Intel loses its ARM wrestling match, kicks out Atom mobe chips Intel has thrown in the towel on smartphone processors after losing round after round against the ARM architecture – the dominant brains of the mobile world. Intel today scrapped the development of its Atom processor codenamed Broxton, which was aimed at powering high-end smartphones and tablets. Broxton was expected to …
Don't bank on it. This hasn't exactly come as a big surprise to the financial markets who pour over the numbers so the share price already has this built in.
On a side note I do miss Tim Worstall writing for the Reg and educating us all about such things.
The share price change is often a reflection of financial analysts understanding which is quite often simplistic. ARM has increased its profits when the world read about Apple selling less so the ARM share price went down. The financial analysts couldn't see that ARM are selling more licenses to other companies so Apple shrinking was not a big deal. Also, ARM is selling licenses on more complex processor designs for a greater profit margin.
This is the kind of news that the financial analysts see as a green light (which in this case they should do!).
Intel leaving mobile makes no difference to ARM as Intel had a microscopic market share.
Intel leaving mobile makes a chunky difference to Intel as they are not wasting money developing expensive low margin chips that hardly sell, and displace a sale of a high margin product when they do.
"someone somewhere in Whitehall is working on a plan that will screw with it"
I thought developing something to the point where it was nearly production ready and then canning it without reaping the rewards because it had cost too much already was a British disease. But no, here's Intel doing it It may well be that they couldn't see production bringing in enough returns to justify the money they've spent on development but then zero returns doesn't do it either.
I suppose the reason is that they've got to cut employee numbers RIGHT NOW to meet unrealistic analysts' expectations. Analysts, of course, don't realise or care that it's the employees who create the returns in the long run. So we'll see yet another tech company hollowed out due to short-termism.
The best thing that financial regulation could do would be to forbid the publication of quarterly results, maybe even the publication of results at less than two year intervals, just to get these idiots from breathing down managements' necks.
"The best thing that financial regulation could do would be to forbid the publication of quarterly results, maybe even the publication of results at less than two year intervals, just to get these idiots from breathing down managements' necks."
That'll never happen. A company can go from boom to bust in less time, so investors insist on more frequent reports or they'll sell out. That's the cardinal rule of corporate ownership: never tick off the investors.
"The best thing that financial regulation could do would be to forbid the publication of quarterly results,."
That'll never happen
Already has in London. Listing rules used to require a minimum of half yearly financials, and at least a trading update every quarter, the quarterly updates are now entirely optional. Some companies choose to do this, but there's no requirement, and a company can be compliant with two sets of results a year.
AC is correct that this reduces visibility, and some companies (including one I worked for) can go from boom to bust in less than six months. But quarterly reporting didn't make that any more obvious. If management are clear and communicate well, I don't think anybody will sell out just because there's no quarterly results. If they aren't clear and don't communicate well, would you then trust their more frequent updates?
Please Doctor Syntax and the rest of you who commented on Orlowskis story about "Why? Apparently, “for every consumer to have a multitude of username and password combinations is not only inconvenient but becomes a security risk.”.
Should not a sentence like "the nation’s entire population of ten men and a donkey" reveal something to you.
Did you not understand that that whole shit, for gullible Brits, was all Orlowski and nothing else.
Shame on you.
"The best thing that financial regulation could do would be to forbid the publication of quarterly results, maybe even the publication of results at less than two year intervals, just to get these idiots from breathing down managements' necks."
Nice thought, but in practice that simply make insider trading much more attractive, and it would also make the markets a bit less efficient because folks can't make rational decisions on zero data. :)
Windows Phone is (was already) Dead. Universal Apps? Not so important now, so a lot of dead wood already in Windows 10, making Windows 7 look the leaner choice, trouble is, MS has now purposely 'borked' Windows Update for Windows 7, making it unusable if you reinstall, restore from an Image. Manually installing KB3102810 worked for a while, not now.
MS literally seem to be putting a hammer to Windows 7SP1 to kill it.
