Tick Tock Tock Says Intel Now As Well

PrMinisterGR

2016-03-23 18:52

It was about time. 10nm will come at around 2018, two years after than expected. 7nm will be even harder, and 5nm is considered almost impossible with current materials/physics. Speed upgrades have slowed down very much the last 5 years, and they will keep doing so even more, until we have a paradigm shift.

#5249270

__hollywood|meo

2016-03-23 19:28

die shrinks are getting really hard, guys, leave us alone for a little bit while we figure this out, okay?

tl;dr version

#5249271

Matt26LFC

2016-03-23 19:28

Been on the cards really hasn't it. Not surprising at all since they've been having nothing but trouble reducing transistor size now. But what can we expect from canonlake now? It was mean't to be the die shrink of Skylake, which if it still is that means basically we get 3 versions of skylake what with Kabylake coming this year. Skylake>Kabylake>Canonlake all basically the same arch!? Or will Canonlake now be a whole new arch on the new process?

#5249274

Denial

2016-03-23 19:32

Been on the cards really hasn't it. Not surprising at all since they've been having nothing but trouble reducing transistor size now. But what can we expect from canonlake now? It was mean't to be the die shrink of Skylake, which if it still is that means basically we get 3 versions of skylake what with Kabylake coming this year. Skylake>Kabylake>Canonlake all basically the same arch!? Or will Canonlake now be a whole new arch on the new process?

Yeah same arch. The next architecture will be Icelake.

#5249275

CPC_RedDawn

2016-03-23 19:36

Still looks like my next CPU upgrade won't be til at least Canonlake or Icelake then. 10nm here we come!!! 😀

#5249287

Ven0m

2016-03-23 19:57

I wonder when they'll figure out how to put multiple layers into CPUs. This should allow huge performance increase without decreasing fab node.

#5249290

PrMinisterGR

2016-03-23 19:59

You'll wait at least a couple of years for that 😀

#5249296

Robbo9999

2016-03-23 20:04

I think there's a mistake in this article if I'm understanding correctly. The article says right at the beginning: "Based on Moore's Law this been proven to become more and more difficult, it's becoming Tick-Tick-Tock". I think it's actually supposed to be Tick-Tock-Tock instead, because a Tick is a die shrink & its becoming harder to die shrink. I sat in front of the computer screen trying to work out why I wasn't understanding the article, it was that incorrect statement at the start of the article that threw off my understanding! The table at the end of the article supports the fact that it's becoming Tick-Tock-Tock, not the Tick-Tick-Tock written in this article. (Just now noticed that title of article is correct though).

#5249302

__hollywood|meo

2016-03-23 20:10

I wonder when they'll figure out how to put multiple layers into CPUs.

now that will be expensive

#5249310

Hootmon

2016-03-23 20:17

I wonder when they'll figure out how to put multiple layers into CPUs.

The main problem with that is heat dissipation from the 'buried' components. You would need some sort of active in-die cooling system or a different breed of switch.

#5249315

Fender178

2016-03-23 20:27

I know my next upgrade will not be until 2020-2021 with one of the E based CPUs. Hopefully my CPU will last that long.

#5249321

Elder III

2016-03-23 20:51

Looks like I'll be fine awhile longer. My wallet is happy at least...

#5249334

Tugrul_512bit

2016-03-23 21:11

Maybe double layer possible with 2 coolers on both sides of mainboard. Then case would look more symmetrical. Wherever you open, you see a cpu. Mainboard needs to be thicker with a folding style opening for maintenance. This also gives more space for more ram slots and other things and needs more hardware components but this would be an enthusiast hardware.

#5249344

fantaskarsef

2016-03-23 21:33

Still looks like my next CPU upgrade won't be til at least Canonlake or Icelake then. 10nm here we come!!! 😀

hehe yeah, was thinking the same somehow

#5249350

-Tj-

2016-03-23 21:42

TigerLake it is then.. Was planing on CannonLake but seeing that chart made me pass. I think at 2019 its enogh for this i7 Haswell. 3yrs is not a long wait, should be 6yrs old then, 🤓 I already have it since its rls back in June 2013.

#5249355

tsunami231

2016-03-23 21:49

wasnt this chart show a few months ago? I kind curious when the "2016 optimize" chips are gona be released. mostly cause i still havent made my new build

#5249356

Fender178

2016-03-23 21:49

TigerLake it is then.. Was planing on CannonLake but seeing that chart made me pass. I think at 2019 its enogh for this i7 Haswell. 3yrs is not a long wait, should be 6yrs old then, 🤓 I already have it since its rls back in June 2013.

Yeah with me I'm going to wait it out until 2020-2021.

#5249377

xIcarus

2016-03-23 22:56

The main problem with that is heat dissipation from the 'buried' components. You would need some sort of active in-die cooling system or a different breed of switch.

That would probably require significant socket redesigns. First thing that comes to mind would be the processor to be inserted upright into some kind of plate protruding from the mainboard. The processor itself having the die on both sides. That would be terrible for reliability though, a falling heatsink would mean a snapped socket. The easier thing would be bigger dies, but as we all know - bigger dies are more expensive to manufacture. Or the even easier thing would be to ditch that faking IGP, it's way too big. Honestly, they could at least make the K versions with a primitive IGP - primitive enough to offer usable video outputs.

#5249413

fearsjohn

2016-03-24 02:17

That would probably require significant socket redesigns. First thing that comes to mind would be the processor to be inserted upright into some kind of plate protruding from the mainboard. The processor itself having the die on both sides. That would be terrible for reliability though, a falling heatsink would mean a snapped socket. The easier thing would be bigger dies, but as we all know - bigger dies are more expensive to manufacture. Or the even easier thing would be to ditch that faking IGP, it's way too big. Honestly, they could at least make the K versions with a primitive IGP - primitive enough to offer usable video outputs.

why some people like me actually want the igp for quicksync purposes. what they could do is offer 2 versions one with igp and 1 without.

#5249436

Irenicus

2016-03-24 04:44

Why does it say Ivy bridge is i3? Ivy bridge is i5?