[Fool] Intel mobile gets another blow: Aicha Evans leaving

[Arachnotronic]

At an average tenure of 2 years per company he would have never completed a CPU.

Correct. However, it's important to make sure that the "8 companies" doesn't double-count any companies.

Pre-1998 he worked at DEC

Then between 1998 and 1999 he worked at AMD. People claim that he is responsible for the success that was K8 but clearly with 1 year there he couldn't have done much.

From 1999 to 2004 he worked at SiByte (which was acquired by Broadcom). From 2004 to 2008 he worked at PA Semi, which was then acquired by Apple, and he was at Apple from 2008 to 2012.

So really, he was with the PA Semi/Apple team from 2004 to 2012 -- that's eight years! He was at SiByte for 5 years. His second stint at AMD lasted from mid-2012 to late 2015.

But hey, I don't want to let facts get *too* in the way of the AMD worshipping that goes on around here.

 

[Phynaz]

^ Thanks for the details.

 

[videogames101]

If Intel mobile has been so garbage is this really a "blow" or a potential "boon"?

There are just as many shit exec's as good ones, and I have no idea whether this guy was good or not.

 

[Phynaz]

I remember lots of people saying the likes, that Keller's departure casts doom on Zen and K12.

Business as usual.

http://seekingalpha.com/news/3171354-veteran-intel-execs-davis-skaugen-depart

This could be twisted into Zen being that bad, that Skaugen didn't see some interesting and demanding fight ahead.

Rick Merritt says this is Venkata Renduchintala cleaning house. Makes sense, he's been on the job for long enough that he's going to make the division into what he envisions.

I don't think this has anything to do with AMD or Zen. This is the new guy getting rid of the old regime (and any potential challengers to the throne).

 

[witeken]

http://www.eetimes.com/document.asp?doc_id=1329351&page_number=3

He speculates Evans got the boot because she failed to get a design win in the iPhone 7 [or iPhone SE?].

That would make a ton of sense.

Ashraf speculates it's because she's frustrated, I'd rather think it's incompetence of her / her team.

 

[StrangerGuy]

Nice thing about x86 is that (unlike ARM) it is a standardized platform.

This means, unlike ARM, we should be able to update our OSes like we do our PCs. This should make for a much more secure system for those keeping their phones long term.

Its too bad most consumers (especially non enthusiast ones) wouldn't even realize this when factoring in their buying decision....and yet it is so important.

Yup, let's blame a ISA designer for Google's Wild Wild West decentralized Android model "anything for quick marketshare" where the OS dev takes next to no responsibility to update and secure the entire user base.

I think I lost several IQs points after seeing that post.

 

[cbn]

I still believe there is a huge opportunity for x86 in mobile. Just not in cheap android phones/tablets but for high-end win 10 devices.

Nice thing about x86 is that (unlike ARM) it is a standardized platform.

This means, unlike ARM, we should be able to update our OSes like we do our PCs. This should make for a much more secure system for those keeping their phones long term.

Its too bad most consumers (especially non enthusiast ones) wouldn't even realize this when factoring in their buying decision....and yet it is so important.

Yup, let's blame a ISA designer for Google's Wild Wild West decentralized Android model "anything for quick marketshare" where the OS dev takes next to no responsibility to update and secure the entire user base.

I think I lost several IQs points after seeing that post.

1. Google is not the point. Since there is no standardization with ARM the problem spans across other OSes as well. (Unlike x86 where we are free to mix and match OS and hardware)

2. ARM is working on standardizing the platform like x86 (see below link):

http://arstechnica.com/information-...orm-as-it-sets-its-sights-on-the-server-room/

ARM finally defines a platform as it sets its sights on the server room
"Server Base System Architecture" defines the building blocks of an ARM server.

by Peter Bright - Jan 29, 2014 4:46pm PST

In an attempt to strengthen the entry of ARM processors into the server market, British chip designer ARM has put together the Server Base System Architecture (SBSA), a definition of a standard platform for ARM-based servers. This move should reduce the abundant variation and complexity that has hitherto been a feature of ARM systems. SBSA was assembled by ARM along with its partners, including HP, Dell, AMD, Citrix, and Microsoft.

