Question Intel Q2: 7 nm in bad shape

Page 8 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

lobz

Golden Member
Feb 10, 2017
1,250
1,244
106
  • Like
Reactions: spursindonesia

Lodix

Senior member
Jun 24, 2016
299
85
101
If the history of Samsung's reactionary press releases tells us anything, it's that their 5/4nm process development is indeed delayed.
Yeah, they usually publish a press release when they start risk/mass production of a given node. And they did announce the risk production of 5LPE a long time ago but nothing about its mass production yet ( apart from this ). Maybe it is just a little delay like 7LPP. But it is misleading anyway.
 
  • Like
Reactions: Tlh97 and lobz

JustMe21

Senior member
Sep 8, 2011
299
22
81
What if Intel bankrolled Global Foundries' 7nm rollout? GFO seemed about ready to go with 7nm, but didn't seem to have enough financial incentive to put it into production.
 

Kenmitch

Diamond Member
Oct 10, 1999
7,740
857
126
That image . . . is that real? Or did someone photoshop that?
It's funny, but it's a photoshop. Intel uses a different shade of blue.

The original photo

 
  • Haha
Reactions: Tlh97

scannall

Golden Member
Jan 1, 2012
1,674
1,037
136
What if Intel bankrolled Global Foundries' 7nm rollout? GFO seemed about ready to go with 7nm, but didn't seem to have enough financial incentive to put it into production.
Or just buy them. GF had 7nm ready to go but didn't want to spend the money on the equipment. That was a gift from IBM that they threw away.
 
Last edited:

Ajay

Diamond Member
Jan 8, 2001
7,064
2,364
136
Intel has plenty of fab space. What they lack, is a viable 7nm process. If they could buy a license for a good 7nm, that might work. But only TSMC has a process that would be useful to them (7N EUV).
Intel would still be behind by the time the got such a process setup and and running HVM. What they would really need is TSMC's 5N process, and then redesign their products while that process is setup.
There is Samsung, but they are having problems as well. Maybe Intel and Samsung could form a joint venture on process tech and catch up to TSMC?
 

NostaSeronx

Diamond Member
Sep 18, 2011
3,031
612
136
Or just buy them. GF had 7nm ready to go but didn't want to spend the money on the equipment. That was a gift from IBM that they threw away.
STMicro(ISDA)/GloFo = 7nm DUV (56nm Cx/Mx SAQP 40nm) <== Not enough customers (+EUV option only targeted SAQP layers)
IBM/GloFo = 7nm EUV (44/48nm Cx/Mx 36nm) <== Not enough money (production wise this was 5LP)
IBM/GloFo = 5nm SNS EUV (44/48nm Cx/Mx 32nm) <== Not enough money (production wise this was 3LP)

22FDX/12FDX/10LP/7LP was from STMicroelectronics. Pretty much everything in the first two leading performance FinFETs were patented and researched in the ISDA by STMicro first; https://patents.google.com/patent/US9685380B2/en , etc.
 
Last edited:

name99

Senior member
Sep 11, 2010
214
201
116
It's half-broken. They have some 10nm parts out, albeit mobile. This wouldn't be an issue if 10nm were based on a new µarch and they could lend a 40-60% IPC increase at those reported clocks which would obviously put it in line with a new architecture, and one we haven't seen in close to 20 years. The rest of your post is explaining stuff I know already or repeating stuff already said. I sold all my Intel shares nearly a year ago. I already knew how bad 7nm was.
It' worth looking at those 10nm parts. We have
- Cannon Lake. (Oh dear, let's forget that right way).
- Ice Lake. Slightly coooler than 14nm+++Lake, but mainly ~same peak performance
- Lakefield. Woohoo, competing strongly against a mid-range Qualcomm chip

This is not a power lineup. But that's not the point, the more important point is the qualifiers.
The biggest Ice Lake that ships is 4 cores, die size of 122mm^2. If the process is yielding great, why were no desktops at say 6 or 8 core ever released?
Why was Lakefield crippled with just one SC core rather than two?
And Ice Lake is being followed up not by a desktop part, but by Tiger Lake, another mobile quad core part.

