Intel Scraps 10nm for Desktop, Brazen it Out with 14nm Skylake Till 2022?

[Aquinus]

Uhm, they've dropped $6-7 in value over the past month...
They're still down about $4.50 from their peak valuation this year.

I'm happy with my gain. I try not to worry about what could have been.

 

[dont whant to set it"']

"all hands on oem/Enterprise lineups" it would seem and yet, was I, the average Joe kept in the shady darkness of almost all node variants on the 14 nm node that those came to be in the diy/consumer market?

 

[silentbogo]

I don't see how AM4 is almost dead. What makes it almost dead?

What do you mean, almost dead? There's a new CPU coming out next year.

Ryzen 3000-series is probably going to be the last sAM4 chip. AMD's initial plans were to support the platform throughout 2020.
Not sure what's coming next, maybe they'll change their mind and keep it for another gen. It's hard to believe, but starting w/ Bristol Ridge we are already past a 3-year mark. If EOL will be extended by another 3 years, sAM4 will beat LGA775 in terms of longevity.

 

[Bones]

well, it’s highly unlikely zen 3 will release on am4, and even so, it’s not expected to be any significant uplift over zen2. So amd will at best only match intel in gaming.

With Intel deciding to stay with 14nm and AMD is already nipping their heels.... Right now?
Putting things off until 2022 to even release another desktop arch?
Yeah, right........

You don't innovate you don't compete, it's that simple and Intel just isn't getting it done. Intel right now is in a deep hole and it's getting really dim down there, they say 2022 and that's a bad sign. Things can change, no doubt but as things are now I don't see that changing anytime soon.
AMD has them by the proverbial nads and they know it.

 

[TheLostSwede]

Ryzen 3000-series is probably going to be the last sAM4 chip. AMD's initial plans were to support the platform throughout 2020.
Not sure what's coming next, maybe they'll change their mind and keep it for another gen. It's hard to believe, but starting w/ Bristol Ridge we are already past a 3-year mark. If EOL will be extended by another 3 years, sAM4 will beat LGA775 in terms of longevity.

Sorry, but where did you get this information? That goes against everything that's known about AM4 and Zen.

 

[Imsochobo]

Sorry, but where did you get this information? That goes against everything that's known about AM4 and Zen.

Amd have said several times Am4 through 2020

 

[silentbogo]

Sorry, but where did you get this information? That goes against everything that's known about AM4 and Zen.

AMD Community | AMD

community.amd.com

 

[TheLostSwede]

Amd have said several times Am4 through 2020

AMD Community | AMD

community.amd.com

Exactly, so Zen 3 in 2020 on AM4. Not on a new socket. I don't know how you two read.

AMD Ryzen 2018-2020 Roadmap leaked: Castle Peak, Matisse, Picasso, Vermeer and Renoir | VideoCardz.com

AMD Ryzen 2000 AMD Ryzen Threadripper (Zen2) codenamed Castle Peak AMD has recently held a special event for retailers and distributors. A new roadmap has been shown

videocardz.com

CPU-Roadmap: Ryzen 9000X3D vs. Core Ultra 200

Die Prozessor-Roadmap für 2024-2026: Welche AMD- und Intel-CPUs sind erschienen und werden voraussichtlich erscheinen? PCGH fasst für Sie zusammen!

www.pcgameshardware.de

 

[Dave65]

Haha, oh wow. And here I was hoping to hold out a little and get 10nm cpu from intel to upgrade for the new console gen from my 6700k. I don’t want to buy into am4 as it’s almost dead too, so it’s very curious to see how hardware releases will play out next year alongside new ps/xbox release.

AM4 is almost dead?
Someone better call AMD and tell them.

 

[TTPPUU]

waiting till 2022+ then

more money saved

 

[Raendor]

AM4 is almost dead?
Someone better call AMD and tell them.

they know it quite well, since they stated support ending in 2020 themselves from the beginning. We’re in October 2019, so 2020 says hello knocking on the door.

