Intel Coffee Lake (8th & 9th Gen Core CPUs) + Skylake-X Refresh & W-3175X MegaThread! FAQ and Resources

[-Fran-]

Oh, 10nm isn't dead! It's just crawling out of the door into Mobile and, at some point, servers!

I wonder when the Desktop crowd will see the scraps?

Cheers!

 

[jimmysmitty]

10nm was never dead and this has been known for quite some time. This is not the first time Intel pushed mobile first then went to desktop. In fact the entire Core series is thanks to the Pentium M which evolved into the Core Duo and Core Solo. They went out January of 2006 then came Core 2 Duo in July on desktop.

 

[-Fran-]

Wasn't Core 2 based of Tualatin's P3 design? Or at least, that's what the rumours say, haha.

In any case, that "10nm is going strong" is lies and everyone knows by now. Their revenue says as much. It's been in a hole for a long time now and it was about time they started getting their act together. It's been... What? 1 whole year already?

Cheers!

 

[jimmysmitty]

10nm going strong is does not mean financially. They could be talking about yields. And sure it would be nice to see 10nm to be out now but it is a complicated task. And no I don't count TSMCs 7nm considering Intels first approach at 10nm was going to be more dense than that by far. Maybe you should go and tell Intel to et their act together? I can go with, I live close to their Chandler fabs.

And yes if you go that far back it was but the first iteration of this was the Pentium M/Core Solo/core Duo.

Normally desktops and laptops share variants however this was a chip that was mobile only. It was Intels testing ground for Core, to see if this new uArch would be any good. And it was. I remember when they sold S478->S479 adapters so you could throw a Pentium M into a desktop and overclock the hell out of it to make it basically beat even top end P4 EEs and Athlon 64 FX chips.

 

[mdd1963]

2021? LOL!

That is just short of pathetic....a full 3 years late?

 

[jimmysmitty]

2021? LOL!

That is just short of pathetic....a full 3 years late?

Actually close to 6 since 10nm was originally planned for a 2015 release.

And sure, you could say its "pathetic". Then again I don't see you releasing a very complex and advanced process node that takes billions in R&D to get working.

Again you have to remember that Intels original 10nm was going to be more dense than TSMCs 7nm is. The denser a process is the harder it is to get it to work. AMD itself dropped out of the process FAB market because they couldn't afford to keep up with the bigger companies.

 

[TechyInAZ]

Yep that's one huge misconception about nm. It isn't all about size.

Intel's 10nm if done the way they wanted to originally, would be better than the current 7nm platform Ryzen is on.

 

[jimmysmitty]

Yep that's one huge misconception about nm. It isn't all about size.

Intel's 10nm if done the way they wanted to originally, would be better than the current 7nm platform Ryzen is on.

I think its still close, just not as dense as they wanted it to be. I think its still more dense than TSMCs current 7nm but will be slightly less dense than TSMCs second EUV based 7nm:

https://www.anandtech.com/show/13405/intel-10nm-cannon-lake-and-core-i3-8121u-deep-dive-review/3

Samsung vs TSMC 7nm Update - Semiwiki

The semiconductor foundry business has gone through a dynamic transformation over the last 30 years. In the beginning the foundries were several process nodes behind the IDMs with little hope of catching up. Today the foundries are leading the process development race at 10nm - 7nm, and will...

semiwiki.com

This was the latest information I could find.

And if it still pans out the way Intel wants, their 7nm is going to be twice as dense as Samsung or TSMCs 7nm EUV and probably more dense than TSMCs 5nm which is said to be 80% more dense than their 7nm:

https://www.pcgamesn.com/amd/tsmc-6nm-5-nm-80-percent-density

So that would put it at 203.4MTr/mm2 vs 237.18MTr/mm2.

Of course this is if it all pans out the way Intel wants but if their 7nm is able to more than double density it would be a good win for them.

 

[TechyInAZ]

Thanks for the links jimmy, interesting stuff.

If my memory serves me correctly, Intel had to forgo several key features of 10nm to make it to production. Since it's been taking them forever to produce 10nm, they were forced to start slicing off some stuff.