Look at this pattern. Everything suggests that they are having yield problems! They're limiting themselves to the smallest parts they can get away with shipping while still matching AMD (sometimes to absurd lengths, as with single core SC on Lakefield). They apparently have no interest in moving the desktop to 10nm. Ice Lake SP has slipped from "1st half of 2020" to "Q3 2020" to (a few days ago) "late 2020" with language vague enough that it's probably early 2021.

It's hard to look at this and conclude "10nm is yielding great, can handle large dies, and is ready to be scaled up across the product line.
 

DrMrLordX

Lifer
Apr 27, 2000
16,023
4,970
136
It's funny, but it's a photoshop. Intel uses a different shade of blue.
Okay, thanks for the clarification. I figured it wasn't real.

@name99

Lakefield only has one Sunny Cove core to keep idle power down. Also you really don't seem to be disagreeing with anything he's said and are just needling the guy. We all know 10nm is b0rked. Yields appear to be bad, die sizes are staying small, etc. It sort of works but not across an entire product lineup, so it is half-broken.
 

Doug S

Senior member
Feb 8, 2020
250
301
96
There is Samsung, but they are having problems as well. Maybe Intel and Samsung could form a joint venture on process tech and catch up to TSMC?
I just posted in another thread I think Samsung would be a more likely partner for Intel, especially if Intel wants to divest its fabs. They are having trouble keeping up with TSMC but that's because they don't have nearly the money coming in from their foundry business TSMC does. Partnering with Intel would give them the money they need to keep up with TSMC, and their slack capacity would make it easier for Intel to migrate off their own processes while their own fabs were being upgraded to the latest Samsung process.

Though I still am not a believer in the idea Intel is going fabless. Things aren't bad enough for that to be necessary yet.
 
  • Like
Reactions: Tlh97 and scannall

Ajay

Diamond Member
Jan 8, 2001
7,064
2,364
136
I just posted in another thread I think Samsung would be a more likely partner for Intel, especially if Intel wants to divest its fabs. They are having trouble keeping up with TSMC but that's because they don't have nearly the money coming in from their foundry business TSMC does. Partnering with Intel would give them the money they need to keep up with TSMC, and their slack capacity would make it easier for Intel to migrate off their own processes while their own fabs were being upgraded to the latest Samsung process.

Though I still am not a believer in the idea Intel is going fabless. Things aren't bad enough for that to be necessary yet.
Yeah, we are both cross-posting. IMHO, being 1 node behind is something I think Intel has shown it can handle, being 2 nodes behind is not sustainable; vis-a-vis it's competition with AMD. The obvious 800 pound gorilla in all this is the fact that Intel missed out on all the mobile (ARM) business. Massive blunder in terms of fab income.
 

mikk

Platinum Member
May 15, 2012
2,733
603
136
10nm is surely not yielding great but the progress over the last months is more than obvious. Tigerlake based on a tweaked 10nm is coming in a couple of weeks with much better clock speeds and by looking at the availability of Icelake-U it gets better and better. Icelake-U closes in to CML-U when it comes to devices. The difference was huge last year or even half a year ago but by know it closes in to CML-U. There are like 350 different ICL-U devices/configurations available compared to 550 CML-U devices in Germany, at some point next year we could see the turnaround where 10nm excels 14nm in availability of mobile ULV.
 

moinmoin

Golden Member
Jun 1, 2017
1,657
1,594
106
10nm is surely not yielding great but the progress over the last months is more than obvious. Tigerlake based on a tweaked 10nm is coming in a couple of weeks with much better clock speeds and by looking at the availability of Icelake-U it gets better and better. Icelake-U closes in to CML-U when it comes to devices. The difference was huge last year or even half a year ago but by know it closes in to CML-U. There are like 350 different ICL-U devices/configurations available compared to 550 CML-U devices in Germany, at some point next year we could see the turnaround where 10nm excels 14nm in availability of mobile ULV.
So they are slowly reaching the state originally planned for 2016/17. Congratulations Intel, I guess?
 

clemsyn

Senior member
Aug 21, 2005
397
8
81
Now the Murthy is out, do you guys think Intel's 7nm and beyond would yield better under Kelleher. I mean she has been working under Murthy (although she doesn't have the final say) with that node.
 