Microsoft have always biased themselves towards Intel, this time it appears Microsoft is having the rug pulled from under them, but maybe Intel saw the writing on the wall, and pulled the plug first.
This stems back to 2012, Microsoft failing to upgrade WinPhone 7 users to Windows 8 Mobile, then getting caught trying to do the same with Windows 10 Mobile, getting asked the question directly and stating 'No - everyone will get a free upgrade' , then realising this wasn't technically possible.
It looks like Joe Belfiore took the can for that one.
I'm sure someone will describe it as 'growing pains' though its more getting old pains.
What a Mess.
For IoTs you don't need a raspberry pi - if there is a powercut you don't want to have to rebuild the sd card (I know this from bitter experience). More useful are the Arduino/teensy boards. For example the recent Arduino mkr1000 has WiFi a reasonable (for this job) ARM processor and even a cryptography accelerator chip. There is even an Arduino with the Intel Quark chip onboard, however Intel have not released the operating system that powers it so you can't get at the internals.
For IoTs you don't need a raspberry pi
Of course you don't but the RPi is creating something a bit like the old PC ISA for IOT (sorry for the abbreviations). ISA (industry standard architecture) is important because it reduces costs and risks.
People and, increasingly, companies are prototyping embedded devices with RPis, Arduinos and the like safe in the knowledge that they should be able to maintain software and hardware down the line.
A couple of years ago a company like Intel might have been able to own this space by providing the ISA. Now I think they will have to work with whatever is being established out there. There's still a huge opportunity for them: what will follow the RPi 3 now that it looks like Broadcom has lost interest?
IoT is not even a thing, even conceptually it is not developed yet. Hacking something up at home does not an IoT make, it's still just a mini PC with Ethernet, a Gaston Lagaffe innovation. How the hell are you guys arguing about whether RasPi or Arduino is best for it?
Did I mention that it isn't even certain whether anyone of the "consumer demographic" will even shell out for IoT stuff, whatever it is when it will be something?
>IoT is not even a thing, even conceptually it is not developed yet.
Actually it is. There are several companies that offer complete systems that interface your phone to many different things such as lights, heating, garage doors, Video cameras, and do so over the internet. And they have done so for many years.
The hype about IoT is where software companies, such as Microsoft, want to grab the action, and the revenue, and promote their IoT efforts linked to their services. As they don't make hardware they are trying to ride on the back of RPi and others and trying to get users and developers to write software that will be tied to MS services.
You are wrong that 'it doesn't exist', it is just that MS tries to ignore what does exist and will promote whatever they do as 'the standard' and 'innovative' when it is neither (yet again).
"Did I mention that it isn't even certain whether anyone of the "consumer demographic" will even shell out for IoT stuff, whatever it is when it will be something?"
That's assuming we will be given a choice. Odds are, we won't, at least outside the 'premium' range.
People comparing arduinos and Pis again. It's really daft.
If you want a device that will run linux with Apache/NGINX/node.js/MySQL etc, or you want to see a HDMI desktop with LXDE or run a CSI port camera or run a choice of different OSes then you need a Raspberry Pi.
If you just want a controller with low power standby that can boot straight into its sketch then you need Arduino.
If you want both, then they work nicely together. You can even write and upload Arduino sketches from a Pi.
It's a bit like trying to decide if you want a car or a motorbike.
"Raspberry PI is a seed that will be flowering"
The Raspberry Pi is a Broadcom BCM2835 evaluation PCB that sold a lot because it has a stupid name. This time last year they has sold 5 million of them? Microchip (who are not even in the top 20 list of semiconductor manufacturers) sell 2.7 million microprocessors every day.
The Raspberry Pi is no more than a pimple on the arse of the global electronics industry.
>Microchip (who are not even in the top 20 list of semiconductor manufacturers) sell 2.7 million microprocessors every day.
And the 8bit PIC is still there biggest seller (both revenue and profit wise I believe). The rasberry pi is much more expensive, complicated and more powerful than the micro controllers that go in most products who often just need things like temperature control. Also with their recent purchase of Atmel I am pretty sure they are top 20 now. Still the margins are tiny as are the unit costs in this market so yes volume is king. That said Microchip has been turning continuous profit for 20+ years so while not perhaps a sexy market it is a legitimate one.
