Hello 'WOS': Windows on Arm now has a price Microsoft's long journey away from Intel reached escape velocity this week, as the first traditional laptop machine with Qualcomm's Arm processor was revealed by Lenovo, in the shape of the Yoga C630 WOS. "WOS" stands for "Windows on Snapdragon", how Qualcomm prefers to call "Windows on Arm", and since it's Qualcomm Inside, …
Anything that steps away from a monoculture is a good thing in my books.
Chipzillla desperately needs more competition if for nothing else but to motivate them to be better, Sadly I don't think AMD has what it takes to be any other than an occasional annoyance.
Give the rise and collapse of the Bulldozer architecture I am forced to anticipate the pattern to repeat itself with Rizen.
I'm not sure I follow. Bulldozer and related microarchitectures were a flop because both performance and power efficiency were beyond comparison (except under very specific workloads). Ryzen is a completely different case. As for ARM, I'm a little sceptical.
In the past, performance and compatibility weren't there. As those improve, the gap in power consumption dwindles. I fear that in the end, if it does become an option, there will be no particular reason to go with ARM except just for the sake of getting something that's not x86. Which is cool in a way, but it would be even better if we could get a practical advantage. And they do need to go the full distance on performance and compatibility, as market has already rejected products such as Windows RT (power efficient, cheap, not really compatible) and Transmeta Crusoe and the likes (low power consumption, low performance, laptops weren't exactly cheap).
I am more enthusiastic for ARM. The reason Intel is a power hog is because the chip, internally, is a RISC. It emulates the IA instruction set in software, i.e. microcode. That is clearly not as efficient as simply running RISC instruction set. This emulation is why all the low-power Intel attempts have failed, and the native RISC instruction set is why ARM is so doing well at the low-power end. But this does not mean that ARM needs to be barred entry to the high power computing - quite the opposite, actually, because it turns out that at the high end, the power efficiency is also very important. For an example, see The Next Platform. The only reason for Intel to stay at the position it is right now is its instruction set, which the exact same thing killing its (energy) performance.
... actually, I need to correct myself. The only reason for Intel architecture to stay at the position it is right now is its instruction set
Intel (firm) knows it very well, which is why it is investing in alternatives - for example (also The Next Platform).
" I am more enthusiastic for ARM. The reason Intel is a power hog is because the chip, internally, is a RISC. It emulates the IA instruction set in software, i.e. microcode. That is clearly not as efficient as simply running RISC instruction set. "
Surely Internally x86 is now RISC. It takes a CISC and converts it to a stream of micro instructions. It's a bit like compression in that sense then, meaning lower bandwidth on the external bus. The conversion to micro instructions would seem to be a fairly trivial task in terms of die size and power.
It takes a LOT of power and electronics to do the translation FAST.
It isn't a simple table lookup translation. Closer to being a compiler... The only RISC instructions passed on are those relative to the specific operation. It isn't the same sequence for different operands.
I use ARM every day in my development environment. I work almost entirely on Raspberry Pi these days.
I would profit greatly from a Windows laptop running on ARM with Raspbian running in WSL.
That said, I already get 12 hours battery life on my Surface Book 2 for watching videos and I also have. Core i7 with 16GB RAM and a GTX 1060.
Nokia basically destroyed their entire telephone business by shipping underpowered machines with too little RAM because they actually believed batter life was why people bought phones. They bragged non-stop about how Symbian didn’t need 200Mhz CPUs and 32MB of RAM and yet, the web did and when iPhone came out and was a CPU, Memory and battery whore, people dumped Nokia like the piece of crap it was. The switch to Windows was just a final death throw.
After all these years, ARM advocates seem to think people give a crap about battery life and are willing to sacrifice all else... like compatibility or usability just so they can not run what they want or be able to use it just because they can’t carry a small charger with them. I honestly believe that until ARM laptops are down to $399 or less and deliver always online Core i5 performance, they won’t sell more than a handful of laptops.
Let’s also consider that no company shipping Qualcomm laptops are making a real effort at it. They’re building them just in case someone shows interest. But really, the mass market doesn’t have a clue what this is or why it matters and for that much money, there are far more impressive options.
