ARMed Devices Are A Steam-roller Catching Up To Intel

“MediaTek has revealed an upcoming Helio X30 system-on-chip (SoC) that will likely be one of the first 10nm fabricated SoCs to arrive when it ships in mid-2017. The deca-core mobile SoC will also be one of the first to use ARM’s new Cortex-A73 successor to the Cortex-A72”
 
See MediaTek’s 10nm Mobile-Focused SoC Will Tap Cortex-A73 and -A32
The dam on ARMed performance is about to burst. OEMs are about to dump ARM’s latest tech at 10nm resolution into smartphones. How long before Lemaker’s Cello becomes obsolete? Now, before it’s shipped in numbers. In a year, the world will be flooded with ARMed alternatives to Intel CPUs in all formats/roles/performance ranges.

Can you hear the rumbling behind you, Intel?

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology and tagged , , , , , . Bookmark the permalink.

14 Responses to ARMed Devices Are A Steam-roller Catching Up To Intel

  1. oiaohm says:

    kurkosdr the value for GNU/Linux varies a lot between countries.

    Lets say I am in the market of selling computers to Malaysia government. Attempt to sell a windows machine up hill battle. Why that market is 95+ percent GNU/Linux with scraps of windows and os x making up the remaining.

    Now lets say we swap across to Malaysia on statscounter and then you notice something strange no Linux. Huge unknown value. Go check Malaysia government customised Linux distribution and find it does not ID itself. So that 7 to 14 unknown swing line around the popularity of OS X and Windows 8.1 is most likely Linux.

    Reality Malaysia shows we don’t have good numbers of Linux Market share. If you take the Linux desktop numbers globally for the volume it covers there is more desktop computers that are GNU/Linux than OS X.

    Remember kurkosdr that desktop number robert is looking at on statscounter excludes android. So there is a question how much Linux has in fact grown or is it Linux holding it own and Windows shrinking against Android.

    http://gs.statcounter.com/#all-os-ww-monthly-201605-201607-bar
    This is the interesting one to look at. Identifiable linux is .94 percent yet there is almost 4 percent that that is not identified. So over all market Linux still is showing up on statscounter as it 1 percent of old. But the Malaysia shows us that a percentage of Linux desktop machines are not identifiable.

    Thing to remember is go back to 2000 pull up same kind of information windows desktop + windows mobile back then make up over 90 percent of it. So Microsoft had quite a fall from grace. At the rate android is going it could in a few years account of 50% of web traffic.
    http://gs.statcounter.com/#all-os-MY-monthly-201605-201607-bar
    Places like Malaysia were we know the government computers are not getting counted right. If everything detected that is windows in Malaysia you only get to 36.9% and android is sitting on 43.32 percent.

    So we have already crossed the point in some markets if you are making software and want the biggest customer base to sell to its Android.

    kurkosdr there is a serous question here. If Microsoft keeps on losing market share the way it is and with OS X and Linux keep on holding market share and Android keep gaining their market share how long before applications on Windows have no choice but to port to OS X, Linux or Android or possibly all 3.

    Valve making steam box and porting games to it kinda makes sense when you notice the fact Microsoft market share is vaporising.

  2. kurkosdr says:

    Chuckle. I notice recently Germany shows GNU/Linux gets 3.5% share of desktop page-views weekdays and 4% on weekends.

    Then, allow me to post a correction:

    Just like 96% of the market does…

  3. kurkosdr wrote, about neglect of GNU/Linux, “Just like most of the market does…”

    Chuckle. I notice recently Germany shows GNU/Linux gets 3.5% share of desktop page-views weekdays and 4% on weekends. Then there are Chromebooks and “other/unknown” and Android/Linux. Uruguay varies from 13 to 14%. That’s not neglect but diversification/absence of monopoly. I like it.