Even as ARM processors have proliferated in smartphones and tablets and are starting to make their first tentative steps into the server room, ARM has not been a platform in the way that the x86 PC is a platform.

Way back in the early 1980s, the IBM PC defined the way the computer booted, initialized its hardware, laid out its memory, and provided access to standard features like graphics and the keyboard. This enabled an ecosystem of PC software to develop. The PC platform was cloned by Compaq and others, and these clones were functionally equivalent to IBM machines. Operating system software that worked on one clone would work on any other, and it would work on the PC itself.

Over the years, the PC platform has changed, but this compatibility has remained as a core feature.

To the chagrin of operating system developers, ARM has lacked a comparable platform. Linux creator Linus Torvalds once described the proliferation of inconsistent, incompatible ARM systems as a "fucking pain in the ass," and implored the ARM community to "push back on the people sending you crap" and devise a common platform. Intel, likewise, has used this diversity to criticize ARM.

Since that statement in 2012 there has been some progress. Microsoft essentially defined an ARM tablet platform for Windows RT, enabling its kernel to work on both Qualcomm Snapdragon and Nvidia Tegra 2 and Tegra 3-based systems. Linux developers have also managed to consolidate their support for some of the diverse ARM platforms.

Without any clear market leader in the nascent ARM server market, this diversity and lack of platform could be deeply problematic. It would prevent easy software compatibility, with each different kind of system needing its own customized kernel.

The SBSA is ARM's effort to address that very problem. An operating system that targets SBSA will be able to run on any SBSA system: it will have the same basic platform components, put together in the same way, with the same kind of firmware, boot process, interrupt and I/O handling, hypervisor, and more. For example, SBSA will require all USB 2 controllers to conform with the EHCI 1.1 specification, all USB 3 controllers to conform with XHCI 1.0, and all SATA controllers to conform with AHCI 1.3.

Of course, the above standard is for servers. But I would assume this standardization could work for non server OSes as well (just like it does for x86). However, until this day comes it is up to the ARM device maker to provide updates....and on the lowest end devices this just doesn't happen.

 

[Nothingness]

Nice thing about x86 is that (unlike ARM) it is a standardized platform.

This means, unlike ARM, we should be able to update our OSes like we do our PCs. This should make for a much more secure system for those keeping their phones long term.

Its too bad most consumers (especially non enthusiast ones) wouldn't even realize this when factoring in their buying decision....and yet it is so important.

Do you think you can install Linux or Windows on any x86-based Android phone or tablet? Do you think these platforms follow the same "standards" as traditional laptops and desktops?

 

[cbn]

Do you think you can install Linux or Windows on any x86-based Android phone or tablet? Do you think these platforms follow the same "standards" as traditional laptops and desktops?

For an x86 Android tablet, provided it has enough RAM, storage and a compatible GPU then I don't see why it couldn't install Windows.

 

[Exophase]

Pre-1998 he worked at DEC

Then between 1998 and 1999 he worked at AMD. People claim that he is responsible for the success that was K8 but clearly with 1 year there he couldn't have done much.

Doesn't that mean he left AMD at their most competitive with the best still to come and things really heating up? You'd hardly say his leaving then was a bad sign for K8 the way people are saying this about Zen, would you? Especially when the move was to a company like SiByte...

I don't think we can really know one way or the other exactly why he left. Maybe something about AMD upper management consistently frustrates him.

 

[Nothingness]

For an x86 Android tablet, provided it has enough RAM, storage and a compatible GPU then I don't see why it couldn't install Windows.

Because of locked firmwares and the inability to change the early boot process.