And oh... connectivity. If always connected was really a core business for Microsoft, why is it that my 2018 model Surface Book 2 15” Computer packs LTE?
"The reason Intel is a power hog is because the chip, internally, is a RISC. It emulates the IA instruction set in software, i.e. microcode. That is clearly not as efficient as simply running RISC instruction set. "
My understanding is that the core is RISC-like, and that the Hardware Control Unit processes most of the instruction set and the microcode is only used when the instruction is not in the CU, or has been patched.
The microcode is obviously not as efficient as running through the CU, but I don't think it would be fair to claim that simply because there is a RISC-like core, that there is an efficiency loss. (ignoring of course the microcode updates for Meltdown/Spectre...)
Bulldozer and related microarchitectures were a flop because both performance and power efficiency were beyond comparison (except under very specific workloads).
I presume you mean that the performance was low at a given power consumption? You are possibly right, but you ignore other things. I have been using AMD almost exclusively in low to mid range builds (not that I do hundreds, but y'know) for a while now because for an equivalent amount of computing grunt, the AMD parts have for a long time been cheaper than the Intel parts, particularly when you consider integrated graphics and motherboard facilities.
For example, for a very long time a cheap AMD board would have 4 or 6 SATA3 ports while the equivalent Intel board might have two SATA3 and two more SATA2 ports if you were lucky. It was a similar story with USB3.
For general desktop use, light gaming and life-extension upgrades the A6, A8 and A10 processors (mainly the A8) have been excellent value for money over the last five or six years, though I note that Intel's pricing has moved in the last nine months or so, still, an A8 (two Bulldozer modules sold as "four core" with R7 graphics) is the same price as a two core non HT Celeron with UHD610 graphics.
Ryzen is now competing with Intel in the higher performance bracket, which is a place AMD haven't really been for a long time. I'm just setting up my first Ryzen system. Interesting times.
M.
Hopefully MS won't fluff it this time so it will be. In the end it's about the software.
They already have fluffed it ... it runs Windows 10 S -- the version that restricts you to only running stuff from the MS App store.
I like the reported battery life, but not at that price. Anyway: I would wait until someone reports that Linux Mint runs on it.
"Apparently Windows 10 S has an app in their store which flicks the switch and changes it to normal Windows 10"
but this thing doesn't have an Intel (or Intel-emulating) CPU so presumably it won't be able to run x86 and x84 native windows apps. So this while thing really boils down to nothing more than a reinvention of windows RT. Can someone explain how I'm wrong?
Theoretically, anything relatively legacy from win32 should run, as long as it can without doing any deep-system work. The APIs are compatible and the instructions are simulated. If what I saw months ago is still correct, 64-bit applications won't work, but with only 4GB of ram, that's not a really big deal.
I wonder how far the battery drops when you're running one of those. I presume quite a bit. I also wonder if the 25 hours of battery when watching local video is perhaps taking advantage of a GPU that is efficient. Usually, that kind of number sparks my exaggeration sensor. I would like to see real world numbers for activity that will actually happen: browsing internet, doing office work, spreadsheets with inefficient formulas courtesy of finance, and doing one of those with several tabs, mail client, and skype running in the background. Then I'll know what I predict.
> Apparently Windows 10 S has an app in their store which flicks the switch and changes it to normal Windows 10 and it's a free download.
That is apparently true for the Intel version of Windows 10S. That has no bearing on whether the ARM 10S has the same facility.
The Snapdragon 850 only has emulation for x86 and _not_ x86-64. As most software over the last decade has moved to 64bit then your favourite programs may not be available on this machine.
In fact the lack of 'normal' Win10 may be deliberate so that Microsoft can ensure that only software that will run under this emulation will be available. Otherwise there may be too many returns because the user tried loading software that could not run.
Anyway: I would wait until someone reports that Linux Mint runs on it.
I did this a couple of years ago with an Acer Aspire One cloudbook; they shipped with Windows 10 which it was far too low specced to run properly on at all, and so retailers were awash with returns which had basically been booted up and then sent straight back. I picked one up for less than half price, put Mint on it and it ran perfectly.
Picked myself up a returned Dell Inspiron 13 a couple of months back for ~£230. Usual price is > £500.
