" Intel is finally shipping its 10 nm processors in high volume"
Define high volume. Seems 10nm has started with a wimper at best. And even Intel says it will not be till later 2020 till more major 10nm cpus come out. Thats assuming all goes to plan and no issues come up.
No word yet on desktop full power -S or even mid power -H parts fabbed at 10nm. Which almost certainly means there won't be, and Intel will release Comet Lake-S/H instead, at 14nm++++. The question is if they'll repeat the same late next year when they release Tiger Lake. Word on the street is that the successor of Comet Lake-S/H is Rocket Lake-S/H, also fabbed at 14nm+++++ but apparently with a new μarch (either Sunny or Willow Cover cores and either Gen11 or Gen12 iGPUs).
Those 15 watt CPUs have a use. Not everyone wants (or needs) a super thick battery draining high end laptop. That being said, I'll stick with my Threadripper on the desktop and skylake on my laptop for the time being.
Exactly - as it is with 10nm plus the re-re-re-re-retooled 14nm they still cant keep up with production. That is mostly due to AMD's competition and Intel having to use more cores per CPU than expected... Anyhow "volume" is being used very loosely here.
A little survey. Who would be interested in Intel's 10nm DG1 GPUs, given that Navi is 7nm, by the next year we will have a next gen, and Intel is going to release an all-new architecture with all-new drivers and all-new bugs? What if it's not even PCIe4?
I suspect that Intel's GPUs are going to outperform current Navi. Big Navi may be a different story though. If you are a good consumer, you won't be loyal to any brand, and instead buy based on price/performance/overall value. If Intel launches a competitive GPU to current NVIDIA/AMD offerings, then I expect many consumers will consider them. If price is the least bit out of whack, or performance isn't up to par compared to the competition, they are going to have a hard time. a 3 horse race is a hell of a lot harder to deal with to begin with, and when you are years behind the competition it becomes more of an issue.
Do you really believe they are going to be a good value with brand new(=buggy) drivers and worse technology? Well, probably Intel will give them out below cost to volume manufacturers...
@peevee no consumers at all, server/heavy workstation market products. Intel plays it smart, proof that you can make a products which beats the crap out of the other competition and your path becomes open for the consumer market. However i am not sure if intel is looking to produce any consumer market gpu at all. But honestly i am pretty sure if they want to they certainly have the money for it. People tend to forget that intel is a giant, and has the financial power to do anything they set their minds at. I keep saying it when you wake a slumbering gaint you will feel the painfull result ;)
intel is not aiming at any consumer based gpu product, they clearly only aim at the computation market which is a thousands of dollars per gpu costing industry Which uptill now is dominated by nvidia. AMD has a small percentage in the high end super server segment currently. But that is not the big profit market and often those super computers get nvidia gpu's as well. AMD has no answer yet for the top end enterprise gpu market, but i am sure intel is going to succeed to compete with nvidia sooner than later. When the giant wakes up, its showing its muscle soon. So question remains do they only go for top end user products or will they enter the smaller profit market (consumer products) as well. We have to wait and see, my two cents is that it will take a while longer before intel thinks about making consumer based gpu. If had owned intel i would try to enter the top segment as well, and i might be wrong but profit wise consumer products is not a good starting market.
Here we go again. Intel's HP 10 nm is 50% denser than TSMC's HP 7FF.
Gaming GPUs do not need PCIe 4.0 and even PCIe 3.0. 2080 Ti is fine with PCIe 2 x16.
Intel has a time to debug Xe while factories are being built. They are not rushing like AMD. But still any new architecture is quite buggy at the beginning. So was with Turing, so as RDNA right now.
Any yet that "wimper" probably covers a quarter of all AMD's CPUs. If Intel has 4 times the volume of AMD they only need to have 1/16th of their CPUs on 10 nm to cover a quarter of AMD's volume.
Did Anandtech even had a Icelake product sample to review? I haven't been following Intel closely ( Not interested in their lies ) but as far as I can tell there isn't a wide availability ( or even availability ) of Intel 10nm products on the market.
They said they were aiming for product to customer's hand for Xmas. And we will see about that. Judging from the way things are going, They are basically stocking up as much 10nm as they can and try to make 10nm looking like yielding normally.
How did the saying go - Fool me once, shame on me. Fool me twice..