There's the problem in a nutshell. Intel has never been able to develop satisfactory low power x86, let alone ultra low power. Their best efforts at low-power have produced exceptionally low performance as well.
This might have something to do with their penchant for trying to solve issues by increasing complexity (much like a certain software company they have close ties with, which coincidentally is also having trouble getting its software to run efficiently on smaller devices.) In the low-power world, complexity kills.
Don't get me wrong -- Intel's x86 team does some amazing engineering, but its core design philosophy is antithetical to the simple requirements of low-power systems.
>This might have something to do with their penchant for trying to solve issues by increasing complexity
It has more to do with x86 being one of the worst instruction sets you can imagine for low power use (even if efficiently emulated). I can agree though complexity is what kills. Intel has tried to move away from x86 in the past but being a victim of its success its a chain that may eventually drag them down.
I never quite got why anyone would have wanted an Atom (x86), powered Samsung Device over say a Exnos, or Qualcomm (Both ARM), versions. When everyone, and their dogs were probably developing against ARM,and not x86.
Then again all is not totally lost, assuming Google really do mean to converge Android into its Chrome OS. Perhaps then it'll work. As for small ultra low power Workstations with perhaps just enough grunt to run some less demanding games? I'd be all for that in a second.
I never quite got why anyone would have wanted an Atom (x86), powered Samsung Device
Consumers really couldn't care less about the chip but manufacturers will generally buy from the cheapest supplier. But even when Intel was effectively giving the chips away manufacturers weren't very interested.
Losing mobile is seen by many, including Intel, as the beginning of the end of x86 dominance. Once x86 compatibility is not seen as essential, Intel can expect to see market share and margins elsewhere start to plummet.
This was an acceptable Android tablet powered by a quad core Atom - It sold quite well as it was cheap (£129) for a device sold in the UK with a 1 year warranty, decent screen and reasonable speakers.
It had a few disadvantages - it sucked power in standby as the Atom chip was nothing like as frugal with power as an ARM chip.
(The low price was in part due to Tesco including non-removable Tesco shopping apps.)
Still slightly ahead of its time but a lot closer with advances in hardware and software than when the Motorola Atrix was launched. we are getting there slowly but the lack of a future low power X86 chip set family will put a spanner in development of future hardware convergence possibilities.
I was hoping for a low power x86 chip in a future ASUS Padfone.
Needed :
large battery, 5,000mAh+
5" MAX or smaller screen
x86 low power multi core chip
4-8 GB ram
128-256 GB memory (samsung have demonstrated a 256 chip)
Micro SD slot
Dual USB C connectors with (Display port support)
WIFI 802.11ac
4G - LTE
Choice of Tablet docks 10" and 13,1" and 15.6" full HD panel large battery, with pogo or similar connectors for keyboard, kickstand (like surface) rather than weight in keyboard (like transformer pad). (all docks to fit the same phone model)
desktop dock with multiple USB ports, Monitor ports (through USB C ?), RJ45 network, (possible version to allow internal SSD or HDD) . could even do a desk dock with a phone handset. Convergence :-D
Nothing too radical in the Specs above (same hardware for all software options):
options of the software below is a bit more of a stretch (unlocked boot loader. but I don't see that happening)
FULL Windows 10 (x64) with dialler to allow it to be used as a phone.
FULL Ubuntu or other main stream Linux Distro (x64) with dialler to allow it to be used as a phone.
Android. N or a version like Phoenix or Remix but with dialler.
with these specs i would have 99% of my daily communication and computing needs covered and for the last 1% i could fire up a VDI or RDP in to a server.
It could even be used as a BYOD device for connecting to a corporate network using VDI or Citrix connector. and an app to connect it to the corporate PBX when docked.
I realise that this would not be a main stream high volume device. but with standardised docks there could be a family of cheaper models for people who just want phone and tablet. the lower models could even run ARM chips for cost and performance and software compatibility.
