Aicha Evans stays after all: convinced by Brian Krzanich

[witeken]

Pretty spectacular.

http://seekingalpha.com/news/317518...ithdraws-resignation-street-debates-amd-china

http://fortune.com/2016/04/22/intel-aicha-evans-staying

PS: To Ashraf Eassa. About Moore's Law and tech leadership.

Yesterday I crunched the numbers and here are my conclusions.

* In short, TSMC and Intel both do something different: Intel is going for >= 2x density improvement, and thus sacrificing TTM because scaling is so hard. TSMC OTOH is going for less aggressive scaling, which has shorter TTM.

* They are not better or worse strategies from the data that I have available. Intel's leadership should still be *at least* 1 year. But that is IF and only IF TSMC operates perfectly (which is doubtful) without yield difficulties, and that's probably also 1 yr against Apple because they will be first, so versus other companies it will be more.

* The big unknown is what technologies the process nodes will adopt. Even if Intel's lead has shrunk to 1 year (density), if they have better transistors, the real lead will still be many years. Intel already has air gaps at 14nm, so that shows how good they are. TSMC's 10nm is finFET but 7nm is unknown, most likely SIGe. Intel's introduction of III-V/Ge and quantum well could be at 10nm, but that is unconfirmed.

We simply don't know.

* Okay, now back to scaling.

TSMC 16nm: 5760 nm²
Intel 14nm: 3650 nm²
(Difference: 1.55x denser)

TSMC 10nm (claim 0.53x so): 3053 nm² (2017) (finFET)
Intel 10nm (claim ~0.5x, let's take 0.52x): 1898 nm² (H2'17) (unknown) [if they achieve 2.1x like expected it will be 1740 nm²]

TSMC 7nm (claim 1.6x = 0.652x so): 1908 nm² (without EUV) (probably H2'18 for iPhone although claim is H1 -- don't forget TSMC was 2Q behind SS at 14/16nm)

So Intel 10nm will be *at least* as dense as TSMC 7nm AND have *at least* 1 year TTM advantage AND have *most likely* a much better transistor.

 

[frozentundra123456]

One can argue about process and this node vs that node, but performance is the bottom line. They are fine in the core line, although I am very interested how samsung 14 nm will perform for zen. But in the low power mobile, they just arent competitive from either a performance or feature standpoint. If in fact their process lead still exists, it makes it unfathomable why they *still* cant bring out a competitive mobile chip, and in fact almost seem to have given up.

 

[PPB]

Spectacular that the guy that should be responsible for intel's ultramobile losses is staying after all? For qualcomm and co. Probably.

Sent from my XT1040 using Tapatalk

 

[ksec]

Unless Intel move the 7380 to their 14nm, I find it hard to believe Apple would pick a modem based on TSMC 28nm. Heck even the newly announced 7480 bares no mention of Intel Fab, which means I presume would still be on TSMC. Even if it is similar in capability 28nm 7380 would be much more power hungry then 16nm.

And why she resign when she have landed the Apple order? The head of IoT also happen to retire at the same time.

 

[waffleironhead]

One rat ready to jump from the sinking ship was lured back aboard with a piece of cheese. Spectacular indeed.

 

[Exophase]

Call me cynical, but after Brian Krzanich announced a diversity initiative stressing an increase in the representation of women and minorities in their workforce (with $300m behind it) I can see why he would put extra attention in convincing a high ranking black woman to stay.

That could have absolutely nothing to do with it but after all of the statements he's made it's hard to remove the possibility from my mind.

Then again, this could also be a matter of trying to downplay the situation to the outside.

 

[zentan]

Yea and stuffs like the extra commission for recommending minorities and other stuffs is just ridiculous. Not to mention the attacks and unproven allegations against Linus by their employee,playing the minority card. I hope it backfires pretty soon on them if they are going to stick with such privilege and appeasement policies
Anyway like you said,it might not be the case here.

 

[Arachnotronic]

Spectacular that the guy that should be responsible for intel's ultramobile losses is staying after all? For qualcomm and co. Probably.