How many times have we seen Intel making the same claims and then not follow up with proper execution? Now we should believe that they are shipping in high volumes, oh give me a break. Isn't it funny that they just released 10th gen 14nm CPUs to be their "bread and butter" also in mobile space during the next year.
Those are usually lower core counts and lower speeds... 10nm Ice Lake Xeons are supposed to support PCIe 4.0 which you would want in server storage and networking. But if it won't show up until around the same time as Milan, will it have much impact?
"The company no longer sets ultra-ambitious goals for scaling each node, but attempts to find a right balance between performance, power, cost, and timing. "
In other words Intel is copying TSMC's approach with 7nm. TSMC took a very low risk approach by only having some upfront process at 7nm and they didn't go to crazy on densities either. Sure the first 7nm process isn't spectacular but they have desktop chips on said process where Intel does not. So I would say very wise move to emulate TSMC's evolutionary instead of revolutionary approach as these nodes get more and more complex.
More importantly, TSMC was able to ship and deliver what they promised for 7 nm roughly on time. They weren't as ambitious but that conservative design helped TSMC leap frog Intel's 10 nm.
The real risk for TSMC's 7 nm node is the EUV variant. This actually simplifies some things (going to single patterning in some cases) but having the EUV tools ready for volume production was the greater variable. EUV has been a technology the entire fab business has been striving for the past decade. Kind of impressive that TSMC got to 7 nm without it.
Pretty much my sentiment as well. I'm excited about TSMC's move to 7nm+ using EUV more than the first generation 7nm as this is where we really start to see 7nm shine. More so I expect to see decently higher clocks for AMD. We all see only one core able to hit those rated boost clocks for any given Zen2 CPU and its very obvious its TSMC's process holding that back. The quad patterning is obviously not extremely consistent albeit impressive they can do it at all.
The use of EUV does affect the performance of the chip that is produced... The world is complicated. Transistors are physical things with defects and characteristics that can depend on the method used to make the transistors. Per the CEO of ASML (the maker of the EUV lithography tools) one reason DRAM manufacturers are looking at EUV for their future DRAM process nodes is the performance of the chips made using EUV compared with those using multipatterned DUV.
NO. Intel want to look like it's copying TSMC. That's different from actually copying them...
The superficial difference are clear - the bragging (OMG, they just cannot control themselves, can they?) about 2x density - the pre-announcement years in advancement rather than not announcing till details are set.
But those are superficial. More important is how process evolves. - Intel's +, ++ etc advances seem like extreme optimizations of an existing process. Meaning that they get better yield and better transistors for that process BUT it's very process specific. There's no evidence that what you learn in moving from 14nm+ to 14nm++ helps you much in going from 10nm to 10nm+. - TSMC's advances are mostly more generic. This year lithography gets better for metal layers. Next year the basic transistor cell gets smaller (SDB, COAG, that sort of thing). The year after that GAA gets introduced. These are all improvements that can be built upon, unlike Intel's improvements.
This COULD change with 10nm (it's not clear even Intel knows what it's doing, so certainly outsiders don't!) But it's not what Intel has been doing so far.
This is like listening to a substance abuser: "I am on a different path and am changing my ways. The past is behind me now. I am looking forward to better things in the future." Its always just worthless talk.
I wouldn't be shocked if Intel re-brands their 5nm process as "4nm" or "3nm" since there's even less correlation between marketing names for processes and any physical dimensions than there was ten years ago.
Not really, actually it has been narrowed over the last couple of years. Intel's 10nm is about 8.5nm on competing processes. Intel's 7nm is closer to TSMC's 6nm process. This is based on the transistor density per square inch.
What 10nm products has Intel shipped for two years? Customer only stuff? I don't know of anything from 2017 that was 10nm... Someone want to report the CEO for lying to the SEC?
The only "two years" I see in the article is the timeline for 7nm. Two years from now in 2021. The slide for process techs even dropped the polite fiction that they shipped 10nm at the very end of 2018 (a crappy mobile cpu without graphics in a few China only laptops).
Splitting hairs, but Intel also released a couple Canon Lake NUCs: NUC8i3CYSN NUC8i3CYSM
I have no idea what kinds of numbers those shipped in -- but I'd bet that total volume was pretty low.
As I commented in the Anandtech Canon Lake review(where Ian reviewed the Chinese laptop back in January 2019) the NUC was expensive and was limited to 8GB of soldered down RAM.