The laptop has a touch screen, tablet mode, full windows 10, 1TB drive, 4GB of memory. Absolutly nothing wrong with it at all. Did not look like it had even been turned on.
Installed Ubuntu on half the disk, as I need both OS's. I manly needed the laptop for testing software with an Intel GPU.
Ubuntu runs well. Windows not so well.
Chrome on both struggle after you have 5 or 6 tabs open.
Could do with at least 8GB of memory though!
Apparently, we thought wrong. I can get a monster battery pack for my slow-sipping Asus X202E (which I adore, runs Windows Server 2016 here as my lab machine) well below that cost, tossing in an upgraded laptop in the deal. Yes, I want to see alternative paths, non-x86 paths at that, but this? at this price?? No.
The chips are cheaper, but you can expect to pay a premium for the combination of low weight + good battery life. 25 hours for a 1.2 kg device is very impressive. People who want something light but with good endurance won't mind paying a bit. If the numbers do hold up then Microsoft has definitely stolen a march on Apple.
"you can expect to pay a premium for the combination of low weight + good battery life"
And thus marketing kills demand for a product which at half the price would be marching out the door to the tune of cannibalising sales of most of the rest of the models
Disruptive products come in at prices which put everything else on its ear. This doesn't and the fact that it's coming from an established maker is a lot of the reason why.
WRT "Always connected" I've been seeing laptops with 3G and LTE connectivity for years. What kills them utterly is the telco fees. After a couple of months of that only the most dedicated roamers keep the SIM installed and everyone else pulls 'em.
"WRT "Always connected" I've been seeing laptops with 3G and LTE connectivity for years. What kills them utterly is the telco fees."
Depends where you live. Here in NL 4G data is very reasonably priced, with roaming throughout the EU. On my Vodafone business contract I get unlimited calls in NL + EU + 15 GB data + 2 extra SIMs (for my laptop and mobile hotspot) for EUR 36 (exc. VAT) per month. Including roaming in Canada and the US would cost me EUR 29 extra. Haven't used WiFi for more than a decade.
https://www.vodafone.nl/zakelijk/shop/mobiel/abonnement/?icmp=nav-subs-simonly
There are also consumer contracts for about the same price with unlimited data.
I mean Intel and AMD have patents on their instruction set architecture.
Of course they could just emulate >15 year old CPUs which would fit the usecase of most coorporate Windows machines very well. After all few Windows-only Software actually uses anything that came out after 2000.
The original x86 patents expired a quite some time ago and they're the ones that matter. Licences may still be required for some of the extensions MMX, SIMD, etc. but those aren't likely to be such a problem.
Anyway, the real fun will come on these machines with Windows removed so that no emulation is required. I assume we can expect bootlockers designed to prevent this.
"After all few Windows-only Software actually uses anything that came out after 2000."
I wish this were true. Unfortunately, some *EXTREMELY* unwise software developers (for business applications) drank the Micro-shaft coolaid and either use C-pound, or (nearly as bad) ".Not" with C++.
If the application you use falls into this category, you're *B0NED*.
However, if the developers were SMART, they used Java [Oracle does this] or MFC/C++ *without* ".Not" and targeted XP or 7 [and not 10]. Yes, it's STILL possible to do that. And very, very wise.
/me points out that with a little effort, MFC applications can be modified to use wxWidgets to run on Mac or X11 systems. There's effort for sure, but it's not "that much" and worth doing. Then you can have a single code base for everything. Yes, _I_ do this.
Once business applications are commonly available for Linux and Mac, people will *STOP* "needing" Windows, and developers will have even MORE reason to make their applications run on non-windows OSs.
A few years back I got all the main Windows tools of the company I worked at the time runing under Wine. I even fixed a few issues with Wine to make sure it worked and swapped out DLLs on installation to be doubly sure.
However there were no additional sales because of this even though half our runtime sales were for Linux.
I did not regret doing this exercise. But afterwards I pushed development to use Qt for the tool frontends.
However with the SDK costs of Qt being so high compared to a VS licence and the new confusion sown by the new owners over the use of LPGL and the dropping of the runtime exception clause I struggle to convince people to use Qt.
The primary issue with picking technology is with Senior management. Invariaby they use Windows and always pick Windows technologies. Senior management in the UK does not usually have a software development background...