  4. kurkosdr says:

    ” That reasoning neglects GNU/Linux entirely ”

    Just like most of the market does…

  5. kurkosdr wrote, ” ARM can have the best chips, but without software from Google’s side, nothing will happen. Unfortunately, Google for the moment seems kinda lost between Chrome OS, running Android apps on Chrome OS and windowed Android.”

    That reasoning neglects GNU/Linux entirely. We won’t go away. GNU/Linux works for people and will run on just about any hardware.

  6. oiaohm says:

    kurkosdr
    MAYBE on a Mac, but those things are expensive
    Yes Windows is expensive compare to Linux based solutions as well.

    The most expensive CAD program on earth only runs on Linux. So there is software in the high end that is Linux exclusive as well. Its about time you start pulling your head out the sand and notice how everything is changing. You have many times stated faults that are not just Linux.

  7. oiaohm says:

    kurkosdr
    Yes there will be “light” apps like AutoCAD 360. But the real thing, aka the really professional application, needs a Windows PC.
    Change apps for versions, AutoCAD 360 for photoshop and Windows PC for Mac and you typed the exact line that was said about the first Photoshop port to windows.

    Something to remember the AutoCAD 360 pro now on android can create new documents. The first version of AutoCAD on phones was only a viewer. So we are seeing more and more functionality of AutoCAD appear on phones and tablets version. This is not the only application were we are seeing this happen.

    Really no point messing with my handle kurkosdr you don’t have the years of experiences seeing mac zelots and what they said to see how lot of what you are saying now sounds exactly like mac zelots of old.

    Chromebooks/chromeboxs/Linux systems running android applications give android applications access to lot bigger GPU than a phone or tablet can have.

    So functionality of a Android application is limited by GPU and CPU power they can access. Those limitations are defined partly by how much cooling.

    The kinda lost state between chromebooks and android is what should worry Microsoft long term. I would not say that Google is lost. Android on chromebooks also equals Android applications accessing bigger GPU units. Even arm based chromebooks use chipsets with larger gpu units than what is phone in mobile phones or tablets.

    Google is not the only player in the game either. Ubuntu is not giving up there attempts. Of course I would say the winner is going to be something android compatible maybe a google product maybe not.

    kurkosdr question how do you get steam games on to a chromebook. The answer is insane amount of difficulty.

    So there is a gap. When Shashlik gets that it can operate without using a full vm running(remove binder requirement) Android applications add on qemu usermode emulation allowing steam games and win32/win64 application in wine to work but this requires a decent mesa stack and wayland. So at this point a standard Linux distribution looks fine.

    kurkosdr there are a list of events that have to happen before I expect Arm Linux laptops in volume. But the list of events required are getting shorter.

    Big one is lot of arm chips don’t have large Pci-e interfaces. I am expecting to see these with 7nm or 5nm arm64 chip production. Arm on desktop will not happen in a big way because of google. Google is backing openpower in server room. The parties backing arm64 in server room are the ones that are likely to trigger arm64 on desktop. Why people developing software to deploy on servers like having keyboard mouse and screen when debugging what they are planing on deploying. This is why Server OS solutions are in with a chance.

    Yes Microsoft in windows 10 adding Linux support is about allowing developers to debug linux stuff on windows without booting into Linux. Of course this will not help when developers are wanting arm64 support. Due to the battle starting between openpower and arm64 in server vs x86 developers could be wanting 3 different hardware solutions for testing before deployment.

    Of course kurkosdr has ignored that Microsoft has been fairly much forced to add Linux support in the form of Ubuntu and what this in fact means. Kurkosdr Microsoft first started the layer Bash for Windows uses to run Android applications on Windows but it become more critical to be able to run close to a normal Linux Distribution applications. So Microsoft to keep some hold on market needs Linux compatibility. The question is long term will we need Windows compatibility the answer is possible no with the way applications are developing on Android.

    The future is no where near black and white kurkosdr as you are making out.

    Google or someone like HP (who is backing arm64 in server room) will trigger arm64 Linux on the desktop. This could be a google OS or it could be a normal Linux Distribution. HP could be Debian if Dell could be Ubuntu.

