Intel 7nm CPUs Delayed by a Year, Alder Lake in 2H-2021, Other Commentary from Intel Management

[btarunr]

Intel's silicon fabrication woes refuse to torment the company's product roadmaps, with the company disclosing in its Q2-2020 financial results release that the company's first CPUs built on the 7 nanometer silicon fabrication node are delayed by a year due to a further 6-month delay from prior expectations. The company will focus on getting its 10 nm node up to scale in the meantime.

The company mentioned that the 10 nm "Tiger Lake" mobile processor and "Ice Lake-SP" enterprise processor remains on-track for 2020. The company's 12th Generation Core "Alder Lake-S" desktop processors won't arrive before the second half of 2021. In the meantime, Intel will launch its 11th Gen Core "Rocket Lake" processor on the 14 nm node, but with increased IPC from the new "Cypress Cove" CPU cores. Also in 2H-2021, the company will launch its "Sapphire Rapids" enterprise processors that come with next-gen connectivity and updated CPU cores.



It's interesting to note that Intel was specific about "CPU" when talking about 7 nm, meaning that Intel's foundry woes only affect its CPU product stack, and not a word was mentioned in the release about the company's discrete GPU and scalar compute processors that are being prototyped and validated. This is probably the biggest hint we'll ever get from Intel that the company's dGPUs are being designed for third-party foundries (such as Samsung or TSMC), and that the Xe dGPU product roadmap is disconnected from that of Intel's fabs.

Intel is accelerating its transition to 10 nm products this year with increasing volumes and strong demand for an expanding line up. This includes a growing portfolio of 10 nm-based Intel Core processors with "Tiger Lake" launching soon, and the first 10 nm-based server CPU "Ice Lake," which remains planned for the end of this year. In the second half of 2021, Intel expects to deliver a new line of client CPU's (code-named "Alder Lake"), which will include its first 10 nm-based desktop CPU, and a new 10 nm-based server CPU (code-named "Sapphire Rapids"). The company's 7 nm-based CPU product timing is shifting approximately six months relative to prior expectations. The primary driver is the yield of Intel's 7 nm process, which based on recent data, is now trending approximately twelve months behind the company's internal target.

Intel's post results call also revealed a handful interesting tentative dates. For starters, "Tiger Lake" is shipping in "a matter of weeks," indicating an imminent launch ahead of the "Back to School" shopping season. Next up, the company's high-performance scalar compute processor, codenamed "Ponte Vecchio" remains slated for 2021-22, and given that it's reportedly being designed for 7 nm, we have our next big hint confirmation that these dGPUs will be built on third-party 7 nm fabs. Intel did mention that the Foveros packaging technology could be further developed over the years, and its upcoming discrete GPUs could combine dies (tiles) from multiple sources, which could include its own fabs.

Given the delays in Intel's 7 nm foundry node, the first Intel client-segment processors based on the node won't arrive before late-2022 or 2023, which means refinements of the current 10 nm silicon fabrication node should support Intel's client-segment product stack for the foreseeable future. The first enterprise 7 nm processors will arrive by the first half of 2023. Intel also mentioned that they expect to see "one full node improvement" from a refined 10 nanometer process, which isn't surprising, given how much experience they have improving their 14 nanometer process.

View at TechPowerUp Main Site

 

[mtcn77]

Intel still owns the BEST chip to chip communication a.k.a EMIB. They could design the worst chips and still win due to communication. It is just like Nvidia mobile ip ported gpus. They have more communication, so all is dandy.

It is just the same with Samsung entry into 3d nand. They took a 40nm process and made history. Everybody knows about 860, nobody knows what came before it eventhough it came in a twice subsequent node.

 

[Crackong]

For 7nm GPU , TomsHardware says in their version of article :

The company will also use external third-party foundries for its forthcoming 7nm Ponte Vecchio GPUs, the company's first graphics chips. Swan noted the GPUs will come in late 2021 or early 2022, portending a delay beyond the original schedule for a 2021 launch in the exascale Aurora supercomputer.

 

[mtcn77]

Could Intel compete on a packaging scenario, TSMC CoWoS vs Intel EMIB, that is what I'm guessing at.

