Intel CPUs: Ten-year plan to includes 1.4nm and a two-year cadence

Dimitrios1983

2019-12-13 17:02

If these goals aren't met with INTEL, PLAN B --------------------> Pour sugar in Lisa Su's gas tank & Vandalize TSMC production. ;-)

#5741179

Reddoguk

2019-12-13 17:03

What a beautiful looking chart. Shame it won't go like that in reality.

#5741185

RooiKreef

2019-12-13 17:18

Hahahaha! For now I will take Intel's word with a grain of salt. But with that said I'm glad they are so positive and I hope their plans work out.

#5741206

Cyberdyne

2019-12-13 18:06

angelgraves13:

... Also their 10nm isn’t as dense as 7nm from TSMC. It was on paper but they’ve since scaled it back to get it out the door. There’s no way they can compete with TSMC or even Samsung, who will be TSMC’s main competitor by 2030. Intel’s architecture is a decade old and they don’t have a new one coming for at least another half a decade or more. Then there’s their fabs....a money pit. They can’t compete on either front...

asturur:

... You guys, again, you want a fast CPU, you do not want a piece of technology. Some of you would buy a 5nm 386 (at least is 32bit ) if it was available.

#5741213

Gomez Addams

2019-12-13 18:13

Robbo9999:

However, TSMC's 5nm does not equal Intel's 5nm - the densities can be quite different (ie transistors per mm^2)..... currently Intel has a higher density per mm for any given Xnm node. But yes, I'm not saying that Intel will be able to match the race against other fabs though.

In my opinion, that is one of the causes of the issues they are having. TSMC, by having a lower device density, are being more conservative and with the dimensions constantly shrinking I believe that is very important. I think they need to be more conservative at every subsequent generation. They can't just say we're going from 14 to 7 so we'll just half every dimension and off we go. Leakage and tunneling do not increase in a linear way so densities can't either. If I were Intel, I would step back a bit and decrease the density and proceed more conservatively. Actually, the key to the process shrinkage roadmap is going to be lithography systems. They are already the primary limiting factor and there are two aspects to that. The first one is the physics of light. Dimensions of just a few nanometers are in the same range as wavelengths of light and reactions of light with a crystal lattice get "dicey" there, to put it in technical terms. That ties into the other factor - masking materials. There is a huge amount of work being done creating materials to use in photomasking that react to very small wavelengths of light and this is very challenging work. Photomasking is basis for how lithography machines work.

#5741215

NCC1701D

2019-12-13 18:15

I will admit to being shocked that Intel doesn't have a new arc ready to release in light of AMD's recent success. I kept thinking, it's coming, they'll make a big announcement soon, but nope. I'm still not counting them out and proclaiming their doom. They have the money and now the kick in the pants, so it will be interesting to see what they do in the next few years. That said, AMD got my money for a 3700x and it's the first AMD CPU I've ever built with. If Ryzen 4K closes the small gaming advantage gap, then Intel is in even more trouble in the DIY build market. Still a lot of folks that I know going for the 9900K(S) series when making their final build decisions.

#5741229

Embra

2019-12-13 18:37

I was hoping that in ten years we might start seeing some totally new tech. They can only shrink so far... and that is not far from where they are now.

#5741232

Corrupt^

2019-12-13 18:59

NCC1701D:

Still a lot of folks that I know going for the 9900K(S) series when making their final build decisions.

I did, system is currently mostly used for gaming. But if by the time DDR5 is released AMD is better I'll probably swap. I actually gained more FPS then I expected, upgrading from a 5820K @4.25Ghz to a 9900KS. Games have become so CPU bound lately :x A swap to AMD was pretty much the same cost as I would've needed faster RAM.

#5741236

JAMVA

2019-12-13 19:06

https://i.postimg.cc/44qqjwJN/UNTEL.jpg What happened to 10+ 🙂

#5741254

NCC1701D

2019-12-13 19:49

angelgraves13:

It takes about 5 years to design and bring to product a new architecture. When would they have time to do that? They hired Jim Keller to fix their security issues. There's no indication that he's started work on a new architecture or that it'll be out before 2025. AMD is changing their architecture every year now. Intel can't compete.