 

[bhtooefr]

Also, the smartphone Atoms aren't IBM compatible, so they can't run PC OSes. The tablet ones, OTOH, are. (But, some tablets used Moorefield, not Bay Trail, so they're stuck running Android.)

 

[cbn]

Also, the smartphone Atoms aren't IBM compatible, so they can't run PC OSes. The tablet ones, OTOH, are. (But, some tablets used Moorefield, not Bay Trail, so they're stuck running Android.)

I remember in the Anandtech Moorestown article Anand mentioned the chip did not support PCI bus and as I recall in the same article or another one because of this it couldn't run (big) Windows 7 (but it saved power because of the omission of this hardware).

Pine Trail still has all of the bells and whistles of a PC platform however. Take the PCI bus for example. Every 12 microseconds it wakes up and polls every IO on the platform. That kills idle battery life, especially when you’ve got a tiny smartphone battery. Pine Trail is useless for smartphones, and that’s where Moorestown comes in.

If you thought this was the netbook Atom squeezed into a smartphone, you’re very wrong. It’s got a completely different memory controller, a true smartphone GPU (the same core, but clocked higher than what’s in the iPhone 3GS) and a ton of power optimizations that just don’t exist in the netbook version. The chipset is also very different. The PCI bus is gone as is anything that could ruin power consumption. Intel did a lot of optimization and a lot of cutting here. What resulted is something that looks a lot like a smartphone hardware platform and nothing like what we’re used to seeing from Intel.

This is Moorestown.

Did Microsoft change this PCI bus requirement for (big) Windows 8? Because from what I gather Clover Trail (Windows support) and Clover Trail+ (Android Support) only differed in their iGPU, but where both based on a phone chip:

http://www.anandtech.com/show/6790/...ualcore-cpu-and-graphics-unveiled-at-mwc-2013

 

[CHADBOGA]

Report: Intel Corp.'s Mobile Chief Bails

Will that be the dead blow? Or will it take until 5G until Intel gets some return on its huge investment?

My analysis is that because Intel was desparate to cut losses by 0.8B to 1B per year, they had to stop investing in the business and now it's sitting somewhere in no man's land.

You have to choose Intel: invest in mobile and gain a lot of MSS. Or don't.

I don't think losing Aicha Evans is a blow, but rather confirmation that Intel's multi-year failures in mobile are far from over.

 

[StrangerGuy]

I don't think losing Aicha Evans is a blow, but rather confirmation that Intel's multi-year failures in mobile are far from over.

It's all speculation, but losing the iPhone baseband socket (very possibly with contra-revenue I bet) let alone a competitive mobile SoC is pathetic to say the least for the mobile division.

 

[CHADBOGA]

It's all speculation, but losing the iPhone baseband socket (very possibly with contra-revenue I bet) let alone a competitive mobile SoC is pathetic to say the least for the mobile division.

What isn't speculation is Intel's failure to achieve anything meaningful in phones and even tablets.

With tablets, their failings haven't been as spectacular as their failings in phones, but overall it is still a disappointing effort to date.

 

[VirtualLarry]

What isn't speculation is Intel's failure to achieve anything meaningful in phones and even tablets.

With tablets, their failings haven't been as spectacular as their failings in phones, but overall it is still a disappointing effort to date.

I, for one, like my contra-revenue $60 Windows 8.1 7" tablets. Sure, they could use a bit more than 1GB RAM and 16GB eMMC, but for the price, they've been pretty phenomenal and useful little toys.

Who could have imagined a full Windows 8.1 PC, in the palm of your hand, for less than $100?

 

[frozentundra123456]

I agree. The sub hundred dollar atom tablets are very useful for the price. The problem is that there was really no adequate follow up.