Nothing too radical in the Specs above (same hardware for all software options)
No, but an awful lot to put in a small package. Development costs would be large and the market small leading to very high unit costs. Mobile devices have thrived on using as much commodity hardware as possible and building to a price.
Physical ports are nice but also more work which is why we're seeing so much emphasis on software solutions: miracast, continuum. Put your phone on a charging mat next to a screen and you're docked.
There's really no substitute for wires. Physics dictates this. Plus when it comes to storage, the lack of a Micro SD slot is a deal-breaker to many. It certainly is to me.
And I wouldn't mind a step back towards portable bricks if it means I can run the thing flat out and not have to hunt the boonies after an hour or two for a charger or battery pack when I'm forced to travel light. Designers really need to ask people if they're willing to trade in weight for working life without having to resort to bodges.
Here's the equation: Neither Android nor Ubuntu care a wit about X86 compatibility, even a modest power saving is worth infinitely more, and there is a practical certainty that 86x prices will be jacked to the moon by the one source supplier given any significant success. Only Windows care about x86 now, and the Windows on mobile is looking bleak indeed, never mind the present. On server, desktop and high end laptop, nobody wants a crippled x86. So the future market for Atom rounds to zero, and no tears shed.
The fact that it was based on X86 says it all really. An ancient cpu arch that can trace it's roots back to the 1970's. It would have been put out of it's misery years ago, but for the monopoly postion held in the PC and server markets. Not a nice arch for hardware designers, nor for software engineers either, when compared to cleaner and more orthoganal architectures such as arm. Intel should have started a complete redesign with sharp pencil and clean sheet of paper years ago, but gee, so much locked in profit from X86...
"Intel should have started a complete redesign with sharp pencil and clean sheet of paper years ago"
The only reason for my specifications insisting on an Intel x86 chip is the software comparability that this gives.
If all the standard win32 programs were rewritten so as to be fully featured universal apps or coded to run on ARM or a new "modern" architecture natively then there would be NO need for Intel x86 chips in Mobile but with current software and legacy programs it is still needed and will be for the foreseeable future. I want to install and run programs locally not be tied to a CLOUD service.
> But a phone which does Continuum for x86 Win32 programs would have sold truckloads,
Which is one reason why Microsoft does not want to do that. MS wants you to buy a PC _and_ an XBox _and_ a laptop _and_ a phone, plus several IoT devices - as long as they all run Windows and buy from their app stores.
Buying just one device that runs Win32 programs (mostly not by Microsoft) is not an option in MS's revenue plans.
Note that Continuum is currently ARM and only runs UWPs which must be bought from MS store (with a 30% cut for MS).
"Intel should have started a complete redesign with sharp pencil and clean sheet of paper years ago"
You mean like Itanium was meant to be? Intel's track record with introducing new architectures is abysmal. i860 was also a flop. Intel just doesn't seem to be that good at anything other than pushing x86 on the desktop.
Zenfone 2 was a great example for Intel phones can be good with unique characteristics such as huge amount of RAM (4GB) and yet, lack of software quality control added to massive draw of power in standby. I know since I own one as my main phone.
Asus shipped everything ARM based since Zenfone 2, using same brand. Zenfone Selfie, Zoom. It seems they were fed up with the bad feedback they get from users.
"...fuel the virtuous cycle..."
That vomit-raising phrase reads as if it was written in Chinese, and then run through Google Translate. Which would be perfectly acceptable.
But if that was actually drafted and approved in native English, then Intel is doomed. It reveals serious brain-rot.
They apparently failed to understand that they are tied to the PC-platform. People want x86 because it comes with a whole ecosystem of hardware that's well standardized. You have a wide variety of operating systems available, and it doesn't matter if your PC was made by company A or B.
This could have lead to a new class of devices, connected Palmtops. Essentially spiritual successors of the Nokia Communicators, but with x86/PC hardware.
