Future iPhones to wield OpenCL acceleration

13

Comments

  • Reply 41 of 72
    Quote:
    Originally Posted by apfel View Post


    Yes and that is a good thing. But a lot of this is just theory, especially if you are using modern features there doesn't remain a lot of cross-platform. You have to use the extensions - and they different. And OpenGL in OSX is a special thing.

    Linux and Windows cross-platform is working "OK". But OSX is another thing. Apple has implemented his own OpenGL Runtime. The IHVs are offering only the drivers (this is the modell microsoft is using). But the problem is, that this runtime likes to work a little bit different from the IHV implementation under Windows and Linux.



    So in practice there remains not a lot of the cross-platform features in the sense, that you have OpenGL-Code that will work in the same way on different plattforms.



    I don't know what will happen with OpenGL 3.1 and how much time it will take but I wouldn't expect to much. Maybe they are taking the geometry shader in the core. But I don't think that we will see the new Object Modell.



    http://www.youtube.com/watch?v=sddv3d-w5p4



    Alright, I finally understand your point: While OpenGL is supposed to be platform independent, rather than proprietary and system-locked like DirectX is to Windows, thanks to the vendor-specific extensions required when writing OpenGL, you believe these vendor-specific extensions nullify OpenGL's "platform independent" promise. Am I right?



    If I'm right, that's an interesting point, but DirectX has only seen any major adoption on the Windows PC, not mobile-optimized devices, which is the focus of the article: the iPhone/iPod touch mobile WiFi platform, as well as integrated systems like the majority of game consoles and Macs; there are no Linux-powered consoles or devices comparable to the iPhone in terms of graphics, but if there were, they'd likely make use of OpenGL as well.
  • Reply 42 of 72
    jeffdmjeffdm Posts: 12,946member
    Quote:
    Originally Posted by wobegon View Post


    But see, you've just brought up Apple's major competitor to netbooks that's already in existence and FAR more popular than netbooks - the iPhone/iPod touch mobile WiFi platform. If all you're gonna do is surf the web, check and write emails, IM, and listen to some music, the iPhone does all that and more and almost more importantly it does all that and more well. Fast boot up, zippy performance, decent storage space.



    ...

    because they realize just how underpowered their netbook is, in addition to all the other poor compromises they make, notably micro, carpal tunnel-inducing keyboards and claustrophobic, migrane-inducing little screens (generally low res too, which, even though these screens are small, give a fuzzy haze over everything).



    I don't understand how someone that has trouble with a netbook's small screen size and keyboard size is really going to be happy with something has a far smaller screen and keyboard.
  • Reply 43 of 72
    Quote:
    Originally Posted by JeffDM View Post


    I don't understand how someone that has trouble with a netbook's small screen size and keyboard size is really going to be happy with something has a far smaller screen and keyboard.



    Because you're not going to be typing as much on an iPhone in general - you'd do extensive writing on the full-sized laptop you sync it with (which yes, I recognize, is a caveat) - and when you are [typing], you are doing so with your thumbs, which, compared to the alternative - tiny BlackBerry keyboard w/ hard to press keys - is more comfortable and faster to type on.



    Finally, in comparison, typing with your thumbs less (on an iPhone in part because it's not meant for lengthy typing and in part because you'd have a full-size laptop to dock it with for said lengthy typing) vs typing with cramped hands more (on a netbook), I would say the former, combo iPhone/full-size laptop, wins on almost every level except overall price.
  • Reply 44 of 72
    mactelmactel Posts: 1,275member
    Ok, now I get it. It is as if someone slammed a hammer to my head (ouch!). OpenCL's rush out-the-door is not so much for the Macintosh lines but for the iPhone lines. OSX core API are used for the iPhone OS, so what is developed for OSX migrates to the iPhone OS.



    By June 2009 we'll see an impressive iPhone with the fruits of the PA Semi acquisition in a new CPU core and a PowerVR core saddled up next to it. Apple wants to dominate the handheld video game arena obviously. Duh!



    Snow Leopard will benefit but I was thinking how much whiz-bang do we need in the OS (Apple?). It makes more sense for the handheld market to take advantage of OpenCL.
  • Reply 45 of 72
    Quote:
    Originally Posted by MacTel View Post


    By June 2009 we'll see an impressive iPhone with the fruits of the PA Semi acquisition in a new CPU core and a PowerVR core saddled up next to it.



    I wonder how long it will really take. Are we too optimistic?



    If Imagination Technologies is only now hiring OpenCL compiler programmers how long until they are hired and get a useful compiler going? And then how long until programmers can start using it?

    And do we really know that Snow Leopard Mac and Snow Leopart iPhone are developed in sync? Or isn't it more likely that the Mac version will be out first and the iPhone version trailing a bit later?



    In a nutshell I doubt any of this will be finished for the iPhone by WWDC. They might have a prototype to show, but it won't be a product until September/October 2009 or even January 2010.





    Co-Processor

    In regards to PowerVR and PA Semi, I'm pretty certain this is also about a co-processor for the Mac not just for the iPhone.

    Think about it: by building their own low-power OSX accelerator Apple can achieve 2 things:

    - speed up aspects of OSX

    - ensure OSX only works on hardware with that chip installed, i.e. Apple hardware.



    The latter will help fight clones very effectively. Sure, OSX needs to be backwards compatible so they can't just have code that only works on the latest hardware with that co-processor. But I can imagine that new hardware will have new features which don't need to be backwards compatible - and won't be because it will require that co-processor.