I imagine this is an unpopular opinion, but, as delayed and challenged as Intel has clearly been, do people really expect them to be stuck on 14nm forever? They have to, at some point, break the 10nm barrier, and now is a reasonable time for them to do it. You can buy Icelake laptops right now from both Microsoft and Dell.
Be skeptical of 7nm, sure, but Intel has objectively put Icelake into retail, and the CEO would be in hot water if he lied about the systems being introduced on an investor call. 10nm is here, and just because it isn't shipping in 95W parts doesn't mean it isn't shipping in high volume (how many desktop parts sell compared to Ultrabooks, anyways?)
>Intel's already predicting 10, 10+, 10++. Given Intel's predictions and what ultimately turned into reality, I'm not going to be surprised if this turns into a 10++++++++ process 6 years or so later, with volume 7nm production nowhere in sight.
"The company no longer sets ultra-ambitious goals for scaling each node"
Which is why the claim is 2x density increase for 7nm? And why they felt a need to talk about 5nm?
Seriously, they have learned NOTHING from the past. In particular they certainly haven't learned to STFU until you have EVERYTHING validated.
Compare how little TSMC says until they are ready to roll. Even Samsung (way too chatty compared to TSMC) keeps their mouth shut a lot better, rather than promising how things will play out five years from now.
Well, objectively 7 nm **is** a 50% higher area density than 10 nm (7*7=49; 10*10=100; 100/49~=2).
That of course assumes that Intel's "nm" numbers actually have a real physical meaning, rather than amounting to nothing more than marketing fiction...
As for the talk of 5 nm, I'm pretty sure the likes of TSMC have already projected out as far as "3 nm" (whatever that actually means in reality) - and if the company is already spending significant R&D resources on a specific schedule for 5 nm, then the CEO would be remiss not to apprise the shareholders accordingly...
TSMC just announced that they are building a whole new fab for its 3nm. The new fab should be operational in 2022. So much for keeping their mouth shut.
One this is for certain ... there is never going to be a 0 nm process. What we are witnessing, is nothing short of the end ... of the Golden Age of Semi-Conductors.
"yields are improving ahead of expectations for both for client and datacenter CPUs"
So you've produced 16-32 core cpus on 10nm in HV? I only see quads and small chips.
Until you are selling at least a 16 core 100w 10nm you haven't started. 7nm after only 2yrs? LOL Only because 10nm was 4yrs late (still not out IMHO).
Not sure how BIG this 7nm gpu is either, as Intel said the first discrete card is SMALL CU count and LOW END. I'm confused by their statements here, as I though wimp desktop gpu was coming first (what a waste, like AMD 7nm NON king first...ROFL), and hope reality is BIG first as it should ALWAYS be if you want to win. Kings first, crap second.
I'm pretty intrigued to see how 10nm impacts Intel's CPU lineup. Even at 14nm, their mobile chips have been untouchable from a power consumption perspective compared to AMD's offerings, so I'm wondering just how this shift will impact core counts and boost rates in that 15W (or lower) envelope.
On the desktop side, here's hoping they can drop their total power draw down a bit while increasing clock speeds, although their current power draw isn't something that I'm particularly impacted by.
Regardless, as others have said, I'll believe it when I see it at this point. Current 7nm processes aren't particularly impressive, so I'm holding my breath that Intel's 10nm can match or beat what's currently on the market.
Everyone keeps talking about clock speeds, and that is exactly the reason why we'll continue to see 14nm cpus. 10nm has lower clock speeds and I don't believe Intel will ever reach 14nm clock speeds ever again. EUV (7nm) might be able to reach higher clock speeds than 10nm, but I doubt they will reach 5GHz beyond short bursts.
Is there a reason nobody is making Pentium M Banias-style desktop setups with whatever the "shipping" 10nm chips are? Or no manufacturer is ignoring Intel spec with a higher TDP configuration laptop with the 10nm chips?
