Intel cuts Atom chips, basically giving up on the smartphone and tablet markets | Hacker News Reader

Intel cuts Atom chips, basically giving up on the smartphone and tablet markets

pcworld.com

122 pointsIvoah59 comments

I might be biased but I think this is a maybe looks good on paper decision but a terrible long term strategy decision. Just because they can't compete on price doesn't mean they can't compete on features.

It was only two months ago that they were showcasing their android platform and practically begging for customers. [1]

I've wanted to do a phone for awhile. I'm sure I'm not the only one. In the last few months I contacted people at Intel but I had no response whatsoever to my multiple inquiries (to you know, become a customer). And I have some of the company's most senior management in my LinkedIn. Now I know why.

This is kind of pathological behavior on Intel's part imo. The big players were never going to come to their platform. Huawei, HTC, Samsung, LG etc have no reason to change from ARM. They are content to keep doing their cookie cutter thing while fighting a brutal price war in a staid market.

On the other hand, the startups out there, those of us who want to do something really out of the box and grow new markets can't get a word in edgewise. It's disappointing. As a technologist, I really hope someone manages to ship an open source handheld one day. That will be the catalyst for the next iteration of smartphone evolution.

[1] http://www.theregister.co.uk/2016/02/23/move_over_continuum_...)

>> open source handheld one day.

I would guess that this is beyond Intel's pay grade and it is a political/security decision.

Also except for selling openness, what else will sell that you couldn't do today ?

You can get a raspberry pi with an atheros chipset wifi dongle and you have pretty much a completely open source stack. It just doesn't look very nice to users.

>> Also except for selling openness, what else will sell that you couldn't do today ?

Anything that doesn't fit into Google's technical design / strategy that sells and customers like it.

Android has heaps of preconceived notions about how a handheld should work that are baked into the system and unchangeable.

If you are looking for exact examples of things that are not currently technically possible on android, how about a phone with a gps button widget that doesn't require you to leave the maps app to enable/disable gps or (not that I would personally pursue this) the ability to play video in a distributable app widget.

You don't have to have the processor to be open to do all those, do you ?

And the pi/atheros is open exactly because they're not a mass market product.

Raspberry Pi had sold 5 million devices by Feb 2015. It's a bit different from 7 figures per quarter, but I'd still call it mass market.

Why would you use Intel products if you wanted to build an open-source handheld?

I'm looking forward to the not-very-distant day that we no longer need to use Intel products so that we can build open-source laptops.

I just want to be able to call all the shots on the design of my product.

Intel has historically been very open on the desktop. Linux and computing in general wouldn't be where it is without that openness.

I suppose we have different ideas about what "open" means. In my view, Linux and computing in general have gotten where they are largely despite Intel's efforts at rent-extraction.

Wintel provided the mass market that enabled the economies of scale that drove down prices. Linux has piggybacked on that.

Linux was written for the 386....

If it wasn't for the likes of IBM and Intel, with Linux developers on their payroll, even though they also do lots of patents and closed source stuff, GNU would still be looking for a kernel.

> GNU would still be looking for a kernel.

Linux came about before there was commercial interest. Linux (or GNU/Linux if you will) would have existed. Whether it would be in the same place it is today is another story.

I was there. We just wanted to have a free UNIX clone at home, to be able to do the work from there instead of expensive connections over the phone line to corporate and university servers.

If that was all you wanted, you'd still be running MIMIX.

It would be great if RISC-V and other open CPU architectures dominate the CPU and SoC market, but that is going to be a long time coming. lowRISC seems to be the best chance for that.

http://www.lowrisc.org/

RISC-V seems to be lacking on the software side, e.g. where is the upstream support for Linux, GCC, or llvm? There is support in FreeBSD, but only because one of the FreeBSD committers was able to work on this full time (I've been providing advice to him as I ported FreeBSD to arm64 & have an office next to his).

I also have issues with the fact that many parts of the ISA are optional. A general purpose OS needs some common level of support by the hardware. If it can't rely on some features it makes it difficult to support the different SoCs.

IIRC there are ISA changes that need to be finalised before the support goes upstream. The support is already developed and available on github though.

Intel has the best documented GPU.

> I really hope someone manages to ship an open source handheld one day.

100% FLOSS mobile phones are available for sale since 2007 :https://en.wikipedia.org/wiki/Openmoko

The current revision (GTA04A5, ARM Cortex A8 1GHz) can be ordered at http://shop.goldelico.com/wiki.php?page=GTA04

The Nokia N950 was also very open https://en.wikipedia.org/wiki/Nokia_N950

There are also other also-run phones: https://en.wikipedia.org/wiki/List_of_open-source_mobile_pho...