Sent from my XT1040 using Tapatalk

Aicha Evans was on the job for less than a year before tendering her resignation. She is not responsible for the mess you currently see, but it is her job to try to clean it up.

 

[TurtleCrusher]

These kind of posts belong in Politics and News. They don't have jack to do with CPUs or overclocking.

 

[ksec]

Aicha Evans was on the job for less than a year before tendering her resignation. She is not responsible for the mess you currently see, but it is her job to try to clean it up.

I get the idea then may be she had enough with company not getting anything done. Anyway I really do hope Intel get their modem act together. iPhone could enjoy the latest mode on both modem and SoC without going into capacity problem.

The problem is even with 7480 it is still no where near as good as Qualcomm

 

[NTMBK]

Call me cynical, but after Brian Krzanich announced a diversity initiative stressing an increase in the representation of women and minorities in their workforce (with $300m behind it) I can see why he would put extra attention in convincing a high ranking black woman to stay.

That could have absolutely nothing to do with it but after all of the statements he's made it's hard to remove the possibility from my mind.

Then again, this could also be a matter of trying to downplay the situation to the outside.

Pretty disgusting that you can't see past her race and gender. I doubt Intel would ask her to turn around their failing mobile division unless she was actually extremely talented.

Hope she manages to make a difference, but I think that the mobile division is holed below the waterline. Maybe they could turn the modem business around, but without a great AP to integrate it with I don't know how much market there is to grab. Might be better off just selling the modem business to Apple, scrapping Atom, and focusing on their competencies.

 

[Exophase]

Pretty disgusting that you can't see past her race and gender. I doubt Intel would ask her to turn around their failing mobile division unless she was actually extremely talented.

I figured someone would lay this back on me. But I'm not the one talking about specifically wanting to improve race and gender representation, and I've heard from Intel employees that their biggest emphasis was on improving the representation in the higher ranks. So really I don't think I'm the one who is having a hard time seeing past her race and gender, when Intel has made it clear that they're concerned about this.

Please don't get the wrong idea here, I am NOT saying that her gender or race makes me suspicious of her qualifications. I have no doubt that she is very talented and suited for the position, but it's pretty unusual to hear about the company's CEO pleading with someone (especially a relative newcomer) to not resign when their division was not performing up to expectations. There could be more to this story, I don't really know, but Evans herself was saying she'd be out of a job if performance wasn't up to snuff. And you don't exactly see Intel pleading with the others to stay. Can you think of another case where a big leader within the company, someone who interacted with the press was slated to be resigning then didn't because the CEO asked them not to?

With Intel being very vocal about their diversity initiatives (that frankly sound a lot like enforcing quotas) they plant these kinds of doubts in people's minds that wouldn't be there otherwise. But I guess anyone who thinks that can just be rejected as bigots or worse, so it's no real problem.

 

[Arachnotronic]

These kind of posts belong in Politics and News. They don't have jack to do with CPUs or overclocking.

If you're not interested in the discussion, why try to ruin it for those who are? Aicha Evans runs Intel's mobile group, which makes mobile CPUs.

 

[Exophase]

If you're not interested in the discussion, why try to ruin it for those who are? Aicha Evans runs Intel's mobile group, which makes mobile CPUs.

Any comment on witeken's call out re TSMC 7nm being marginally less dense than Intel 10nm? Do you agree with the math?

 

[Arachnotronic]

Any comment on witeken's call out re TSMC 7nm being marginally less dense than Intel 10nm? Do you agree with the math?

Tough to call, honestly. They will both be using SAQP for critical metal layers at those nodes so it's not as though either company will have a fundamental lithography advantage. At 14nm Intel used SADP while TSMC used LELE; the former forces more strict design rules but affords tighter pitches, while the latter doesn't get you as much density but is easier to design to.

I think it will be very close between the two and the main differentiating factor will really be yields & transistor performance. I don't have much insight into what Intel might be doing in terms of 10nm as far as performance enhancements because, in what seems to be typical fashion for them these days, they are not offering any insight into what changes they plan to make at 10nm over 14nm.

