Intel's 8th Generation Core Processors Won't Be Based on 10nm but 14nm

[DocSwag]

Source: http://www.anandtech.com/show/11115/intel-confirms-8th-gen-core-on-14nm-data-center-first-to-new-nodes

Quote

A quick news piece on information coming out of Intel’s annual Investor Day in California. As confirmed to Ashraf Eassa by Intel at the event, Intel’s 8th Generation Core microarchitecture will remain on the 14nm node. This is an interesting development with the recent launch of Intel’s 7th Generation Core products being touted as the ‘optimization’ behind the new ‘Process-Architecture-Optimization’ three-stage cadence that had replaced the old ‘tick-tock’ cadence. With Intel stringing out 14nm (or at least, an improved variant of 14nm as we’ve seen on 7th Gen) for another generation, it makes us wonder where exactly Intel can promise future performance or efficiency gains on the design unless they start implementing microarchitecture changes.

Quote

Despite this, if you were to believe supposed ‘leaked’ roadmaps (which we haven’t confirmed from a second source as of yet), the 8th Generation product ‘Cannon Lake’ is more geared towards the Y and U part of Intel’s roadmap. This would ring true with a mobile first strategy that Intel has mirrored with recent generations such that the smaller, low power chips are off the production line for a new product first, however we'd also expect 10nm to also be in the smaller chips first too (as demonstrated at CES). Where Cannon Lake will end up in the desktop or enterprise segment however remains to be seen. 

So this is interesting... We thought that Intel was simply transitioning from Product-Architecture-Optimization from the traditional tick-tock cycle, but now we're gonna be on 14nm for 4 full generations. This was definitely unexpected and leaves us wondering how much difficulty Intel is going through for 10nm. As well, it also leaves us wondering if we won't be seeing Cannonlake for consumers until 2019. Intel keeps touting that they're 14nm process is superior to competitors 14nm and 16nm processors and is on par with competing 10nm processes, but if Intel keeps delaying then others may be on 7nm when Intel is on 10nm, etc. What exactly will happen though still remains to be seen.

[shadowbyte]

amd pls

[Trav_X]

What does that change for Coffee Lake

[DocSwag]

1 minute ago, shadowbyte said:

amd pls

Zen is 14nm tho

Zen+ maybe in 2018 on 10nm?

[shadowbyte]

1 minute ago, DocSwag said:

Zen is 14nm tho

Zen+ maybe in 2018 on 10nm?

hopefully

[DocSwag]

1 minute ago, Trav_X said:

What does that change for Coffee Lake

Coffee Lake wasn't really confirmed before this. Just a rumor. Now I guess it's something that's real...

[79wjd]

5 minutes ago, shadowbyte said:

amd pls

It's also in Intel's best interest to get smaller nodes as that means cheaper production costs. 

[FratStar]

But didn't we know this already since they moved away from the Tick Tock cycle?

[MageTank]

14 minutes ago, DocSwag said:

Coffee Lake wasn't really confirmed before this. Just a rumor. Now I guess it's something that's real...

But Intel had Coffee Lake on their official roadmap like, 4-5 months ago. 

 

We also knew that it wasn't going to be 10nm either, because of the low yields (Cannonlake will be 10nm, but only for mobile). Even it's wikipedia page stated this ever since October. While it's nice to have a legitimate confirmation, it shouldn't really surprise anyone given how long this information has been circulating (and with Intel themselves mentioning their departure from the standard tick-tock formula in favor of process-architecture-optimization/refinement)

[iamdarkyoshi]

I wonder if they had reliability issues with the smaller size. Might need more time to perfect. Also, are they just not going to use any TIM this time? Seems like something they would do...

[The Sloth]

7 minutes ago, MageTank said:

But Intel had Coffee Lake on their official roadmap like, 4-5 months ago. 

 

We also knew that it wasn't going to be 10nm either, because of the low yields (Cannonlake will be 10nm, but only for mobile). Even it's wikipedia page stated this ever since October. While it's nice to have a legitimate confirmation, it shouldn't really surprise anyone given how long this information has been circulating (and with Intel themselves mentioning their departure from the standard tick-tock formula in favor of process-architecture-optimization/refinement)

 

just image one day we might have an i7 in something like a raspberry pi. 

[TrigrH]

Tick-Tock-Cock-Flop

[MoonSpot]

41 minutes ago, DocSwag said:

Zen is 14nm tho

Zen+ maybe in 2018 on 10nm?