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
59 Comments
Back to Article
Marlin1975 - Monday, October 28, 2019 - link
" Intel is finally shipping its 10 nm processors in high volume"Define high volume. Seems 10nm has started with a wimper at best. And even Intel says it will not be till later 2020 till more major 10nm cpus come out. Thats assuming all goes to plan and no issues come up.
shabby - Monday, October 28, 2019 - link
And define which processors, those party 15w ones? PfftSantoval - Monday, October 28, 2019 - link
No word yet on desktop full power -S or even mid power -H parts fabbed at 10nm. Which almost certainly means there won't be, and Intel will release Comet Lake-S/H instead, at 14nm++++. The question is if they'll repeat the same late next year when they release Tiger Lake. Word on the street is that the successor of Comet Lake-S/H is Rocket Lake-S/H, also fabbed at 14nm+++++ but apparently with a new μarch (either Sunny or Willow Cover cores and either Gen11 or Gen12 iGPUs).eek2121 - Monday, October 28, 2019 - link
Those 15 watt CPUs have a use. Not everyone wants (or needs) a super thick battery draining high end laptop. That being said, I'll stick with my Threadripper on the desktop and skylake on my laptop for the time being.drothgery - Monday, October 28, 2019 - link
You mean the highest-volume CPU category for Intel?evilpaul666 - Tuesday, October 29, 2019 - link
I don't follow them super closely, but the new 15W parts didn't seem particularly impressive compared to the previous Coffee Lake ones.goatfajitas - Monday, October 28, 2019 - link
"seems 10nm has started with a wimper at best"Exactly - as it is with 10nm plus the re-re-re-re-retooled 14nm they still cant keep up with production. That is mostly due to AMD's competition and Intel having to use more cores per CPU than expected... Anyhow "volume" is being used very loosely here.
peevee - Monday, October 28, 2019 - link
A little survey. Who would be interested in Intel's 10nm DG1 GPUs, given that Navi is 7nm, by the next year we will have a next gen, and Intel is going to release an all-new architecture with all-new drivers and all-new bugs? What if it's not even PCIe4?eek2121 - Monday, October 28, 2019 - link
I suspect that Intel's GPUs are going to outperform current Navi. Big Navi may be a different story though. If you are a good consumer, you won't be loyal to any brand, and instead buy based on price/performance/overall value. If Intel launches a competitive GPU to current NVIDIA/AMD offerings, then I expect many consumers will consider them. If price is the least bit out of whack, or performance isn't up to par compared to the competition, they are going to have a hard time. a 3 horse race is a hell of a lot harder to deal with to begin with, and when you are years behind the competition it becomes more of an issue.peevee - Monday, October 28, 2019 - link
Do you really believe they are going to be a good value with brand new(=buggy) drivers and worse technology? Well, probably Intel will give them out below cost to volume manufacturers...bronan - Sunday, November 3, 2019 - link
@peevee no consumers at all, server/heavy workstation market products.Intel plays it smart, proof that you can make a products which beats the crap out of the other competition and your path becomes open for the consumer market.
However i am not sure if intel is looking to produce any consumer market gpu at all.
But honestly i am pretty sure if they want to they certainly have the money for it.
People tend to forget that intel is a giant, and has the financial power to do anything they set their minds at. I keep saying it when you wake a slumbering gaint you will feel the painfull result ;)
bronan - Sunday, November 3, 2019 - link
intel is not aiming at any consumer based gpu product, they clearly only aim at the computation market which is a thousands of dollars per gpu costing industryWhich uptill now is dominated by nvidia.
AMD has a small percentage in the high end super server segment currently. But that is not the big profit market and often those super computers get nvidia gpu's as well.
AMD has no answer yet for the top end enterprise gpu market, but i am sure intel is going to succeed to compete with nvidia sooner than later. When the giant wakes up, its showing its muscle soon. So question remains do they only go for top end user products or will they enter the smaller profit market (consumer products) as well.
We have to wait and see, my two cents is that it will take a while longer before intel thinks about making consumer based gpu. If had owned intel i would try to enter the top segment as well, and i might be wrong but profit wise consumer products is not a good starting market.
regsEx - Monday, October 28, 2019 - link
Here we go again. Intel's HP 10 nm is 50% denser than TSMC's HP 7FF.Gaming GPUs do not need PCIe 4.0 and even PCIe 3.0. 2080 Ti is fine with PCIe 2 x16.
Intel has a time to debug Xe while factories are being built. They are not rushing like AMD. But still any new architecture is quite buggy at the beginning. So was with Turing, so as RDNA right now.
tawm - Tuesday, October 29, 2019 - link
>Intel's HP 10 nm is 50% denser than TSMC's HP 7FF.Wrong.