"Once business applications are commonly available for Linux and Mac, people will *STOP* "needing" Windows"
We still have coders insisting on producing stuff that _ONLY_ works on MSIE and who then whine that everybody has a windows box available, even android users.
Softbank, as far as I can read, is basically Chinese. I've got to admit, it's getting very hard to keep track of who owns who these days... https://www.zerohedge.com/news/2016-07-18/did-china-just-buy-most-important-company-world - https://www.design-reuse.com/news/44316/softbank-arm-china-stake.html
I think Intel cant afford ARM. I voted against the sellout to Softbank as I though ARM would be worth 4 or five times the offer price in a few years. Seeing their post buyout release of info on there NN software I thought I'd underestimated. Seeing their ML stuff I KNOW I underestimated.
Qualcomm is a preferred supplier for the US military so no sale without their approval.
Anyway, that ship has more or less sailed: there are too many ARM licencees, including Intel, out there now with the right to design their own ARM-based archs.
I don't see any particular need. Lots of iPhones out there with 1 to 3GB. Very few Windows applications would need anything like 1GB of working memory. Extra RAM has traditionally been used as a cache to compensate for slow magnetic disks but it makes very little difference if you go with a good fast solid-state drive.
It is a source of mystery to me why the world is still flooded with desperately slow PCs which take forever to boot up or even to launch an Office app, but I don't think RAM size is the limiting factor.
>It is a source of mystery to me why the world is still flooded with desperately slow PCs which take forever to boot up or even to launch an Office app, but I don't think RAM size is the limiting factor.
A lot of -most -- software still waits for events using busy/check loops (polling for data). Performance increases are entirely dependent on making the storage device or network as fast as possible, getting access latency to zero rather than designing around it. Add to this a lot of modern software uses system resources like malloc pools as if they're infinite with zero penalty for fragmentation and the inevitable software bloat and its small wonder that the systems take for ever to come up unless you've got a really fast SSD.
Some profligacy with resources is to be expected, they're not only relatively cheap but also its the result of the convenience versus efficiency tradeoff. The perennial problem we have -- and we've had it since the first PCs came out -- is that everything is flung against the stops, turned up to '11', so no matter how good your system is its never good enough.
martinusher - thanks, that sheds some light. I still come across laptops, even new ones in store, that don't seem any faster than the Gateway 2000 I had 20 years ago! Well, maybe a bit faster, but not the 20x you'd expect from the raw technical specs. Sad, really.
I've also come across "cloud" services that do a huge amount of work in browser scripts until they grind to a halt... and yes a higher-spec PC worked around the problem... sigh.
you're right.
I had the 'fun' of bringing back an old athlon xp machine back to life with 2GB memory. it had originally had only 128MB and I seem to remember it did all the office stuff just fine.
I couldnt even install anti virus because the installer intro (yes the flashy, zoomy, animated *intro* - not the actual software, but just the feckin intro), required sse2, which the cpu doesnt support.
standard office tasks havent changed in 20 years - why the bleedin F do computers need more and more resources to do the same workloads we did back then?
dont get me wrong, i'm not against advancements in technology, but i deplore the need to 'improve' software with shit it doesnt actually need, only to put demands on hardware that serve no practical purpose. this isn't progress.
No, Windows RT died purely because it was locked down to a (cr)app store. There was no other reason.
Install Linux ARM on it for example and the i.e Debian repository has almost every bit of software you will need.
To be honest, I know quite a few people who are happy with their Jailbroken Windows RT. Microsoft delayed the ARM laptop market by 5 years with this stupidity. That is a lot of wasted energy and thus environmental damage, thanks Microsh*ts.
I think it was a big mistake using Windows in the name of Windows RT. Compounded by launching Windows RT tablets at the same time as Windows Tablets and all the prelaunch talk was about the soon to be launched Windows tablets was that you would be able use all existing windows applications. So it became Q. What does RT mean? A. It's means it's Windows but you can't run Windows applications.
True enough but the difference is that ARM licenses out their designs and other companies can fabricate their own chips.
With Intel, Intel does everything. They have ultimate control.