    The big difference is arm64 on desktop by Linux will not be like arm done by windows. Running x86 binaries will be possible on Linux arm64 systems including windows binaries.

  8. kurkosdr says:

    MAYBE on a Mac, but those things are expensive

  9. dougman says:

    “But the real thing, aka the really professional application, needs a Windows PC.”

    No one “needs” Windows anymore. 80% of what people do on Windows, can be done with Linux/Android/Mac, etc..

  10. kurkosdr says:

    Wasn’t this stated last year and the year prior? But nothing has changed.

    Well, duh. As I ‘ve said before, any kind of “arm on the desktop or laptop” will happen because of Google. ARM can have the best chips, but without software from Google’s side, nothing will happen. Unfortunately, Google for the moment seems kinda lost between Chrome OS, running Android apps on Chrome OS and windowed Android.

    Obscure DIY oddities like running Debian on cheap ARM boards like the OdroidC2 might serve their owner well (or at least the owner has managed to convince TLW she doesn’t need 1080p video support) but are not going to make ARM on the desktop or laptop mainstream. But Pog is allowed to fantasise.

    @ohioham
    Yes there will be “light” apps like AutoCAD 360. But the real thing, aka the really professional application, needs a Windows PC.

  11. oiaohm says:

    http://www.eetimes.com/document.asp?doc_id=1329202

    Scary part is the speed TSMC and ARM is moving down though the nm. 2017 is when 7 nm is market to be tapped out. 5nm somewhere between 2018-2022 by current forcast.

    7nm should allow 4Ghz+ chip clock speeds without heating chip up.

    kurkosdr please learn to do more research before posting. AutoCAD 360 ever heard of it. Yes it has a arm64 version for Android. You said “Too much effort” yes its a lot of effort to rewrite some of those specialist applications for Android but the problem is we are seeing instances of this happening. kurkosdr you are making the same arguement what was made about photoshop and other multi-media at one point that it would not be ported to Windows from mac that happened so we should not be surprised if this happens with android.

    http://www.shashlik.io/what-is/
    kurkosdr other thing is we are seeing Android apps become universal between Linux desktop, chromebooks and android phones.

    Too much effort. Microsoft essentially took the professional and gamer crowd out of ARM’s reach by refusing to allow third-party Desktop Apps (win32) to be compiled for Windows RT.
    Lets look at the Linux side on this for a moment. Crossover is coming to arm devices.
    https://www.codeweavers.com/products/crossover-android
    First is going to be x86 chromebooks and android running windows applications by the translation level. But work is under way to allow wine core to operate on arm64 and work looking into qemu to see if user mode emulation it has can be reused with wine in a more performing way.

    So just because Microsoft did not do something does not mean it will apply on the Linux side. Yes Windows x86 binaries running on android,chromebooks and linux machines arm64 based should be expected at some point question here is level of over head.

    https://wiki.debian.org/QemuUserEmulation
    Yes is very simple to forget Linux systems don’t have to give a rats what arch an application binary is written for its just a question of overhead. Native binaries don’t require emulation so run faster. Linux is a very much different beast to windows. Linux systems installing arm64 binaries next to x86 ones is perfectly fine on lots of systems.

    The Android world has standardized to ARM, and the Android OEMs are not going to jump to x86 or MIPS64 or whatever else and risk compatibility and performance hit issues with native apps, no matter what kind of products non-ARM vendors release.
    This is contains mistake. Changing cpu type does not have to risk compatiblity the core is Linux so cpu type emulation is on the cards. The issue is purely performance. The reality is x86 does not run arm qemu usermode emulation fast enough. MIPS64 is still a question mark. Many things about MIPS64 means it does run emulation quite well.