However Intel promoted bog standard Android devices. Exactly the kind of device Intel has a great disadvantage at, since Android is ARM country. Many applications come with their own ARM binaries to actually do stuff. Those need to be emulated. In effect the user will have a device which looks and feels precisely the same as cheaper ARM-based competitors. Having closed boot loaders also eliminates all the remaining advantages.
It would be nice to see Intel with chips in the microcontroller + WiFi space where I think the IOT hometurf is really at. But somehow I find it hard to believe, as Intel has a prodigious past in burning just these kind of bridges. Intel does not have many microcontrollers to offer compared with the compition, and all things ARM (XScale) has been ditched long time a go.
Intel likes the idea of IoT, but has thus far completely failed to understand it. The Edison fiasco is the proof.
"Hey, let's create an IoT system. It'll have a processor - what've we got? Oh an Atom will do - and sensors, and an internet connection. What more do you want?"
Well, let's have a think. A development system that doesn't cost a large multiple of the price of a MicroBit or a PiZero might be a good start. A board that doesn't have a grossly overspecced processor, most of whose cycles will be idle. A processor that doesn't chew battery like it's going out of fashion. Seriously - the baseline Edison board probably out-specs what's in your PVR! A gig of RAM? Four of flash? This is basically a system-on-chip version of a PC, not an IoT controller.
Basically, Intel "gets" IoT, like a gorilla "gets" needlepoint.
but then I'm part of a generation that had to wrestle with the Intel/MS curse where MS told us programming was easy (when it never was) and then made it bloody near impossible to find out what was wrong with the code as it fucked about in the wrong segment. OK you dont get that nowadays but I'd bet the world would be a much better place if we hadn't had it at all.
"then made it bloody near impossible to find out what was wrong with the code as it fucked about in the wrong segment"
If 64K was enough for everyone I would have stuck with a 6502 and paging. That was one of the reasons I didn't rate the "PC" until the 386 came along, seemed like a huge backward step having been coding in C on an Amiga. :)
Hilarious. 5G doesn't really exist. Unlike 2, 3g & 4G it's more about a flexible infrastructure and higher level protocols than the "air interface" or Modem Chips.
Intel has already "lost" the IoT market, which is mostly bonkers stuff no-one wants. How many TV,s Routers, WiFi points, setboxes etc use Intel?
The power issues when there is no screen and GPU even more favour ARM in embedded computers. MicroChip bought AVR, they and NXP produce 50cents and cheaper ARM based SoC. I can't see x86 or Intel in that market.
Memory chips :) That's what they did before the 4004!
> the end of Microsoft's mobile phone business?
It was never even the Beginning of the Windows Mobile Phone Business because "Windows 10 Mobile"/"Windows Phone" is (or was, as the consensus is that MS is giving up on WinPho) running on ARM. The only thing connecting "Windows 10" and "Windows 10 Mobile" is marketing whalesong. See also: Windows 8.1 devices (all use Snapdragon, i.e. ARM) and Windows 10 Mobile devices (all use Snapdragon, Rockchip or Mediatek i.e. ARM)
The Surface has has given up on non-Intel some time ago for reasons unclear, so there may be problems in that particular segment (Which according to my entirely unscientific survey exclusively consists of random C-level personnel receiving presents from the Microsoft Country Manager. Said personnel then proceeds to bother IT support because "my surface is not working correctly", leading to revenue sinkholes)
As for "cheap windows tablets" that are usable? Err.... where?
Companies building in a new field (like IoT) need confidence that their suppliers will be around for the long haul. Having see Intel abandon the phone and tablet market will make those companies reluctant to trust Intel to stay in the new markets - especially if the unit cost has to be low.
Intel have just effectively said "We only supply X86 chips to PC and server makers. We will not be in any other business EVER."
"We only supply X86 chips to PC and server makers. We will not be in any other business EVER."
I'll take this opportunity to say Good.
The cost in compromise for convergence is way too steep in both directions.
I don't want to run Crysis on my phone and I don't play Angry-Birds on my tower.
So if Intel want to pull their entire mobile workforce and chain them all to the performance x86 processor wheel then more power to 'em.