Intel Management Engine and other creepy stuff like that will be a little harder (but not impossible I suppose) to slide into every fabricated ARM chip.
The point about Qualcomm having patents is fair. But I imagine, if it's successful, that other manufacturers will do something similar -- licensing or working around the patents to various effect. (And if they can't, we'll have swapped an x86 monoculture, where Intel and AMD compete, for a Qualcomm monopoly.)
a global monoculture (ARM everywhere)
Not so much because of the different nature of the different ARM designs and custom extensions. More importantly ARM is already a fairly open architecture with lots of companies designing their own chips based on ARM designs which makes it easy to put design specific hardware for specific tasks (encryption, encoding, etc.)
So, the real risk is fragmentation so that stuff optimised for a Snapdragon may not run at all on Exynos, or a Kirill or an Apple A10 or a Mediatek, etc.
I think the answer is to replicate a windoes API under Linux, that frees you from any particular architecture, and allows either an emulation shim to be used or better, rapid recompilation of windows apps for linux.
In the end the difference between emulating x86 all the way to the operating system internals or simply emulating to the point of the OS interface is one of efficiency.
I could understand paying a little more than the US for a device if it was for valid reasons.
£850 is just over $1.100 right now according to Google. You can't put it down to tax since both those prices will be before tax.
It isn't necessarily additional transport costs as they are not necessarily made in the US.
I'm struggling to thing of whatever this 30% surcharge is caused by - apart from stupidity.
Apple already have OS X running on ARM which the call IOS (no not cisco IOS).
MS have to get Windows running on ARM or they will loose the low power, always on war that the big A already has a huge head start on.
Windows on ARM will help MS get back into the tablet and ultra portable (phone and phone as compute) space which will be the next battleground.
This is MS showing the world they are still relevant and giving Intel a huge kick up the bum for good measure. MS have tried lots of alternatives and they know they need to maintain their legacy connections in order to stay relevant against both Apple and Google.
WoA will flop. Nobody asked for it, and it only suits MS as it will force people to their App Store and servces. x86 emulation is patchy at best and from all reports, very slow. Customers will just be very confused, and desptite the claimed battery life, no real world tests have lived up to the claims. These days, nobody is that far from a power point anyway.
One problem nobody seems to be look at though is for 'always on' 4G most will need to pay for yet another monthly data package, and with Windows hungry telemetry collection, you could eat through that data very quickly.
One problem nobody seems to be look at though is for 'always on' 4G
Trust me, the vast majority of Canadians thought of that immediately. The greed of our wireless oligopoly will ensure that only the very well-heeled will consider this.
The rest of us will rely on ubiquitous WIFI coverage, which is usually free.
"ubiquitous WIFI" ???
u·biq·ui·tous
yo͞oˈbikwədəs/Submit
adjective
present, appearing, or found everywhere.
Maybe in a few parts of Canada. Toronto and Montreal are mere specks in this great land of ours. And not every coffee shop is part of a chain.
It isn't going to work for you when you're sitting in your car at the side of a road. Or in a client's steel warehouse. It isn't going to work at your cottage. It isn't going to work in a lot of places.
And even when you're in a place where it works, you might be in a place where "high speed" is 100-300 kbps, like North of 60, or in the shadow of a mountain, or any where in Manitoba outside of Winnipeg or Brandon.
Like Chromebooks, these will be useful as "portables" in schools, universities, and offices where they can stay on-premises in an area of known good reliable and fast cell phone or wi-fi coverage.
Which means battery life is immaterial, since they can easily be plugged in.
And weight weight savings below 2 kg won't matter.
Wider use in mainstream Canada will require eliminating the dependence on always being connected. I think that will be readily possible within a 3-4 years.
install Linux, or just skip the Windows tax entirely and buy bare metal.
Interesting alternative could be Chromebooks. New models should run Linux apps beside Android and native ones. Actually, you can semi-officially install Gnome Software Center already, but it is still in "beta" (you have to turn it in "dev mode" or what is the name for that). What is still missing (in development) are proper GPU drivers. But it is already around the corner.
FYI, running Linux on ARM isn't anything new.
But Linux is an operating system that has difficulty finding acceptance when it is given away free.
Shilling for Linux in Windows discussions has been tried many times over the years and it's done nothing to help its acceptance.