I don't have the answers to those questions. Literally have no idea what Intel's been working on behind closed doors other than road maps, or if they're waiting solely on Jim Keller to have free time to build their new architecture. Last I heard that was what he was working on for them, but I can't pin that down to anything solid since I don't have the inside info other than leaks and heresay from the internet. I'm just not willing to say Intel can't compete with any absoluteness. I'd argue that they're still competing pretty well in the current market, even with AMD's recent success. I hope they can and will continue to remain competitive. That's a best case scenario for us consumers. AMD did big things with a lot less money. Intel can do the same with a lot of money. Time will tell. I think the pressure from AMD will at least speed their efforts.

#5741263

Loobyluggs

2019-12-13 20:10

NCC1701D:

I'm just not willing to say Intel can't compete with any absoluteness.

They do other stuff besides CPU's on Desktops, and their business model is one of the best in the USA as they literally have an effective market control, without breaking the law. It really is quite astonishing.

#5741274

svan71

2019-12-13 20:41

Intel CPUs: Ten-year plan to includes 1.4nm and a two-year cadence Makes me wonder in 2014 was their 5 year plan 14nm,14,nm,14+nm,14++,nm,14nm++++++ so much for charts and plans.

#5741276

GamerBoyManuel16

2019-12-13 20:44

svan71:

Intel CPUs: Ten-year plan to includes 1.4nm and a two-year cadence

Makes me wonder in 2014 was their 5 year plan 14nm,14,nm,14+nm,14++,nm,14nm++++++ so much for charts and plans. Tbh let's hope they manage it to get back on track Otherwise there is just TSMC and Samsung left for making modern Chips 🙁 It would be better if there is any / more competition

#5741297

dfsdfs1112

2019-12-13 22:16

Intel tells this story on slides from 2013 And you eat all the nonsense

#5741301

tsunami231

2019-12-13 22:33

I would like to see this, in 10 years lets see how true this slide is

#5741310

user1

2019-12-13 22:54

good to see they have a provision for backporting designs, in the event of node failure, dont really know why they didn't have that before, 14nm icelake would have been much better than skylake 6.0

#5741312

H83

2019-12-13 23:03

NCC1701D:

I will admit to being shocked that Intel doesn't have a new arc ready to release in light of AMD's recent success. I kept thinking, it's coming, they'll make a big announcement soon, but nope. I'm still not counting them out and proclaiming their doom. They have the money and now the kick in the pants, so it will be interesting to see what they do in the next few years. That said, AMD got my money for a 3700x and it's the first AMD CPU I've ever built with. If Ryzen 4K closes the small gaming advantage gap, then Intel is in even more trouble in the DIY build market. Still a lot of folks that I know going for the 9900K(S) series when making their final build decisions.

That´s not exactly true. Intel has a new CPU arc and they have already released but only for laptops:https://www.anandtech.com/show/15092/the-dell-xps-13-7390-2in1-review-the-ice-lake-cometh Also this arc has been ready for at least 2 or 3 years, the problem is that the same arc is tied to their broken 10nm process and because of that it´s being continually delayed for desktop. Intel really dropped the ball with their 10nm process.

#5741321

DmitryKo

2019-12-13 23:58

It's still not "doom and gloom" for Intel. Even though they really botched up their 10 nm node to the point of cancelling 10 nm desktop processors, they've managed to keep their 14 nm node competitive on the performance front, even with restricted transistor budgets, so their x86 CPU market share is still above 80% at this point. DYI enthusiasts like me are disappointed by the recycling of the same architecture and lack of exciting new features, but it's no different than the last 10 years, when AMD was unable to come with competitive CPUs. If Intel also botches their 7 nm node in 2022, which will result in another round of supply shortages, that would become a real problem; by then, Zen4 shall arrive with DDR5, PCIe 5.0, and AVX-512 on the 5 nm TSMC node, and OEMs will be already tempted to switch to AMD. But even if they botch it, backporting to previous nodes is the new backup plan for Intel - which starts with the porting of 10 nm Willow Cove architecture to the 14 nm Rocket Lake in 2021, with improved IPC, PCIe 4.0 and TB3/USB4. There are all indications that Intel learned from their mistakes though, and it doesn't seem like AMD is counting on them making further mistakes either. https://wccftech.com/amd-we-never-dreamed-we-would-be-ahead-of-intel/ https://seekingalpha.com/article/4312224-advanced-micro-devices-inc-amd-presents-barclays-global-technology-media-and?part=single

#5741345

anticupidon

2019-12-14 02:33

I rather like to see a 5 years plan, more affordable prices for CPUs without security holes. Oh, wait, that's AMD.

#5741363

Neo Cyrus

2019-12-14 03:47

Hay guise, I have 1.4nm bridges to sell, sign here.