Yojimbo - Monday, October 28, 2019 - link
Any yet that "wimper" probably covers a quarter of all AMD's CPUs. If Intel has 4 times the volume of AMD they only need to have 1/16th of their CPUs on 10 nm to cover a quarter of AMD's volume.ksec - Tuesday, October 29, 2019 - link
Did Anandtech even had a Icelake product sample to review? I haven't been following Intel closely ( Not interested in their lies ) but as far as I can tell there isn't a wide availability ( or even availability ) of Intel 10nm products on the market.They said they were aiming for product to customer's hand for Xmas. And we will see about that. Judging from the way things are going, They are basically stocking up as much 10nm as they can and try to make 10nm looking like yielding normally.
Uroshima - Monday, October 28, 2019 - link
"Xeon CPUs for server storage and network" ... just me or this does not sound as a serious data crunching type of Xeon?Jorgp2 - Monday, October 28, 2019 - link
That sounds exactly like the role of Atom server CPUs.They might push a 10nm Tremont based replacement for denverton
visualzero - Monday, October 28, 2019 - link
How did the saying go - Fool me once, shame on me. Fool me twice..How many times have we seen Intel making the same claims and then not follow up with proper execution? Now we should believe that they are shipping in high volumes, oh give me a break. Isn't it funny that they just released 10th gen 14nm CPUs to be their "bread and butter" also in mobile space during the next year.
visualzero - Monday, October 28, 2019 - link
Sorry, meant to post this to the top level.Santoval - Monday, October 28, 2019 - link
"Fool me once, shame on me. Fool me twice..."It's actually the other way around ("Fool me once, shame on you. Fool me twice shame on me").
GreenReaper - Monday, October 28, 2019 - link
It shouldn't be acceptable for Intel to be fooling paying customers even once!twotwotwo - Monday, October 28, 2019 - link
(It's a call transcript where it's hard to get every comma right; maybe "server, storage, and network" was the intent?)Freeb!rd - Monday, October 28, 2019 - link
Those are usually lower core counts and lower speeds... 10nm Ice Lake Xeons are supposed to support PCIe 4.0 which you would want in server storage and networking. But if it won't show up until around the same time as Milan, will it have much impact?Scabies - Monday, October 28, 2019 - link
>We are on track to launch our first 7 nm based product, a datacenter-focused discrete GPU, in 2021Unless the thing gets Raja'd
Vitor - Monday, October 28, 2019 - link
LOL at 10++ being already mentioned.FreckledTrout - Monday, October 28, 2019 - link
"The company no longer sets ultra-ambitious goals for scaling each node, but attempts to find a right balance between performance, power, cost, and timing. "In other words Intel is copying TSMC's approach with 7nm. TSMC took a very low risk approach by only having some upfront process at 7nm and they didn't go to crazy on densities either. Sure the first 7nm process isn't spectacular but they have desktop chips on said process where Intel does not. So I would say very wise move to emulate TSMC's evolutionary instead of revolutionary approach as these nodes get more and more complex.
Kevin G - Monday, October 28, 2019 - link
More importantly, TSMC was able to ship and deliver what they promised for 7 nm roughly on time. They weren't as ambitious but that conservative design helped TSMC leap frog Intel's 10 nm.The real risk for TSMC's 7 nm node is the EUV variant. This actually simplifies some things (going to single patterning in some cases) but having the EUV tools ready for volume production was the greater variable. EUV has been a technology the entire fab business has been striving for the past decade. Kind of impressive that TSMC got to 7 nm without it.
FreckledTrout - Monday, October 28, 2019 - link
Pretty much my sentiment as well. I'm excited about TSMC's move to 7nm+ using EUV more than the first generation 7nm as this is where we really start to see 7nm shine. More so I expect to see decently higher clocks for AMD. We all see only one core able to hit those rated boost clocks for any given Zen2 CPU and its very obvious its TSMC's process holding that back. The quad patterning is obviously not extremely consistent albeit impressive they can do it at all.peevee - Monday, October 28, 2019 - link
" EUV more than the first generation 7nm as this is where we really start to see 7nm shine. More so I expect to see decently higher clocks for AMD."Why? It is a fabrication technology, how is it going to affect things like interference, leakage and resistance?
Yojimbo - Monday, October 28, 2019 - link
The use of EUV does affect the performance of the chip that is produced... The world is complicated. Transistors are physical things with defects and characteristics that can depend on the method used to make the transistors. Per the CEO of ASML (the maker of the EUV lithography tools) one reason DRAM manufacturers are looking at EUV for their future DRAM process nodes is the performance of the chips made using EUV compared with those using multipatterned DUV.name99 - Monday, October 28, 2019 - link
NO. Intel want to look like it's copying TSMC. That's different from actually copying them...The superficial difference are clear
- the bragging (OMG, they just cannot control themselves, can they?) about 2x density
- the pre-announcement years in advancement rather than not announcing till details are set.
