a man can get? Fuck, no. Raped retail is the best a man can get.
What part of this don't you understand? If two threads is good, and three threads is better, obviously five threads would make us the best fucking space heater that ever existed. Comprende? We didn't claw our way to the top of the always-on backdoor game by clinging to the two-die industry standard. We got here by taking chances. Well, failing to address our IME backdoor is the biggest chance of all.
Here's the report from "Engineering" aka actually Marketing. Someone put it in the bathroom: I want to wipe my ass with it. They don't tell me what to invent—I tell them. And I'm telling them to stick two more backdoors in there. I don't care how. Make the Hypervisors so thin they're invisible. Put some on the die. I don't care if they have to cram the hijack in perpendicular to the user's choice, just do it!
You're taking the "security" part of "secure computing" too literally, grandma. Cut the strings and soar. Let's hit it. Let's roll. This is our chance to make customer-fuicking history. Let's dream big. All you have to do is say that five backdoors can happen, and it will happen. If you aren't on board, then fuck you. And if you're on the board, then fuck you and your father. Hey, if I'm the only one who'll take risks, I'm sure as hell happy to hog all the glory when the five-backdoor overpriced handjob becomes the data-raping tool for the U.S. of "this is how we snoop now" A.
People said we couldn't go to Hell. It'll cost a fortune to manufacture, they said. Well, we did it. Now some egghead in a lab is screaming "Marketing's crazy?" Well, perhaps he'd be more comfortable in the labs at Nvidia, working on fucking actual discrete video sytems that work. Intel HD-video, my white ass!
Maybe I'm wrong. Maybe we should just ride in AMD's wake and make affordable security-conscious-slightly working CPUs. Ha! Not on your fucking life! The day I shadow a penny-ante outfit like AMD is the day I leave the marketing game for good, and that won't happen until the day Intel dies!
Re: (Score:2)
5 years later... (Score:4, Informative)
It was explained that while 10nm will have more masking steps than 14nm, and the delays that bogged down 14nm coming late to market will not be present at 10nm â" or at least reduced. We were told that Intel has learned that the increase in development complexity of 14nm required more internal testing stages and masking implementations was a major reason for the delay, as well as requiring sufficient yields to go ahead with the launch. As a result, Intel is improving the efficiency testing at each stage and expediting the transfer of wafers with their testing protocols in order to avoid delays. Intel tells us that that their 10nm pilot lines are operating 50% faster than 14nm was as a result of these adjustments.
https://www.anandtech.com/show/8991/intel-at-isscc-2015-reaping-the-benefits-of-14nm-and-going-beyond-10nm [anandtech.com]
Kinda expected (Score:1)
When there was no longer any desktop 10 nm parts for both 2019 and 2020 in the roadmaps. Eventually 7nm would be ready and it wouldn't make any sense to bother any longer.
Since it happen in server parts I guess either it's capacity or cost which make it not a thing for desktop.
I thought 7nm was below physics limits (Score:2)
I thought anything smaller than 10nm wasn't possible, because of the EMI between the pathways. Is this a problem that got solved through better technology, was I wrong about the limit of shrinking or was I just mistaken about the EMI limit.
Re: (Score:1)
The 7/10nm figures being thrown around by Intel and AMD have nothing to do with the actual sizes of transistors and gates.
https://en.wikichip.org/wiki/10_nm_lithography_process
Re:I thought 7nm was below physics limits (Score:4, Informative)
Well, first the node numbers are now marketing numbers, and are quite detached from physical dimensions. What counts is transistor density, actually Intel 10nm has a higher transistor density than TSMC's 7nm. But the difference is that the first one is still essentially vapor, while the second one has been shipping for about one year, initially in small quantities, but they are on the rise.
Re: (Score:1)
Well, first the node numbers are now marketing numbers, and are quite detached from physical dimensions. What counts is transistor density, actually Intel 10nm has a higher transistor density than TSMC's 7nm. But the difference is that the first one is still essentially vapor, while the second one has been shipping for about one year, initially in small quantities, but they are on the rise.
I wonder how the TSMC 5nm compares to the Intel 7nm? TSMC 5nm is in risk production already, and Intel could possibly catch up with 7nm.
Re: (Score:2)
90 nm was the old limit. Special tricks were applied just to get that far. But that would have been assuming visible spectrum light sources I presume.
I guess if they can find a way to focus xrays then they'll go there too.
Re: (Score:2)
You're talking about the technology used to fabricate. I was talking about the actual operation of a chip and it's self-interference
Not a whole lot of confidence (Score:2)
Personally, I do not have a whole lot of confidence that Intel will hit its 2021 target for 7nm (the node that the rest of us will call 5nm.) They just basically admitted that 10nm is delayed _another_ year. They hope to jump straight into 7nm, that is EUV with ~32nm metal 1 pitch, without passing Go, without collecting experience at 36nm metal 1 pitch. EUV is going to just work, and solve all their problems in one grand sweeping coup. Good luck with that!
Re: (Score:2)
Who cares what Apple does?
Re: (Score:2)
Apple switching to Arm processors isn't something to really worry about. They're move from PowerPC to Intel was relatively pain free. If you want to worry about something is Apple's lack of focus on their computers. The last few years especially they haven't really seemed to care about the computer line, especially the Mac Pro, and their user base is slowly going away.
Instead of bringing out new up to date models of everything, and getting rid of the trashcan Mac Pro for a tower or something upgradable, App