 

[Verpal]

Let me guess......

10nm +++ that boost up to 5.5Ghz with big.LITTLE design that chew tons of power but miraculously competitive against 5nm EUV.

Maybe Intel should just consider shipping bare die locked CPU with cooler soldered on the top, at least user won't complain about temperature.

Lets hope 10nm +++ won't happen, Intel can't be this stubburn, they seems to realize AMD's threat already.

 

[X71200]

Intel still owns the BEST chip to chip communication a.k.a EMIB. They could design the worst chips and still win due to communication. It is just like Nvidia mobile ip ported gpus. They have more communication, so all is dandy.

It is just the same with Samsung entry into 3d nand. They took a 40nm process and made history. Everybody knows about 860, nobody knows what came before it eventhough it came in a twice subsequent node.

Even the chips with the beefiest of that connection, the LGA 3647 Xeon Golds with 3 links and octa channel RAM can be had for cheap from eBay because Intel no longer really rules like they did back in the days of Nehalem. Higher destinies inherently carry more leakage as usual, look at their thermal ratings, almost nobody wants Intel anymore. As for Samsung making it big with the 860 SSD and 3D-nand, you're wrong there as well. The 840 Pro and others were big because of how good they were back when Sandforce was still selling around, and almost anybody in a tech forum was well aware of it.

 

[chodaboy19]

If intel can at least get 10nm shipping across all market segments they could have a viable chance to survive. Not sure why they are warning everyone about 7nm when 10nm is not even out in volume. The best guess is that 7nm is in deep trouble and intel is managing expectations from their clients to their investors.

 

[watzupken]

I am not convinced that Alder Lake will be able to save Intel in the consumer space at this point unless proven otherwise when I see the actual results. In particular when we look at desktop segment, the big/little core config will not make much sense. Intel further muddied the water of Alder Lake by creating SKUs with and without big/little config. Looking at the TDP for Alder Lake S from the rumors of up to 125W, I feel its going to be 14nm++++ strategy once again, i.e. sacrificing efficiency for performance. The small cores probably will hide their poor power efficiency under light load.

As for Intel outsourcing their GPU to external foundry, it is not unexpected. Considering the maturity of the 14nm fab and yet they are not able to keep up with demand, it is likely worst for 10nm. In addition, I also feel that Intel may not have the expertise to fab big/ complex GPUs, unlike the likes of TSMC, Samsung and GF.

If intel can at least get 10nm shipping across all market segments they could have a viable chance to survive. Not sure why they are warning everyone about 7nm when 10nm is not even out in volume. The best guess is that 7nm is in deep trouble and intel is managing expectations from their clients to their investors.

They are obligated to because this is a significant event. Intel surely have shared the 7nm roadmap with their investors previously. With the delivery of 7nm products off track, they will need to correct their roadmap and make it transparent to investors. And as you can tell, the moment the news of another 6 months delay in 7nm delivery got announced, the stock got hammered.

 

[mtcn77]

As for Samsung making it big with the 860 SSD and 3D-nand, you're wrong there as well. The 840 Pro and others were big because of how good they were back when Sandforce was still selling around, and almost anybody in a tech forum was well aware of it.

3D-nand changed the basics. No enterprise drive has a better mixed 4K benchmark. To make matters clear, 850 Evo 250 is the 1st in 2.5" drives while 840 Pro 256 is 29th on ssd.userbenchmark. Funny you would mention that.

 

[X71200]

3D-nand changed the basics. No enterprise drive has a better mixed 4K benchmark. To make matters clear, 850 Evo is 1st in 2.5" drives while 840 Pro is 29th on ssd.userbenchmark. Funny you would mention that.

Rofl, are you seriously looking Userbenchmark ratings of anything? The people who think DX11 and 12 are irrelevant in 2020. I'm talking about the relevancy of the 840 Pro when it came out. This has nothing to do with performance, you talked about a drive making its name. The people in tech forums like TPU, [H], etc, were all big on it because it was also better than SandForce. 3D-Nand is not everything, you could put it on a crap drive today and be under the ground with 4k.