Not a RISC vs CISC argument here. Just saying that while the whole ARM community can focus on performance per watt, I'd personally prefer Intel to focus on performance per tick. I can always get a bigger heat-sink.
To compete in the phone market Intel needed to produce processors with 8 to 10 cores, at 64 bits, running at least 2 GHz and do that at a cheap price.
So who the hell would buy they Core i7 desktop processors, if the mobile chip offers everything for a fraction of the price? They were just undercutting their own much more expensive desktop processors.
What it means though, is Intel is going to focus on profiting from the declining PC market.
But then in a few years time, a PC won't be Windows on Intel, it will be Android on ARM. Because Android has more apps, more power, less price, more choice.
So what then Intel?
You have to slim down and get in the fight because the fight is coming to your home market.
"So what then Intel?"
Focus on performance. Because there's a physical limitation regarding computational efficiency. If raw power is needed for things like media or complex graphical manipulation (the kind GPUs aren't set up to do), then Intel's still the kind because their silicon is geared for this. It may run hot, but so does a V8.
I just bought a Dell Venue 8 7000 tablet that uses an Atom Z3580 chip (64 bit/2.3 GHz/quad core). It runs Android flawlessly and gets a good 10 hours per charge battery life. So I don't think this has much to do with how good or bad the Atom chip is as much as it shows just how crowded the mobile processor market has become. Also with low-power Skylake CPUs and their successors on the way, there probably is no longer a need for the Atom line of chips.
Dear Elreg, we have a problem.
In this article: http://www.theregister.co.uk/2016/04/29/eu_login_youtube_national_id_card/
Andrew Orlowski writes this;
"Why? Apparently, “for every consumer to have a multitude of username and password combinations is not only inconvenient but becomes a security risk.”
My problem with that is that apparently, judging from the comments by Brits, some, too many, actually believe some official EU member or individual in the EU actually wrote or thought or had such an opinion.
A hook planted by Orlowski, and wow did it sink.
My first response to that dumb sentence by Orlowski was "I must admit I would like to see the source to that sentence." and that vent through, still there. (no down votes or up votes, surprisingly).
My admittedly more straight forward comments like "Hang on, was that dumb sentence the work of Orlowski. and "Re "the nation’s entire population of ten men and a donkey", this article is written by a Orlowsky. to be fair.", were rejected-
And right now this comment "Please Folks, why can't you see that nobody but Orlowski wrote this "Why? Apparently, “for every consumer to have a multitude of username and password combinations is not only inconvenient but becomes a security risk.” is still, since 9 hrs. waiting moderation, Give the man some help.
Meanwhile, before dealing with the northern donkeys. In the USA Trump goes on telling twats that everybody is ripping the USA in trade while the rest of the world (perhaps even Trump, not sure) knows each and ever trade agreement was written in the USA in the interest of some and forced down the throat even in the Capitol.
I expect more from you Brits even when Brexit distracts and divides you.
Then for "the nation’s entire population of ten men and a donkey", while not an expert on Estonia I cannot relate donkey to much anywhere in Northern Europe, hardly England either but wait, perhaps Orlowski, Poland/Russia, that fits indeed.
I remember a Russian who went to great lengths in telling me how much better Moscow is than my town, as there are so much more people. And I said - wow, fantastic, I newer knew, you don't say, wow wow, Mexico City must be just absolutely fantastic. Not a dumb guy, his face revealed he got it.
Now there is humour of course, when I read an American account of Brits - leftist faggots with rubbish teeth and rundown shoes, I take it with a smile. But there is a difference, that shit is written by commentards and not by people who write, like you, for a publication like ElReg.
I think ElReg should have a serious discussion with you Andew Orlowski.
It's Orlowski's playground, and I don't agree with all the moderation he has done either, but at present there is still enough useful+enjoyable articles&comments at El Reg to outweigh the burden of additional self-censorship when making a post on Orlowski's (or Trevor's) patch. It does diminish the credibility of El Reg in my eyes, but I don't think that should be cause any sleepless nights at Vulture Towers. :)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