Even now, Cherry Trail has not really caught on, and when it is used, even at 14nm, it is barely an improvement over Bay Trail, at least in cpu performance. Now I would gladly move up to a more expensive device if Cherry Trail were a marked improvement or even better, if they would put core M into 250 dollar tablets instead of 899.00 surface models. But they got their foot in the door, at the cost of contra revenue, and now have not taken advantage of it.

 

[Doom2pro]

Going after mobile would be suicide for AMD. Intel has burnt billions on it, with nothing to show for their efforts. Nvidia has similarly struck out, buying and then having to write off Icera. Even companies which had successful mobile chip businesses like TI have walked away because the margins are too slim and competition too fierce.

Being squeezed between integrated manufacturers like Samsung, giants like Qualcomm and zero margin firms like Mediatek sounds like a death sentence for AMD.

Intel has crap graphics IP, nVidia has crap CPU IP and AMD has Good Graphics, CPU and overall good SoC IP...

Why the hell not?

I'd rather have a smart phone with AMD x86/Arm coupled with Radeon Graphics than nVidia/Adreno crap.

 

[Shivansps]

AMD graphics IP is not doing great on low TDP enviroments, so i dont see Intel getting this for low tdp enviroments, and i find hard to belive that AMD wants Intel to get even better at notebooks/desktop graphics.

I, for one, like my contra-revenue $60 Windows 8.1 7" tablets. Sure, they could use a bit more than 1GB RAM and 16GB eMMC, but for the price, they've been pretty phenomenal and useful little toys.

Who could have imagined a full Windows 8.1 PC, in the palm of your hand, for less than $100?

http://t i n y u r l . c o m/jgwvz63 some great ones for a bit more.

 

[beginner99]

But I dont think anyone can reasonably say someone is leaving Intel's mobile division because the work is done and the products are great.

True but if someone leaves it could just mean he got a better offer. Not more, not less.

I still believe there is a huge opportunity for x86 in mobile. Just not in cheap android phones/tablets but for high-end win 10 devices. Both AMD and Intel could create a huge market in that segment but Intel wanted something not meant for them, they got burned and destroyed AMDs mobile plans with them.

Not sure intel destroyed that but else I agree. There would be a huge space for x86 in mobile. With continuum they could run full windows 10.

And as I already wrote couple of times a dock could be provided for connectivity but also better cooling. So the phone could contain a core-m like CPU and run at low frequency in mobile and then up it's TDP when in this "cooling dock".

It's obvious that businesses would be really, really interested in this. Managers and executives need mobiles already anyway and a laptop. Why not save money and give them only 1 device they can use as both? There could also be a mobile dock with a screen, battery and keyboard. Kind of an empty laptop with a slot to push-in your phone.

In actual real world programs that were decently optimized about the same for ARM and x86 Saltwell was a little behind Cortex-A9 and Krait in perf/MHz while offering similar MHz to Krait (and a little more than A9s) so I don't think it really blew away ARM phones in single core performance. And then there was the whole matter of relying on binary translation for Android apps that didn't support x86 natively, something which destroyed performance. Intel brushed this off by saying ~80% of Android apps didn't even use the NDK, but that was certainly not the case when weighted by app popularity.

Also, Medfield was single-core with HT while ARM phones were already quad-core. So in anything even moderately threaded it hurt pretty badly.

I owned a Motorola Razr I with Medfield in it. It got stolen on vacation after couple of month. I then bought the LG Optimus G which is based on the same platform s the Nexus 4. It was marginally more expensive but with better specs (screen, quad-core vs single-core). I'm pretty happy about that decision. The LG is the superior phone. It's still my current phone btw.

The main difference between the 2 besides the form-factor was that the Medfield one had annyoing long load times for many apps. This was probably due to the binary translation. In app, there was no noticeable differences.

 

[PingSpike]

I, for one, like my contra-revenue $60 Windows 8.1 7" tablets. Sure, they could use a bit more than 1GB RAM and 16GB eMMC, but for the price, they've been pretty phenomenal and useful little toys.