    Loongson Technology from china developed extensions to the MIPS64 instruction set that reduce emulation overhead massively. So android phone will arm64 or something that can run arm instructions effectively by emulation. The thing to be scary if you have a processor that can run arm instructions effectively by emulation it can most likely run other archs like x86 effectively by emulation. Stupid as it sounds is that it is possible that something running by emulation could end up using less power. Mobiles phones it a battle of power usage as well.

  12. dougman says:

    “In a year, the world will be flooded with ARMed alternatives to Intel CPUs”

    Wasn’t this stated last year and the year prior? But nothing has changed.

  13. kurkosdr wrote, “Microsoft essentially took the professional and gamer crowd out of ARM’s reach by refusing to allow third-party Desktop Apps (win32) to be compiled for Windows RT. Canonical will pretend to bring Linux Desktop apps to ARM, but with Ubuntu ARM devices being a cute oddity, it won’t make a difference.”

    M$ has nothing to do with the rise of ARM except to try to catch the wave which they failed to do. The endgame of Wintel is a tiny market compared to ARMed instances. Mobility, IoT, set-tops have all gone heavily to ARM and now ARM is a serious competitor to Wintel on clients, workstations and servers. ARM has designed cores that are focused on mobility but several makers are adding large caches and RAM to compete head to head with Intel and it’s all about GNU/Linux, not Android/Linux. What has been done with cross-compilation and emulation is about to be done on real ARMed hardware. TLW is impressed despite a couple of glitches, mostly due to my ineptitude. e.g. Today she complained she could not edit some files. I’d forgotten to add her to the “ftp” group. She’s liking audio, and she can live with compromised video. Odroid-C2 is far superior to several older PCs we have at home and she loves the price. You know Odroid-C2 is a toy made with ARM’s stuff from 2014, eh? The Cello will be many times more powerful. I’d originally conceived Odroid-C2 as a thin client but it is working as a thick client beautifully. What’s coming out ARMed later in 2016 and 2017 will be perfectly competitive on price/performance. It doesn’t matter how much faster Intel can make a chip if ARMed stuff is good enough and cheaper.

  14. kurkosdr says:

    Intel’s problem is that:

    1) Most people don’t need to upgrade their existing PCs more often than a 5-year cycle, save for specialized cases like professionals and hard-core gamers.

    2) The Android world has standardized to ARM, and the Android OEMs are not going to jump to x86 or MIPS64 or whatever else and risk compatibility and performance hit issues with native apps, no matter what kind of products non-ARM vendors release.


    So, with that knowledge in hand, the following conclusions can be derived:

    1) Intel will keep enjoying a predictable market of professionals and gamers that will keep buying Intel and will help Intel fund their future R&D, for the simple reason that all those special apps, be it Skyrim or Cakewalk or AutoCAD, will not be lightheartedly rewritten to Android. Too much effort. Microsoft essentially took the professional and gamer crowd out of ARM’s reach by refusing to allow third-party Desktop Apps (win32) to be compiled for Windows RT. Canonical will pretend to bring Linux Desktop apps to ARM, but with Ubuntu ARM devices being a cute oddity, it won’t make a difference.

    2) Intel isn’t going anywhere in Android.

    3) Which leaves out Windows PCs, and specifically their use by non-gamer non-pro users. There is where the next fight is. Not of course for Canonical, but for Google. Can Google make the Windows PC obsolete for the non-pro and non-gamer user? That’s their next OS bet. This is the reason they experiment with Chrome OS-Android hybrids, windowed Android, Android apps on Chrome and whatnot. Although to be honest, it looks like an internal bloodbath to me. Google doesn’t have a clear-cut strategy to take on Windows PCs for the non-gamer non-pro user.

    So, before you go all premature jubilation, remember that it all depends on Google as much as it depends on ARM. ARM can release the best silicon in the market, but if Google fumbles in the software front, they won’t steal users from Windows PC.

    Hold the victory speeches for later… if Google doesn’t provide a riverbed for the water from the ARM dam to flow to Intel valley, the water will disperse…

Leave a Reply

Your email address will not be published. Required fields are marked *