Thought AMD/Global Foundries was going to try for a jump to 7nm and skip 10nm.  Didn't they pickup assets to do so from IBM a while ago?

[Guest]

21 minutes ago, nerdslayer1 said:

just image one day we might have an i7 in something like a raspberry pi. 

crazy right, soon rpi can run crysis

[Rune]

Ew.

[ARikozuM]

1 hour ago, Trav_X said:

What does that change for Coffee Lake

Decaf Lake?

[iamdarkyoshi]

9 minutes ago, ARikozuM said:

Decaf Lake?

Those will be the non overclockable ones

[Qwertious]

My next build will be a Coffee Lake machine, nothing is is better than a super optimised process node. The 14nm node sounds perfect. A super optimised 7nm node will be half a decade away probably!

[dizmo]

They're already planning to start the 7nm fab. So. Clearly it's on it's way.

I wonder if Samsung or Intel will get there first. TSMC is already well on their way to 7nm production.

 

8 minutes ago, Qwertious said:

My next build will be a Coffee Lake machine, nothing is is better than a super optimised process node. The 14nm node sounds perfect. A super optimised 7nm node will be half a decade away probably!

Half a decade isn't a long time

 

[79wjd]

12 minutes ago, dizmo said:

They're already planning to start the 7nm fab. So. Clearly it's on it's way.

I wonder if Samsung or Intel will get there first. TSMC is already well on their way to 7nm production.

 

Half a decade isn't a long time

 

Isn't TSMC/GFs 16nm/14nm more like Intel's 22nm~?

[Liltrekkie]

Remember when AMD basically sat on their butt for 4 years on the consumer CPU market and everyone bitched at them because they made no improvements to their processors? 

 

That's basically what Intel is doing here. Skylake was the introduction, like bulldozer (only it didnt suck as far as IPC goes but hear me out), Kaby lake is is Piledriver, some improvements but eh. And then cannon lake will be like the 9590 from AMD. 

 

I seriously do not imagine that while on the same node, cannon lake will offer much in terms of IPC improvements or efficiency. It might offer a new feature or two and maybe 1% IPC improvement but that's it. 

 

Everyone should raise up their pitchforks to intel like they did AMD. 

[Juanitology]

2 hours ago, djdwosk97 said:

It's also in Intel's best interest to get smaller nodes as that means cheaper production costs. 

It may cost less in terms of silicon/chip but it means investing billions of dollars in new fabs and fab equipment. Just yesterday intel announced it will spend 7 billion dollars on one fab to make it 7nm-ready. 

[Liltrekkie]

3 minutes ago, Terodius said:

It may cost less in terms of silicon/chip but it means investing billions of dollars in new fabs and fab equipment. Just yesterday intel announced it will spend 7 billion dollars on one fab to make it 7nm-ready. 

Intel earned a record 56 billion with a net income of 11.2 billion in 2015 or something.  

 

I really don't think Intel spending 7 billion is an issue at all. If anything, them spending 7 billion means some 7 nm process at least works in some regards. They wouldnt spend money upgrading to 7 nm if the node didn't work. 

 

Intel is just sitting on their butt with the consumer CPU market until 7 nm, which isn't okay at all.

[Juanitology]

42 minutes ago, Liltrekkie said:

Intel earned a record 56 billion with a net income of 11.2 billion in 2015 or something.  

 

I really don't think Intel spending 7 billion is an issue at all. If anything, them spending 7 billion means some 7 nm process at least works in some regards. They wouldnt spend money upgrading to 7 nm if the node didn't work. 

 

Intel is just sitting on their butt with the consumer CPU market until 7 nm, which isn't okay at all.

Except that 7 billion is for just one fab, and they have multiple ones making chips for every node. doesn't matter if you're intel or apple, 7 billion dollars is A LOT of money and that cost plus the billions of R&D that go into figuring out how to scale out the process plus the billions of R&D just for making the micro-architecture are a lot of upfront costs. I It's an investment that they factor into the costs of their chips. That's why we pay 300 USD for a piece of silicon that costs 20 USD in materials to make, because there's a median of like 100-150USD worth of R&D and infrastructure amortization in each CPU they sell and the rest is profit. 

[Space Reptile]

4 hours ago, DocSwag said:

Zen is 14nm tho

Zen+ maybe in 2018 on 10nm?

they are "working" on 7nm , so theres hope