Who could have imagined a full Windows 8.1 PC, in the palm of your hand, for less than $100?

I got a $5AR tablet like that from tigerdirect while they were shutting down. I've only powered it on once. I can't think of a use for it and I've tried! My kindle fire HD wifi goes out to lunch all the time but it has a much better screen. I couldn't resist for the price of a sandwich.

 

[ksec]

I think the writing has been on the wall for a long while, and Intel literally confirm it when they announced 7480. Why?

It has long been suspected that Intel are using IP from CEVA, and that announcement confirms it. This makes Intel Modem not that much different from Samsung or Mediatek. Although the announcement only confirm the DSP IP were used, I suspect other part of the baseband modem were using CEVA as well.

The other thing is 14nm. Which is Intel's own 14nm. All the modem baseband so far has been on TSMC 28nm. I was surprised when Intel announced 7360 was 28nm as well. Although Intel never gave an official answer, investor and media was getting the design for 7360 was set a long time ago and couldn't have time to port those to a much different Intel 14nm.

On paper it was ok, Intel promised their design would give similar performance to other modem at TSMC 20nm. And, again most would expect once the modem design was move to Intel 14nm it would give them a slight edge.

But that didn't happen. Not only did the new 7480 wasn't on Inte 14nm, it was still stuck with TSMC 28nm. Not even 20nm, not 16nm FFC, but 28nm!!

The features set here isn't as important as most, ( Ashraf Eassa ) thinks. Mainly because the baseband modem mainly consists of two group, that is Apple, and many others. Most use intergrated solution like Snapdragon, or goes along with Mediatek. The actual baseband modem only market is shrinking in market% size. The only big baseband modem client is Apple. And Apple has always been VERY conservative on Network baseband tech. Mainly because they test out the baseband with literally every Major Mobile network on earth before rolling out those features. And therefore running behind in LTE speed and features isn't a major concern for Apple.

But Power consumption is a major concern, Qualcomm has shown to be very aggressive in pushing LTE boundary, and they have also execute their roadmap to perfection. The new DSP core in X16 modem is again one step ahead of the new CEVA DSP core. And it will be on TSMC 16FFC.

So basically I am sure Apple has received every engineering sample from Intel and they were not happy with it. The roadmap Intel gave were delayed after delay. Cost and power consumption were not up to compete. The Fab and tools Intel have are also likely to be not as good for Modem design compared to TSMC. Apple likely hyped the Intel modem to the press themselves to get better deal from Qualcomm.

 

[ksec]

On the 5G notes, let just define what is 5G first, because everyone seems to have different view on this.

Personally I say whenever we have a new air interface, like LTE which uses OFDMA instead of WCDMA in 3G, I call this a new G.

Since we dont even have a agreement on 5G air interface yet. It will likely be a long time (2-3 years at least ) before it is tested and ready for standard, the implementation etc will take another few years.
* There is a exception for this when one of the front running air interface is a evolved/ tweaked version of OFDMA.

What most newspaper and network say trail and testing 5G in 2017 is actually 3GPP rel 14. More like 4.9G. Which includes two major inclusion. LTE-U to help and gain adoption of Smell Cells. And Massive MIMO, but will increase the total capacity of the network by at least 3x, ( Theoretically it is up to 10-20x, but real life and promise of tech always fall short, being conservatives think 3x is very achievable, with real world closer to 5x ) which means users are likely going to see speed improvement as well, as they get more air time slot.

This alongside with more spectrum alliocation to LTE should give the user feeling like a new G. Alongside with some Lataency optimisation.

What this 4.9G/5G means for Intel? I have yet to see anything from Intel on consumer side of things. But they are ready to take on some backend Giants, mainly on the Cloud RAN /software defined network stack, where X86 and FPGA intergration will give flexibility and saving cost.

 

[ksec]

Also worth mentioning, Intel Head of IoT and 5G Doug Davis is also retiring.