 

[silentbogo]

Exactly, so Zen 3 in 2020 on AM4. Not on a new socket. I don't know how you two read.

I read just fine. "Unofficial roadmaps", "leaks" and "rumors" are not my thing. Until AMD states that Zen3 is coming in 2020 and will be on AM4, all rumors and leaks are irrelevant. We are nearly at the due date and AMD still has lots of stuff to do for the current gen, like releasing 3950x, finishing work on TR-3000, getting some Zen2-powered APUs ready to go and sorting things out in mobile segment. Lots of work, not much time.
If supposed Zen 3 (Ryzen-4000 series I mean) actually manages to launch next year and actually is on AM4, it'll be simply a 7nm+ refresh of Zen2 with minimal changes and hopefully a slight clock boost.
I also do hope that AMD changes their mind and prolongs sAM4 EOL by another few years, but they did give us a heads-up many moons ago and they did officially repeat this EOL statement on many occasions, including this year.

 

[TheLostSwede]

I read just fine. "Unofficial roadmaps", "leaks" and "rumors" are not my thing. Until AMD states that Zen3 is coming in 2020 and will be on AM4, all rumors and leaks are irrelevant. We are nearly at the due date and AMD still has lots of stuff to do for the current gen, like releasing 3950x, finishing work on TR-3000, getting some Zen2-powered APUs ready to go and sorting things out in mobile segment. Lots of work, not much time.
If supposed Zen 3 (Ryzen-4000 series I mean) actually manages to launch next year and actually is on AM4, it'll be simply a 7nm+ refresh of Zen2 with minimal changes and hopefully a slight clock boost.
I also do hope that AMD changes their mind and prolongs sAM4 EOL by another few years, but they did give us a heads-up many moons ago and they did officially repeat this EOL statement on many occasions, including this year.

So you refute rumours and then you throw in a bunch of assumptions with no basis on any facts... Right, makes total sense...

 

[Vayra86]

Woopsie, told ya That didn't take long then.

Maybe they'll try to yank 14 gens out of 14nm now, just for the cohesion, I mean why not.

So you refute rumours and then you throw in a bunch of assumptions with no basis on any facts... Right, makes total sense...

Well, paper launch is a real thing, right - but its still a launch. And that is an important nuance I think @silentbogo is overlooking here.

Definitely something other than 'an announcement' or 'a press release' like we've seen daily from Intel

 

[chodaboy19]

They will keep patching their 14nm+++++ with PCIe 4.0 and other standards to keep it up to date. I think if all this is true then it makes sense from a business perspective to stop wasting money and effort to get 10nm on track and just skip ahead to 7nm. Painful in the short term but intel doesn't really have a choice. Intel can't sell what it doesn't have!

 

[Nihilus]

I doubt Zen 3 will be AM4. Zen 2+ on the other hand, most definitely.

 

[biffzinker]

Hopefully Intel works out the remaining kinks fabbing mobile chips on their 10nm process for the move to 7nm. Imagine this happening again with a repeat of delays, and cancellations of upcoming chips on the 7nm process.

 

[TheLostSwede]

I doubt Zen 3 will be AM4. Zen 2+ on the other hand, most definitely.

There is no Zen 2+, just like there is no spoon...

 

[silentbogo]

So you refute rumours and then you throw in a bunch of assumptions with no basis on any facts... Right, makes total sense...

I'm stating that platform ends in 2020 and there is no official info on Zen 3 to verify that it's going to be sAM4. I see no assumptions or rumors here.

 

[Vayra86]

Hopefully Intel works out the remaining kinks fabbing mobile chips on their 10nm process for the move to 7nm. Imagine this happening again with a repeat of delays, and cancellations of upcoming chips on the 7nm process.