But those are superficial. More important is how process evolves.
- Intel's +, ++ etc advances seem like extreme optimizations of an existing process. Meaning that they get better yield and better transistors for that process BUT it's very process specific. There's no evidence that what you learn in moving from 14nm+ to 14nm++ helps you much in going from 10nm to 10nm+.
- TSMC's advances are mostly more generic. This year lithography gets better for metal layers. Next year the basic transistor cell gets smaller (SDB, COAG, that sort of thing). The year after that GAA gets introduced. These are all improvements that can be built upon, unlike Intel's improvements.
This COULD change with 10nm (it's not clear even Intel knows what it's doing, so certainly outsiders don't!) But it's not what Intel has been doing so far.
outsideloop - Monday, October 28, 2019 - link
This is like listening to a substance abuser: "I am on a different path and am changing my ways. The past is behind me now. I am looking forward to better things in the future." Its always just worthless talk.5080 - Monday, October 28, 2019 - link
Crazy, by the time they have 7nm TSMC is already on 3nm+drothgery - Monday, October 28, 2019 - link
I wouldn't be shocked if Intel re-brands their 5nm process as "4nm" or "3nm" since there's even less correlation between marketing names for processes and any physical dimensions than there was ten years ago.haukionkannel - Tuesday, October 29, 2019 - link
Intel 7nm maybe Denver than TSMC 3nm... that marketing nm has gone out of any meaning Many years ago...5080 - Tuesday, October 29, 2019 - link
Not really, actually it has been narrowed over the last couple of years. Intel's 10nm is about 8.5nm on competing processes. Intel's 7nm is closer to TSMC's 6nm process. This is based on the transistor density per square inch.mooninite - Monday, October 28, 2019 - link
What 10nm products has Intel shipped for two years? Customer only stuff? I don't know of anything from 2017 that was 10nm... Someone want to report the CEO for lying to the SEC?DanNeely - Monday, October 28, 2019 - link
The only "two years" I see in the article is the timeline for 7nm. Two years from now in 2021. The slide for process techs even dropped the polite fiction that they shipped 10nm at the very end of 2018 (a crappy mobile cpu without graphics in a few China only laptops).MrCommunistGen - Monday, October 28, 2019 - link
Splitting hairs, but Intel also released a couple Canon Lake NUCs:NUC8i3CYSN
NUC8i3CYSM
I have no idea what kinds of numbers those shipped in -- but I'd bet that total volume was pretty low.
As I commented in the Anandtech Canon Lake review(where Ian reviewed the Chinese laptop back in January 2019) the NUC was expensive and was limited to 8GB of soldered down RAM.
drothgery - Monday, October 28, 2019 - link
Yeah; and you can definitely get Ice Lake laptops today (there aren't many of them, and they're in premium-priced products, but you can get them).GreenReaper - Monday, October 28, 2019 - link
Hey, it made it to AnandTech! A good option, if you needed AVX-512 on a 2018 laptop for some reason. https://www.anandtech.com/show/13405/intel-10nm-ca...AshlayW - Monday, October 28, 2019 - link
Suuuuure, Intel. I believe you. :Dsorten - Monday, October 28, 2019 - link
Gotta love the marketing slides, with no y-axis scale and not using a true zero baseline.Drumsticks - Monday, October 28, 2019 - link
I imagine this is an unpopular opinion, but, as delayed and challenged as Intel has clearly been, do people really expect them to be stuck on 14nm forever? They have to, at some point, break the 10nm barrier, and now is a reasonable time for them to do it. You can buy Icelake laptops right now from both Microsoft and Dell.Be skeptical of 7nm, sure, but Intel has objectively put Icelake into retail, and the CEO would be in hot water if he lied about the systems being introduced on an investor call. 10nm is here, and just because it isn't shipping in 95W parts doesn't mean it isn't shipping in high volume (how many desktop parts sell compared to Ultrabooks, anyways?)