Ajay

Diamond Member
Jan 8, 2001
7,064
2,364
136
Now the Murthy is out, do you guys think Intel's 7nm and beyond would yield better under Kelleher. I mean she has been working under Murthy (although she doesn't have the final say) with that node.
Who knows. There is some sort of problem with Intel's process development group. What it is, I don't know. I do know that they have lost some key people (like Bohr, who 'retired').
 
  • Like
Reactions: CHADBOGA

CHADBOGA

Platinum Member
Mar 31, 2009
2,010
590
136
Who knows. There is some sort of problem with Intel's process development group. What it is, I don't know. I do know that they have lost some key people (like Bohr, who 'retired').
Has process tech at the smaller nodes become so complicated that the number of people capable of advancing it, has shrunk dramatically and if you don't get some of those small handful of process tech geniuses, it won't matter how much money you throw at things?

I wonder if Bohr didn't have the knowledge/brain power to go much beyond 14nm, even if he had been a superstar for the last 30 years in process tech?

Has TSMC been lucky enough to snap up most of the people capable of advancing process tech in a timely fashion right now?
 

maddie

Diamond Member
Jul 18, 2010
3,159
1,829
136
Has process tech at the smaller nodes become so complicated that the number of people capable of advancing it, has shrunk dramatically and if you don't get some of those small handful of process tech geniuses, it won't matter how much money you throw at things?

I wonder if Bohr didn't have the knowledge/brain power to go much beyond 14nm, even if he had been a superstar for the last 30 years in process tech?

Has TSMC been lucky enough to snap up most of the people capable of advancing process tech in a timely fashion right now?
Remember, we're always at the state of the art. The Wright bros were just as innovative as the latest designers.
 

clemsyn

Senior member
Aug 21, 2005
397
8
81
Has process tech at the smaller nodes become so complicated that the number of people capable of advancing it, has shrunk dramatically and if you don't get some of those small handful of process tech geniuses, it won't matter how much money you throw at things?

I wonder if Bohr didn't have the knowledge/brain power to go much beyond 14nm, even if he had been a superstar for the last 30 years in process tech?

Has TSMC been lucky enough to snap up most of the people capable of advancing process tech in a timely fashion right now?
Honestly, I think Intel will have a very hard time with advancing their process tech. They need to advance their shrinking nodes to compete. I don't think Swan is the right person for the job, he is a finance guy. I also don't think the current board is up to the task, what Intel needs is a an active shareholder to shake up this company and put an engineering person for CEO. I like AMD but we need Intel for competition.
 

JasonLD

Senior member
Aug 22, 2017
253
203
86
Honestly, I think Intel will have a very hard time with advancing their process tech. They need to advance their shrinking nodes to compete. I don't think Swan is the right person for the job, he is a finance guy. I also don't think the current board is up to the task, what Intel needs is a an active shareholder to shake up this company and put an engineering person for CEO. I like AMD but we need Intel for competition.
Brian Krzanich was an engineer. :rolleyes:
 
  • Like
Reactions: spursindonesia

clemsyn

Senior member
Aug 21, 2005
397
8
81
Brian Krzanich was an engineer. :rolleyes:
Yes he was, but was more interested in drones :). The board made a poor decision in making him CEO. They had quite a few choices back then. Brian also axed potential CEO's and hired Murthy to do the process development.
 
Last edited:
  • Like
Reactions: CHADBOGA

ASK THE COMMUNITY