 

[mtcn77]

I'm talking about the relevancy of the 840 Pro when it came out. This has nothing to do with performance, you talked about a drive making its name. The people in tech forums like TPU, [H], etc, were all big on it because it was also better than SandForce. 3D-Nand is not everything, you could put it on a crap drive today and be under the ground with 4k.

Okay. That was how enterprise drives were counted on back then, instead of burst r/w. Funny you would mention that.

 

[windwhirl]

Intel's silicon fabrication woes refuse to torment the company's product roadmaps,

More like they refuse to stop tormenting the company

 

[X71200]

Okay. That was how enterprise drives were counted on back then, instead of burst r/w. Funny you would mention that.

The Evo drives never made it to the "enterprise" back then, heck even those people in the forums dashed them because they weren't relying on long-term TLC sustainability yet. Today you have TLC enterprise drives. Enterprise was still using stuff like SLC Sandforce 2285's, I still have some of those drives in my old laptops and whatnot. Seriously, you're making fun of yourself while trying to make fun of others.

 

[mtcn77]

The Evo drives never made it to the "enterprise" back then,

Why should it, I just stated they changed the hierarchy. You had pro series doing mixed r/w stuff and amateurs that compared their naked read performance on a 0% worn pristine drive. When 850 came, suddenly you had a drive do both simultaneously.

 

[AnarchoPrimitiv]

AMD should be on 5nm by the time Intel is on 7nm, right?

 

[chodaboy19]

AMD should be on 5nm by the time Intel is on 7nm, right?

AMD and intel's fabrication processes are not a 1:1 match.
Intel's 10nm is closer to TSMC's 7nm and Intel's 7nm is closer to TSMC's 5nm.

 

[seronx]

AMD should be on 5nm by the time Intel is on 7nm, right?

AMD should be on something that has the density of Intel's 7nm sooner than Intel. Genoa, Raphael, Rembrandt should be on a TSMC N5 process which is comparable to Intel's P1276/7nm.

 

[Jism]

Intel Bulldozer FX.

 

[DemonicRyzen666]

AMD and intel's fabrication processes are not a 1:1 match.
Intel's 10nm is closer to TSMC's 7nm and Intel's 7nm is closer to TSMC's 5nm.

Where does every one get this?
Far as I know this just what intel claims and there is no actual proof

There density comes from having several different designs for caches L1 caches transistors are different from L2 and so is it's L3 it's all for space saving.

AMD an other manufacturers use a more uniformed transistor on the nod.

That is why AMD an other always end up 20% larger dies

 

[Caqde]

So based on what we are seeing here. Alder-Lake S will go against the 5nm Zen 4 Desktop CPU's (Late 2021/Early 2022)..... And the first 7nm Intel CPU's will be up against a future 5nm+ "Zen 5" chip in Late 2022/Early 2023... Yup doing good Intel. A bit later and you can play against a potential 3nm "Zen 6" chip. Keep the delays coming....

 

[londiste]

Where does every one get this?
Far as I know this just what intel claims and there is no actual proof

At least some design specifications for manufacturing processes are known.

7 nm process - Wikipedia

en.wikipedia.org

It is not just Intel claiming this, industry analysts have a pretty good idea what is going on. If there are products out on a manufacturing process the density and specs can be verified to some degree.

 

[jeremyshaw]

The problem with the 7nm delay is Aurora, IMO. Intel already got a reprieve once, if they screw it up again, do they finally lose the contract? Will AMD snatch all three Exaflop contracts in the US?

 

[thesmokingman]

Well that was a shocker!

Like really, who believed they were actually gonna release 7nm this year?

AMD and intel's fabrication processes are not a 1:1 match.
Intel's 10nm is closer to TSMC's 7nm and Intel's 7nm is closer to TSMC's 5nm.

Do you ever get tired of spouting this rubbish?

That's like Nikola saying they make better electric semis than Tesla. lmao

 

[cucker tarlson]

It is just the same with Samsung entry into 3d nand. They took a 40nm process and made history. Everybody knows about 860, nobody knows what came before it eventhough it came in a twice subsequent node.

sadly,I do.

 

[Deleted member 185088]

How the hell did they mess up so badly, like I said on Anandtech, since Sandy Bridge in 2011 they didn't do anything major, they just got a lot of money doing the same thing over and over.