FreckledTrout - Monday, October 28, 2019 - link
Intel will be perfectly fine. People are very skeptical, rightfully so, but make no mistake Intel will get out of this ditch.JoeyJoJo123 - Monday, October 28, 2019 - link
>Intel's already predicting 10, 10+, 10++.Given Intel's predictions and what ultimately turned into reality, I'm not going to be surprised if this turns into a 10++++++++ process 6 years or so later, with volume 7nm production nowhere in sight.
name99 - Monday, October 28, 2019 - link
"The company no longer sets ultra-ambitious goals for scaling each node"Which is why the claim is 2x density increase for 7nm?
And why they felt a need to talk about 5nm?
Seriously, they have learned NOTHING from the past.
In particular they certainly haven't learned to STFU until you have EVERYTHING validated.
Compare how little TSMC says until they are ready to roll. Even Samsung (way too chatty compared to TSMC) keeps their mouth shut a lot better, rather than promising how things will play out five years from now.
boeush - Monday, October 28, 2019 - link
Well, objectively 7 nm **is** a 50% higher area density than 10 nm (7*7=49; 10*10=100; 100/49~=2).That of course assumes that Intel's "nm" numbers actually have a real physical meaning, rather than amounting to nothing more than marketing fiction...
As for the talk of 5 nm, I'm pretty sure the likes of TSMC have already projected out as far as "3 nm" (whatever that actually means in reality) - and if the company is already spending significant R&D resources on a specific schedule for 5 nm, then the CEO would be remiss not to apprise the shareholders accordingly...
5080 - Monday, October 28, 2019 - link
TSMC just announced that they are building a whole new fab for its 3nm. The new fab should be operational in 2022. So much for keeping their mouth shut.name99 - Monday, October 28, 2019 - link
And what have they said about tihe ACTUAL 3nm process?https://www.tsmc.com/english/dedicatedFoundry/tech...
THAT is how it’s done.
TEAMSWITCHER - Monday, October 28, 2019 - link
One this is for certain ... there is never going to be a 0 nm process. What we are witnessing, is nothing short of the end ... of the Golden Age of Semi-Conductors.TheJian - Monday, October 28, 2019 - link
"yields are improving ahead of expectations for both for client and datacenter CPUs"So you've produced 16-32 core cpus on 10nm in HV? I only see quads and small chips.
Until you are selling at least a 16 core 100w 10nm you haven't started. 7nm after only 2yrs? LOL Only because 10nm was 4yrs late (still not out IMHO).
Not sure how BIG this 7nm gpu is either, as Intel said the first discrete card is SMALL CU count and LOW END. I'm confused by their statements here, as I though wimp desktop gpu was coming first (what a waste, like AMD 7nm NON king first...ROFL), and hope reality is BIG first as it should ALWAYS be if you want to win. Kings first, crap second.
Rudde - Tuesday, October 29, 2019 - link
Intel will release a small gpu on 10nm and later a bigger one on 7nm. I believe both are aimed at datacenter.Dizoja86 - Tuesday, October 29, 2019 - link
I'm pretty intrigued to see how 10nm impacts Intel's CPU lineup. Even at 14nm, their mobile chips have been untouchable from a power consumption perspective compared to AMD's offerings, so I'm wondering just how this shift will impact core counts and boost rates in that 15W (or lower) envelope.On the desktop side, here's hoping they can drop their total power draw down a bit while increasing clock speeds, although their current power draw isn't something that I'm particularly impacted by.
Regardless, as others have said, I'll believe it when I see it at this point. Current 7nm processes aren't particularly impressive, so I'm holding my breath that Intel's 10nm can match or beat what's currently on the market.
Rudde - Tuesday, October 29, 2019 - link
Everyone keeps talking about clock speeds, and that is exactly the reason why we'll continue to see 14nm cpus. 10nm has lower clock speeds and I don't believe Intel will ever reach 14nm clock speeds ever again. EUV (7nm) might be able to reach higher clock speeds than 10nm, but I doubt they will reach 5GHz beyond short bursts.5080 - Tuesday, October 29, 2019 - link
That's not true. Lower nano meter number = higher efficiency + higher performance + power efficiency.Less power also results in the generation of less heat and thus allowing us to increase the clock speeds further.
TheLightbringer - Tuesday, October 29, 2019 - link
I'll believe it when I see it.But Intel have been toying with the 10nm and not delivering for a long time. And sadly press and investors keep believing in them.
evilpaul666 - Wednesday, November 27, 2019 - link
Is there a reason nobody is making Pentium M Banias-style desktop setups with whatever the "shipping" 10nm chips are? Or no manufacturer is ignoring Intel spec with a higher TDP configuration laptop with the 10nm chips?