 

[Exophase]

Tough to call, honestly. They will both be using SAQP for critical metal layers at those nodes so it's not as though either company will have a fundamental lithography advantage. At 14nm Intel used SADP while TSMC used LELE; the former forces more strict design rules but affords tighter pitches, while the latter doesn't get you as much density but is easier to design to.

Is TSMC making this change with 10nm or 7nm? I guess at this point they've already only really taken on huge designers for their new nodes, they'll just have to deal with the added difficulties.

I think it will be very close between the two and the main differentiating factor will really be yields & transistor performance. I don't have much insight into what Intel might be doing in terms of 10nm as far as performance enhancements because, in what seems to be typical fashion for them these days, they are not offering any insight into what changes they plan to make at 10nm over 14nm.

Whatever 10nm brings, I'm going to be super surprised if it includes III-V. I feel like there have been a lot of statements now that have spoken against that possibility.

 

[Arachnotronic]

Is TSMC making this change with 10nm or 7nm? I guess at this point they've already only really taken on huge designers for their new nodes, they'll just have to deal with the added difficulties.



Whatever 10nm brings, I'm going to be super surprised if it includes III-V. I feel like there have been a lot of statements now that have spoken against that possibility.

Exophase,

TSMC has mentioned triple patterning for 10nm, so I would expect they are implementing LELELE triple patterning there. 10nm is not expected to be a long node; it is basically being done to service Apple and a few other companies are riding the coattails.

As far as 10nm, don't expect III-V materials from any of the players, including Intel. I expect the big materials changes from everybody (i.e. moving away from FinFET) will happen at what Intel/Samsung call the 7nm node and what TSMC calls its 5nm node (i.e. nodes that will go into mass production in ~2020).

 

[Exophase]

TSMC has mentioned triple patterning for 10nm, so I would expect they are implementing LELELE triple patterning there. 10nm is not expected to be a long node; it is basically being done to service Apple and a few other companies are riding the coattails.

This seems obvious, but I don't think I really quite appreciated the extent to which Apple is driving TSMC's development.

If Intel really did manage to become a foundry for Apple it could be highly disruptive.

 

[Arachnotronic]

This seems obvious, but I don't think I really quite appreciated the extent to which Apple is driving TSMC's development.

If Intel really did manage to become a foundry for Apple it could be highly disruptive.

Apple is basically the reason that you're seeing TSMC move so aggressively, I hear that internally the biggest priority is to keep Apple happy. TSMC understands that as far as customers with the volumes to justify such intensive leading edge technology go, Apple is pretty much it.

I don't think Intel will be a foundry for Apple.

 

[TurtleCrusher]

If you're not interested in the discussion, why try to ruin it for those who are? Aicha Evans runs Intel's mobile group, which makes mobile CPUs.

If that's the case, it still goes into politics and news. A stretch would be placing it into SFF, Notebooks, Pre-Built/Barebones PCs.

 

[Exophase]

If that's the case, it still goes into politics and news. A stretch would be placing it into SFF, Notebooks, Pre-Built/Barebones PCs.

My commentary about Aicha Evans probably didn't belong here at least >_> Sorry about that, I regret my posts about that :/

 

[witeken]

Exophase,

TSMC has mentioned triple patterning for 10nm, so I would expect they are implementing LELELE triple patterning there. 10nm is not expected to be a long node; it is basically being done to service Apple and a few other companies are riding the coattails.

As far as 10nm, don't expect III-V materials from any of the players, including Intel. I expect the big materials changes from everybody (i.e. moving away from FinFET) will happen at what Intel/Samsung call the 7nm node and what TSMC calls its 5nm node (i.e. nodes that will go into mass production in ~2020).

I would still disagree.

The problems all have to do with scaling, not with the transistor itself. Intel is very good at making very powerful transistors and technology, like air gaps.

So I still bet 10nm, but we'll see. Although I could be very telling that they're not bragging about how good their 10nm node is.

 

[386user]

exophase i agree completely and it was a relevant comment

i am an intel shareholder as an aside