Cortex A8 is a proprietary design, no? How is that open source?

This is a path that must be followed with great attention. Giving up on low-end, entry-level parts often looks great in short term results but may doom the company in the end.

One of the reasons we deploy to x86 servers now rather then SPARC, POWER or MIPS is that we develop on x86. When Sun and IBM chose to exit the low-end space and pursue only large clients, they doomed their platforms to the legacy of those clients who already had their systems running on their platforms.

How many fresh applications are designed with SPARC or POWER in mind? How many are designed for zSeries mainframes?

I think that's a misunderstanding.

Atom chips have always been crippled to keep them from cannibalizing more expensive chips.

Skylake is a fine tablet chip, in fact, that's really what Skylake is good for. They are probably producing them in high enough numbers now that they can give up on Atom.

Atom isn't crippled so much as it's developed from a very simple (small, low-power, in-order execution) baseline and so hasn't had the time to reach the level of performance maturity and sophistication that the "big cores" (Haswell, Skylake, etc) have. If Intel had the ability to get Skylake performance in an Atom power envelope, they would.

Gotta love an article that leads with sunk cost fallacy. Anyway, axing Atom doesn't mean the end of Intel targeting tablets. Core M is fitting into fanless tablets thinner than an iPad.

Separately, the title is misleading when it says "Intel cuts Atom chips". While "cuts" may be technically correct, Intel isn't giving up on Atom, just canceling the smartphone SoC's that were planned to use it. Silvermont (the basis of the soon to come out new Knights Landing Xeon Phi) and (so far as I know) it's successor Goldmont are still going forward as scheduled. From the (significantly better) AnandTech article titled "Intel's Changing Future: Smartphone SoCs Broxton & SoFIA Officially Cancelled":

Also not discussed in greater detail is Intel's future plans for their overall Atom lineup. With Apollo Lake announced just earlier this month, it's clear that Intel's Atom efforts have not been cancelled entirely. We will still see the new 14nm Goldmont cores appear in low-cost PCs under Apollo Lake, most likely in several 11-to-13 inch high volume devices.

http://www.anandtech.com/show/10288/intel-broxton-sofia-smar...

Anandtech coverage: http://www.anandtech.com/show/10288/intel-broxton-sofia-smar...

Could not they just use their superior manufacturing capability to make e.g. 24 core arm that use less power then competition ? If not for them, it will atleast bite the market share from competitor. This is quite analogous to "Nokia should have embraced android much ago."

Shamelessly stolen from the comments at http://anandtech.com/show/10288/intel-broxton-sofia-smartpho...

"The short answer: margins.

The long answer: Intel could absolutely build a killer smartphone SoC on the 14nm process. However it would need to be Core based; the Atom CPU line isn't performant enough. They may need to go back to Ivy Bridge or strip down Skylake a bit to get there, but they could do it.

The issue is that smartphone SoCs have very low margins. MediaTek, Rockchip, etc are crushing the market. Atom is designed for margins, not performance. If Intel built a winning Core smartphone SoC, then they'd have to sell it for significantly lower margins than what their other Core parts sell for. And that in turn would put pressure on the rest of their chip stack; why should an OEM pay $250 for Core-M when you could get a similar smartphone SoC for $30?

For what it's worth Apple faces a similar challenge. But being vertically integrated means they don't have to share with others, and the total profit off of iPhones/iPads can offset the higher costs involved in developing and fabbing the A-series SoCs. However Intel doesn't have that luxury, and Asus would never let Intel have all of the profit.

Smartphone SoCs are a race to the bottom. And Intel as a business has nothing to gain from taking part in that race. The winner is the guy who got badly hurt but didn't die; a Pyrrhic victory."

So why haven't we seen the same story apply for the chromebook market(which btw is 10% of laptops) ? It has open and fair competition between Atoms and Arm's.

10% of laptops where?

Beside the HN quotes that everyone in US is using them, I never saw a single person using a Chromebook in Europe.

Surely I see them on display in German stores, usually just one, being down priced every week until it finally sells. Take a couple of months and another one eventually shows up again on the same display spot.

On the street, trains, coffeshops, airports, never spotted one.

They're said to have a big fraction of the US K-12 education market, where Chrome OS is a good fit, and they're cheaper than full service x86 laptops and much cheaper than Apple laptops.

Sure, they're "said to". But is there a reliable figure somewhere about usage share of that segment?

but this happens time and time again in business, where business's are so afraid to cannibalize their own products that they miss the next big wave.