Maybe the Linux guys paid people to run it they could get some traction in the mainstream market place? Just a suggestion.
> But Linux is an operating system that has difficulty finding acceptance when it is given away free.
That must be why Linux is only found in several billion devices including the majority of servers, 80% of phones and >99% of Supercomputers.
> Shilling for Linux in Windows discussions ...
Actually it is about a device, the Lenovo Yoga C630 laptop. It is perfectly reasonable to ask if it can run other operating systems, such as Linux or RISCOS, though it is rather pointless because these operating systems do not need 32bit x86 emulation as most software is already available in native ARM. It is only Windows that most software is stuck on Intel/AMD, and much of that is now 64bit only so won't run on this device.
> Maybe the Linux guys paid people to run it they could get some traction in the mainstream market place? Just a suggestion.
Paid shilling? No, that is a Windows thing.
most legacy applications were so deeply reliant on x86 quirks that only a subset could run, and customers didn't want a subset. Windows RT was short-lived.
So what's changed? That is still the situation and Windows' ecosystem is still heavily legacy-encumbered. *aaS hasn't significantly altered the situation.
Now, if one could, say, recompile code to the architecture of choice, then that would be fantastic. It's a shame there's no way to do that. Oh, wait...
Of course, you'd have to ditch Redmondware, which is a catch-22 situation if you're moaning about 20+ year-old applications not running. The problem, as always, is the users; you can't really blame them for wanting stability and a familiar working environment but it's causing all manner of crap upstream. The irony is that if you build out your workstations from scratch and have /home sitting on networked storage, you can have interface stability for as long as you like. The trick is to wean the users off of Windows just that one time.
So what's changed?
Windows RT was Microsoft's attempt to have its cake and eat it: cheap devices that wouldn't cannibalise sales of Windows. But the absence of any kind of emulation meant that no one would buy Windows RT, because there was no software, and no one would develop software because no one was buying the hardware. Emulation, especially in hardware, changes everything.
I'm not conviced hours of local video playback is a very good indication of CPU power efficiency. As that's mostly CPU idle as the dedicated video decode in the iGPU decodes and sends the video to the display. Does the CPU even get involved sending the video from RAM to the iGPU or is it DMA by the iGPU. And is video being played just a loop of video which fit's in RAM.
Intel is also using hours of video playback as a performance camparison for it's new laptop CPUs (I think 16 hours is there claim)
I'm not conviced hours of local video playback is a very good indication of CPU power efficiency.
ARM is less about the CPU and more about hardware acceleration anyway. Even if we know nothing about the x86 emulation 25 hours video playback on a 1.2 kg machine is impressive.
It is impressive and 50% longer than intel's* latest in the same type of test. Just I doubt it has much relation to how long it will last when you doing something other than watching a local video.
Of course you can't really benchmark "normal" usage. For example a Daily Mail reader will have much shorter battery life than a BBC News reader, and a Reg reader will get the longest (as they will have something like NoScript or AdBlockers)
I only run NoScript as nearly all adverts are dependant on scripts to display themselves these days. Also I don't get images on a lot of sites either as them seem to use scripts to load them in, so I avoid the page jumping up and down which I find really annoying when I use somebody else computer.
*I always get a down vote when I mention Intel in a negative way.
1. Device made my lenovo, well that is a minus straight away
2. ARM by Qualcomm, another minus
3. Runs windows, multiple minus due to slurp, lack of security, binary incompatability I could go on but you get the idea
4. Price, oops another minus, you can pick up a laptop case and SOC with the same RAM and fast secondary storage for a quater of the retail price of this
5. Move away from Intel, that should be a a big plus
Overall the idea is good but the implementation and implimentor leave much to be desired, windows was designed for x86 and that means that unless it is nolonger actually windows then you are shooting yourself in the foot.
Personally I would suggest that using say Armbian with a x86 emulator on a fast ARM SOC with lots of primary and secondary storage in a full laptop case is going to be a better investment certainly you can make your own right now for less than this product and you just have to pay the once rather than everytime you use it or even just have it connected to the web.
"Overall the idea is good but the implementation and implimentor leave much to be desired"
Yes, but not necessarily for the reasons you give.