Honestly I think part of the reason Intel's 10nm failed is exactly the time to market itself. Every year they had to postpone it was another year for 14nm to flourish and gain perf increases, reducing the gap to what 10nm would be balanced at. At the cost of power, but who cares about that for desktop or in the baby steps that figure takes over the years. I mean, it took Intel a decade to get their CPUs from 77W to 65W errrr 95W errr probably much more if you enter the BIOS...

And every year, the low-power optimized 10nm offering would turn more and more ridiculous. I mean who wants a low clocked low power chip that does worse but saves people 10 minutes of battery time? And also in marketing, how do you explain shedding as much as 25% of your clockspeeds to a simpleton? From AMD we accepted it because FX was so shit at 5Ghz and we had a decade to come to terms with it... though even for Zen, the main complaint was clockspeed limitations in the consumer segment!

 

[notb]

While on desktop their mindshare will still carry them for the next couple of years, on the server side of things this is starting to become something critical.

Worry less, read more.

The 2 important platforms: servers and mobile get 10nm.
Desktops stay on 14nm.
IMO this was quite obvious from the start. We knew the new node will be problematic and not all segments will get it.

It shouldn't be a huge problem. Realistically, Intel will be able to make competitive 14nm CPUs up to 8 cores, so the mainstream is covered (office/home desktops). The high-end stuff that can benefit from 7/10nm is so small Intel just doesn't bother.
It could have been different if they were able to ask the prices they used to. AMD aggressive pricing means the margins are down and high-end desktops aren't attractive anymore.

What this will mean is: AMD will rule in this segment and likely increase their prices.

 

[TheLostSwede]

I'm stating that platform ends in 2020 and there is no official info on Zen 3 to verify that it's going to be sAM4. I see no assumptions or rumors here.

So what's this then?

If supposed Zen 3 (Ryzen-4000 series I mean) actually manages to launch next year and actually is on AM4, it'll be simply a 7nm+ refresh of Zen2 with minimal changes and hopefully a slight clock boost.

You're filling in blanks with your own assumptions and guesses there. Is that any better than the rumours?
Zen 3 is expected to offer the same boost in performance that Zen 2 did over Zen/Zen+, but you're free to believe otherwise.
So far, AMD has delivered and there's no reason to doubt that they'll not continue to do so for at least a little while longer.

 

[Vayra86]

Worry less, read more.

The 2 important platforms: servers and mobile get 10nm.
Desktops stay on 14nm.
IMO this was quite obvious from the start. We knew the new node will be problematic and not all segments will get it.

It shouldn't be a huge problem. Realistically, Intel will be able to make competitive 14nm CPUs up to 8 cores, so the mainstream is covered (office/home desktops). The high-end stuff that can benefit from 7/10nm is so small Intel just doesn't bother.
It could have been different if they were able to ask the prices they used to. AMD aggressive pricing means the margins are down and high-end desktops aren't attractive anymore.

What this will mean is: AMD will rule in this segment and likely increase their prices.

Mobile? Where's that product stack then? What they've got is a POC, at best.

 

[Renald]

I'm laughing so hard right now

I remember the Roadmap from 2 years ago : "it's ok, we're finishing the process in 10nm, and going further".
Then they got short on stock due their inability to provide laptops to businesses.
And know they totally flushed Desktop.

So, AMD is already crushing on Desktop market, and they will start to attack on Mobile since they have no real opposition in Desktop (Zen 3 is on the rails).

 

[notb]

You're filling in blanks with your own assumptions and guesses there. Is that any better than the rumours?
Zen 3 is expected to offer the same boost in performance that Zen 2 did over Zen/Zen+, but you're free to believe otherwise.

I don't see how his pessimistic guess is any worse than your optimistic one.

At this point there aren't even sensible leaks about Zen 3. It's all just hopes and sarcasm - depending on which side of bed you prefer.

So far, AMD has delivered and there's no reason to doubt that they'll not continue to do so for at least a little while longer.

In the 10% market share their able to cover with TSMC supply.

Mobile? Where's that product stack then? What they've got is a POC, at best.

Mobile segment means laptops.