PCs were also a race to the bottom. I wonder what would happen if Intel or AMD bought Compaq or another PC manufacturer maybe in the 1990s and dominated the PC market by the mid-2000s.

Intel probably wouldn't have been able to sustain the DNA required to also be successful in that market, and they would have had severe antitrust problems. AMD would have had the same first problem (but probably worse since they've historically had much poorer management), and would have likely lost a lot of their business selling to other PC manufacturers they'd now be directly competing with.

Nokia's biggest mistake was embracing Android. They had a killer OS with MeeGo, running on a flagship phone, the N9. It could've toppled Android at the time, but sadly, management smarts weren't equal to those of their engineers.

So perhaps Intel is in the process of making an analogous mistake...

Nokia's biggest mistake were the internal politics between Symbian, Maemo and to MeeGo transition (Gtk -> Qt).

When the first Maemo prototypes were shown to us, many asked why a radio modem wasn't part of it.

fta: > but Intel failed to unseat market leader ARM.

Intel should, if anything, license ARM designs and make the best slices of silicon out there for ARM. What I'm saying is, talking about unseating is the wrong way to think about ARM.

It wouldn't be smart of ARM to help Intel and harm it's own ecosystem and especially the chip fabs,because once they are weakend enough, Intel could return to using it's own cores.

Intel is an ARM architecture licensee.

I believe they sold that license about the time they went for Itanic.

Ah, I should have checked. I did read through the ARM architecture page† on Wikipedia and it does not mention that fact. It does say,

"In 1994, Acorn used the ARM610 as the main central processing unit (CPU) in their RiscPC computers. DEC licensed the ARM6 architecture and produced the StrongARM. At 233 MHz, this CPU drew only one watt (newer versions draw far less). This work was later passed to Intel as a part of a lawsuit settlement, and Intel took the opportunity to supplement their i960 line with the StrongARM. Intel later developed its own high performance implementation named XScale, which it has since sold to Marvell."

I'm not sure how much sense it makes for Intel to be both an ARM licensee and develop Atom.

† https://en.wikipedia.org/wiki/ARM_architecture#Core_licence

Wow, this is really surprising. Atom chips were steadily increasing their value from generation to generation and were within spitting distance of ARM in terms of mojo / watt and power consumption. Intel's process advantage should really let them dominate that market.

I wonder if this implies we'll see ARM chips built in Intel fabs?

Probably not the type of chip you meant, but as an example here's an ARM CPU on the Intel 14nm process: https://www.altera.com/products/soc/portfolio/stratix-10-soc...

I don't necessarily think this is a bad decision. Intel currently dominates high performance computing and data center chips, and if they lose that edge they will really lose.

Maybe it makes sense for them to instead double down on that and try to maintain their lead in those areas. Mobile markets are high volume but lower margin.

It seems like with Intel's superior process and production scale they could easily enter into the ARM market. Is it just not profitable enough for them to do this? Or are their fabs too committed to producing x86 chips? Missing out on mobile entirely seems like a pretty big strategic mistake.

Intel bought DEC's ARM business in 1997, tried doing ARM, and then gave up.

https://en.wikipedia.org/wiki/StrongARM

To be fair, it was a different world (almost 20 years ago... I feel old).

It doesn't matter anyway -- after several years concentrating their immense muscle on battery savings and integrated graphics, they have caught up a lot. CoreM / Skylake can fit most use-cases where ARM stuff would be called for, at least in the high-margin market.

This is a little puzzling given their 5G investment at mwc - https://newsroom.intel.com/news-releases/intel-accelerates-p...

So is Edison still on?

That or an article linked in a side bar said they were still chasing the IoT.

Edison is based on Atom?

https://en.m.wikipedia.org/wiki/Intel_Edison

No, the Quark SoCs are 486ish cores, not Atom.

Edison has both Atom and Quark cores.

The Edison has both an Atom running Linux and a Quark running an RTOS

I have my doubts on the future of the Edison. I was going to use it in a project but ended up giving up and switching to ARM after so many bugs went unresolved across multiple platform releases. The power consumption vs performance is incredible, but it just feels like they're not taking it seriously beyond single quantity orders for individual makers

Dumb question ... Broxton = Atom? So Atom will be no more?

No. The Atom core's going to be used pretty heavily, just not as a tablet/phone core. It's just been moved to pretty much solely the entry level laptop/desktop market.

Isn't the HoloLens using an Intel Atom processor?

Isn't Core M a tablet chip?

Reply on news.ycombinator.com