"windows was designed for x86"
WinNT was originally designed for alpha. x86 was expected to die by the end of the 1990s.
"and that means that unless it is nolonger actually windows then you are shooting yourself in the foot."
This part doesn't make any sense.
The first problem is the price.
The second the ram load (it's small even for a linux system and will result in swapping with most GUIs)
The third is the dearth of windows apps for it (which MAY be fixed, or may not)
The fourth is Lenovo's emphasis on "always connected" - which feeds back into the first. You have to be well heeled to _afford_ the operational costs of a LTE-connected computer. It will eat the data allowance of most telcos in half a day and end up expensive or crippled after that (remember the recent kerfuffle about Californian "unlimited data" plans which turned out to be anything but?)
Reasons why lenovo are not to be trusted are availible on this site for those they do not keep up with the industry and is the case for Qualcomm also
Clearly I assumed that ARMwindows that doesnt run x86windows code better than under linux would be an obvious issue. The only benefit for windows is the software already availible for that OS which this is not going to run on an ARM but might under an emulator. Said emulator is going to run better under linux due to the improved optimsation/footprint.
If a windows GUI is what you want then they have those for linux too so why bother with thew M$ overheat and slurp
RAM at 4GB this is the current SOC standard amount so you could replace the CPU RAM for an SOC costing less then £100 leaving you with the case and screen and some secondary storage, for what they want you could buy enough to make 4 laptops of similiar spec and have a choice of OS. Again the SOCs dont have a problem running in less than 4GB under linux and the current crop have fast secondary storage so can swap data out as needed.
Always connected, lenovo want to sit where CromeOS created a market but when comparing google and lenovo, the spying of the later is bizarrely more insidious.
> WinNT was originally designed for alpha.
No. Alpha was one of the later ports and it ran in a non-preferred mode (little endian) of that CPU. NT was originally to be developed for Intel 860. This failed so it was moved to MIPS workstations. It was later ported to x86, Itanium, POWER, and Alpha. It was only kept on Alpha after DEC sued Microsoft for using DEC designs of a VMS replacement, the settlement included many $millions (alleged $100million) paid to DEC, keeping the Alpha port and joint promotion.
https://en.wikipedia.org/wiki/Intel_i860
"""Microsoft initially developed what was to become Windows NT on internally designed i860XR-based workstations (codenamed Dazzle), only porting NT to the MIPS (Microsoft Jazz), Intel 80386 and other processors later. Some claim the NT designation was a reference to the "N-Ten" codename of the i860XR.[13]"""
The second the ram load (it's small even for a linux system and will result in swapping with most GUIs)
Not sure that's the case in Linux to be honest. I've been running OpenSuse / KDE in 4GB on two machines (one desktop, one laptop) for some time, and the only time the desktop started swapping was when I needed to do a bit of slightly complex video editing (Kdenlive) on that machine, rather than my slightly better endowed one. Before I retired it, my 2GB EeePC ran (32 bit) OpenSuse KDE reasonably well, very very little swapping.
I have recently added 4GB to the desktop machine, and only because the children tend to leave themselves logged-in when they are tag-teaming homework, each with a copy of LibreOffice and Firefox running as a minimum. Even so, swapping was rare, but I get nervous when sysguard reports less than 1G free memory, and having some "free" memory for file buffers is always a performance boost, even when running from SSD as this machine does.
M.
How does snapdragon compare to atom?
How many surface3s did MS sell?. it was a similar product (Atom X7) and it flunked. If you bought it without keyboard it was useless due to all the bugs. after 3 years most of the bugs have been ironed out, and it is like a very slow tablet running win10 and if you are masochistic you can install virtual box and run linux under a vm, or use Windows Subsystem for Linux. For the same amount of money you get a way better balanced intel i5 laptop.
WoA will not reach critical mass and be a bug ridden frustration.
Microsoft's offerings always seem to be wading through treacle compared to Linux, ARM is better at many power related things but the main issue Windows has is being manacled to Microsoft updates and their hopelessly non-standard office systems: which are designed more to lock you in than to use.
