Intel Clarifies on 10nm Desktop CPUs: Still on the Table, Likely in 2021

[btarunr]

Intel in a quick rebuttal to the earlier reports from Monday, clarified that desktop processors based on the 10 nm silicon fabrication node are still on the company's roadmap. "We continue to make great progress on 10 nm, and our current roadmap of 10 nm products includes desktop," the company said in its one-liner. Monday's reports predicted a horror story where Intel would drag its 14 nm "Skylake" derived microarchitecture through to 2022, at which point it would be 7 years old.

The Tom's Hardware report that posts the statement, however, pins 14 nm to still last till 2021, if not the 2022 date predicted in the HardwareLuxx report. Intel will sell "Comet Lake" through 2020, succeeded by "Rocket Lake," which takes up much of 2021. Towards the end of 2021, Intel will release a desktop processor based on its matured 10 nm++ silicon fabrication node, which will lead the company into 2022, when it finally launches 7 nm EUV-based desktop chips.



View at TechPowerUp Main Site

 

[Fouquin]

And thus completes the cycle of FUD.

 

[GoldenX]

Still pretty lame, that leaves AMD a full year to do as it pleases with an actual product stack.

 

[Metroid]

meaning intel still want you to sit still and wait for fairies.

 

[ncrs]

I love Intel PR. Optimistic as always...

 

[FordGT90Concept]

2021 which coincides with what Jim Keller is working on.

 

[halo9]

View attachment 134167
I love Intel PR. Optimistic as always...

“We continue to make great progress on 10 nm” Haha, yep right on schedule.

 

[seronx]

Umm.. Tigerlake is planned for a desktop release in 2020...
8-cores with Gen12 32 EUs(1x2x16) for 35W/65W/95W.

Cometlake-S/Tigerlake-S share the same socket.

 

[notb]

meaning intel still want you to sit still and wait for fairies.

No. Intel wants you to move on from DIY desktops. Business-wise it's the worst market at the moment.

Still pretty lame, that leaves AMD a full year to do as it pleases with an actual product stack.

On desktops: sure. Intel won't have a direct answer to 12+ cores.
All consumer 10nm supply is given to the mobile division. And here it's AMD who's a year behind.

Which makes me wonder how much 3950X will cost with no alternative from Intel; -)

 

[R0H1T]

All consumer 10nm supply is given to the mobile division. And here it's AMD who's a year behind.

6 months realistically speaking, ICL right now is competing with WHL & losing. When AMD 7nm APU hit, they should have a good number of design wins.

 

[GoldenX]

The fact that Intel needs HEDT to compete with a desktop part...
Pricing can start to get bad if AMD wants to.

 

[1d10t]

Suddenly dekstop doesn't matter anymore, funny these statement would have been different 2 - 3 years ago
Mobile the most important, yeah, thats all everyone need , anything more demanding you can always rely on cloud

 

[Vayra86]

As posted in the other topic about discontinued 10nm.

Where is the product? I've tried gaming on roadmaps but it just doesn't work.

Intel, stop lying, just admit defeat and move on. We know, shareholders won't like it. Let us know when you'll release that fantastic 2.3 Ghz base CPU, because we're still waiting. On a quad core part. In 2019-Q4.

 

[64K]

I think for 10nm Intel has made the transition from embarrassing to joke territory at this point.

iirc Intel originally projected release for 2016 and now..........2021........maybe.

 

[bug]

Do people actually care about statements like these?
When I'm eyeing an upgrade, I just look at what's on the market. If there's something that's significantly faster than what I have, available for the right price, I buy. I don't care whether it says Intel or AMD on the box, I don't care about nm (but I do care about TDP), I don't care about marketing departments' prowess.

 

[Basard]

Can't get them sweet, 5Ghz+ clocks out of it yet perhaps?

 

[birdie]

Here's the original statement:

"We continue to make great progress on 10nm, and our current roadmap of 10nm products includes desktop."

What people are overlooking is that this is a nonbinding assertion which doesn't mean anything at all as roadmaps change all the time. Intel can't even say for sure whether they will have desktop 10nm parts. From the look of it they won't, or, as mentioned in the article, those "desktop" parts could include CPUs for NUCs which are formally a desktop product, yet technically they are not.

 

[Rahmat Sofyan]

still on track ... yeahhh ..

 

[phanbuey]

Here's the original statement:

"We continue to make great progress on 10nm, and our current roadmap of 10nm products includes desktop."

What people are overlooking is that this is a nonbinding assertion which doesn't mean anything at all as roadmaps change all the time. Intel can't even say for sure whether they will have desktop 10nm parts. From the look of it they won't, or, as mentioned in the article, those "desktop" parts could include CPUs for NUCs which are formally a desktop product, yet technically they are not.

Yup... good point.

 

[efikkan]

And thus completes the cycle of FUD.

I don't know why tech news outlets gave this claim another news cycle, it's the same BS that was in circulation several months ago.

10nm desktop parts are not canceled, not yet anyway. Intel added driver support for 10nm desktop parts just 1.5 months ago, they wouldn't do this if they had already canceled the parts. This support specifically mentions 95W desktop and entry workstation parts.

Plans are always subject to change, but as of right now Intel plans to launch Tiger Lake on desktop. If Tiger Lake is expected to ship around ~Q4 2020, then Intel will make their final decision for full-scale production by ~March-April next year. Until Intel knows, nobody can know, only speculate.

 

[HD64G]

Intel is feared of angry shareholders and think they are sheeps to believe all their bs. 10nm is dead. Only mobile and some server models will be available. if they ever put any of those in desktop it will be in prebuilt systems for office purpose since they won't even dare sell any CPU that is slower than the previous gen since the clocks cannot get close to 4.5GHz without consuming more than their 5GHz made in 14++. Be sensible people. Intel's next chance to combat AMD will be whatever Keller is making for 2022 at the earliest and for their 7nm if it ever fruits. Or else they will make them in Samsung fabs as rumors circle around they are going to do for their GPUs.

 

[_Flare]

I don´t believe the whole production of all 10nm chips Intel will ever produce, will ever reach the numbers of the desktop broadwell i5-5675C and i7-5775C, and that was a strategic decision, but this time it´s entirely because of the f*cked up process.

 

[HenrySomeone]

Well, if this actually turns out to be true, AMD might just be able to catch up core-per-core or even take a slight lead; of course if their rumors about additional 10% IPC improvement and 8core ccx-es regarding 4000 series are true. If they won't manage now, then they never will...

 

[Vario]

Do people actually care about statements like these?
When I'm eyeing an upgrade, I just look at what's on the market. If there's something that's significantly faster than what I have, available for the right price, I buy. I don't care whether it says Intel or AMD on the box, I don't care about nm (but I do care about TDP), I don't care about marketing departments' prowess.

For a lot of people, speculating on semiconductor rumors is similar to watching a sports team or political party.

 

[bug]

Well, if this actually turns out to be true, AMD might just be able to catch up core-per-core or even take a slight lead; of course if their rumors about additional 10% IPC improvement and 8core ccx-es regarding 4000 series are true. If they won't manage now, then they never will...

People forget that even the mighty AthlonXP/64 wasn't better than P4 in all aspects. But that didn't prevent either of those CPUs from being the better choice overall.
Desktop Ryzen is in a pretty good spot right now. Better IPC would be welcome, but it's not where AMD should pour a lot of resources at the moment. Those are better spent on server offers and, more importantly, on mobile Ryzen.