ODF on Linux has always got to be a more efficient choice than running a bloated antiquated predatory OS and programs whose sole purpose is to lock you into monopoly software and suffer endless tedious updates. At least Win10 is fairly stable these days, a point that Linux reached 20+ years ago.
A decent Linux on ARM laptop would be the way forward.
Basically anything beyond say 12 hours or so is a niche market for both laptops and phones. People aren't likely to pay over the odds for a different architecture that's much slower than x86 to get it. They'll rather buy a cheaper Intel laptop and a second battery or powerpack to give them their 25 hours.
You're exactly right there.
It is senseless marketing to focus on over-achieving something that need only be totally acceptable.
A race for low weight is pointless after a sufficiently low weight for the target use is achieved.
A race for thinness is pointless after sufficient thinness is achieved (and may even backfire because it means the victim, er uh, customer need to buy an ugly case to keep it in.
And a race for battery life is pointless after a realistic 18 hours is achieved.
"Basically anything beyond say 12 hours or so is a niche market for both laptops and phones. "
Um......no. Go away troll.
Having had brick-size mobile phones with lifespans measured in single hours, then having those move to "almost a workday", then beyond that, it was battery lifespan driving mainstream acceptance of phones for many years and consumer resistance to early smartphones was largely centred around the short battery life. Battery size now is limited by fire factors and what airlines will allow onboard - so all further gains have to come from efficiency, not bigger batteries.
As for laptops... I had a NEC portable back in the 1980s with a 4*80 char LCD that would run a month on a set of 8 AA cells. nothing since then has approached that battery life or lightness.
One of my pet hates when travelling has always been having to curate the battery to ensure it lasts the trip. It may sometimes be 3-5 days between charges in some areas. There's definitely a market for longer runtime.
I didn't say there was NO market, I said it was a niche market.
How many of us travel to areas where there is no access to electricity for 3-5 days at a stretch (but doing something where we still need our electronic devices, so roughing it on a camping trip doesn't count) That sounds like the very definition of 'niche' to me.
> It may sometimes be 3-5 days between charges in some areas.
Do those areas have LTE available (or public WiFi) for the 'Always Connected' that this device is intended for ?
Can you not recharge from your car or RV or are these areas so deep in the woods that they can't be reached by roads?
Any decent enterprise class notebook has an LTE modem. That's kind of a given anymore. You mean there are some that don't?
Now for $550 or $650 it would probably be a winner. For over a thousand, its in the same price range as an HP Pro or Elitebook or a Dell Latitude for something unproven and subject to MS whims about the architecture, which is something I wouldn't want to gamble on if I'm an enterprise purchaser or software developer. Plus, how much performance is going to be lost in the emulation layer?
I can't say I've been too impressed with emulating x86 on ARM thus far. Its still more than I would probably pay for something without an x86, but with a long battery life it might be worth it for anyone working in the field.
I wouldn't assume this laptop really has significantly better battery life than x86 laptops with long battery life until some real benchmarks were done for something besides video playback. Video playback uses dedicated GPU hardware and depends too much on screen brightness so you can't really compare across laptops unless you know the video being used and brightness setting in nits.
They'd need to test both native and non-native code, too. If one of the applications you need for work isn't available for ARM, you might get LESS battery life with this than you would from a competitor Windows laptop due to the translation overhead.
I also doubt we're going to see a rush of Windows developers porting to Windows on ARM until it has been around for several years and looks successful, given that last time Microsoft went this way they quickly killed the product and any developers that ported to it then wasted their time.
If Microsoft want to get away from x86 then step 1 is change all the toolchains to emit platform neutral binaries. e.g. LLVM bitcode. When the user runs the exe for the first time the OS can compile it. If software is platform neutral then over time perhaps the hardware won't matter too much.
At present, a version of Windows for ARM will go as well as a lead balloon. Just like all the other times MS have tried to port away from x86 without laying the groundwork.
...who here actually cares about running Windows on ARM?
I use Windows extensively, and quite like it; but let's face it - the reason for running Windows on anything is the available software. Which only works on x86. Yes, yes, I know there's some promised-land of emulation, but in literally every case I have experienced the overhead has been obscene, driving either the power consumption up or the user experience down.
I just really don't get who their target market is. Want a device that can play video for ages and ages? Get a Linux or *shudder* Android device.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