    The point for Apple is to be compatible with Windows, but not offer Windows hardware compatibility with OSX.
  • Reply 46 of 72
    Quote:
    Originally Posted by apfel View Post


    The mobile Devices are using OpenGL ES. There they did the Cleanup. Everyone hoped that they make also the Cleanup in OpenGL3, but it didn't happen how I wrote more than one times. So OpenGL and OpenGL ES are not any more identical. It can easily happen that you have to rewrite your Code.



    PS3:

    The PS3 has its own native API. There is something like an OpenGL ES Wrapper, but there are so many extensions that there is no compatibility. This implementation is so slow, that no one is taking it. You take the native API.



    Wii:

    So from the things I heard (but I am not sure), the Wii also doesn't use OpenGL. The Wii is using its own API, but there are many similarities to OpenGL.



    http://www.opengl.org/news/comments/...use_opengl_es/



    Quote:

    Sony confirms PlayStation 3 to use OpenGL ES



    At Sony Computer Entertainment’s PlayStation meeting the company president showed off the development kit hardware and confirmed and the choice of OpenGL ES as the graphics API in order to facilitate rapid developer adoption and content creation. The Sony Playstation 3 will be using OpenGL ES for the 3D graphics API, NVIDIA’s Cg shader language and tools (CG compiler, FX composer, ShaderPerf and PerHUD), and the development tools will also incorporate support for the COLLADA format for art asset interchange so that developers can share interactive 3D art among multiple platforms. This also opens the possibility that after the PS3’s launch, the platform could be opened to general development to anyone, not just game developers. (In marked contrast to Xbox).



    Posted on 07/25 at 06:43 AM



    http://en.wikipedia.org/wiki/PSGL



    Quote:

    PSGL is a 3D computer graphics API based on OpenGL ES for the Sony PlayStation 3. A previous version of PSGL was available for the Sony PlayStation 2 but was largely unused.



    http://www.khronos.org/message_board....php?f=5&t=673



    Quote:

    Quote:
    Originally Posted by MikeJM


    The GDC slides ask for feedback but give no direction as to how to do so.



    It would be great if PSGL was GLES 1.1+ compliant and PSGL 2.0 was GLES 2.x compliant. Fewer discrepencies in APIs make it much easier to develop on multiple platforms. At my company we work on next/current gen mobile, pc, and console products and it is is a constant struggle to keep cross platform libraries clean. Being able to move between things like ATI Imageon 238x and Playstation 3 (almost) seemlessly would just make life easier.



    Thanks.





    PSGL is compliant with GL ES 1.0. It is very close to be 1.1 compliant, but not quite (Some features added to the fixed pipeline are not useful to a full programmable hardware). It has all the features of GL ES 2.0, but using the Cg language and runtime instead of GLSL ES (therefore not compliant with ES 2.0)



    We share your need for fewer discrepancies in APIs, so we are working closely with nVidia on Cg. For instance Cg 1.5 will be identical on PS3 and all the other platform supported by this run-time and language (OpenGL on Windows, DirectX on Windows, OpenGL on Linux, OpenGL on Mac, OpenGL on Solaris).

    Since everything is about shader nowadays, we are also very interested to find out how to create a cross-platform COLLADA FX profile, rather than the multi-platform support that we currently have.



    Regards



    -- Remi



    Quote:
    Originally Posted by TSCondon


    I share MikeJM's interest in full compatibility with GLES 2.x, for a couple reasons:



    1. Cg is not as cross-compatible as GLSL; in particular, it's seriously lacking on ATI graphics cards. And since ATI is an unavoidable block of the PC game market, we have to support both brands of graphics card, and GLSL fits the bill better than Cg, as it is equally supported on both.



    2. GLSL is an open standard, so in the long run, if Sony switches to a different provider for their graphics card (for the PS4, for example) then presumably that card will have solid support for GLSL, and the old titles can take advantage of the new card just as easily.



    3. There are a lot of resources for programming in GLSL. Schools are more likely to teach GLSL than Cg. With Microsoft pushing DirectX so hard, there need to be platforms where OpenGL/GLSL rules unquestioned; otherwise, DirectX will keep edging out OpenGL, which is bad for Sony, while a Dev platform that puts a check on Microsoft's dominance by providing OpenGL is a very good thing for them.



    4. Nvidia provides all their PC-market chips with support for OpenGL 2.0, which means they already have GLSL compilers and support . I can't imagine it would take much to port it over to the PS3 firmware.



    5. I happen to like the syntax and style of GLSL better than CG/HLSL



    Even if GLSL runs sub-optimally on the PS3, it still seems like the benefits of supporting GLSL are worth it. Anyway, just my $0.02



    From what I could find Nintendo's API is proprietary and duplicates what it needs of OpenGL w/o opening it up.



    It's quite clear that from the OpenCL Working Group listing of Companies, the PS3 will embrace OpenCL whether they want to choose Nvidia or AMD/ATi for their GPU needs and whether they want to choose IBM, Intel or AMD for their CPUs.
  • Reply 47 of 72
    Quote:
    Originally Posted by MacTel View Post


    Ok, now I get it. It is as if someone slammed a hammer to my head (ouch!). OpenCL's rush out-the-door is not so much for the Macintosh lines but for the iPhone lines. OSX core API are used for the iPhone OS, so what is developed for OSX migrates to the iPhone OS.



    By June 2009 we'll see an impressive iPhone with the fruits of the PA Semi acquisition in a new CPU core and a PowerVR core saddled up next to it. Apple wants to dominate the handheld video game arena obviously. Duh!



    Snow Leopard will benefit but I was thinking how much whiz-bang do we need in the OS (Apple?). It makes more sense for the handheld market to take advantage of OpenCL.



    It's for OS X 10.6 and will benefit 2008 Hardware from the Server, to the Desktop to the Handheld platforms. Any earlier system specs that will benefit are minor, but will still be something on the Core2Duo Intel systems.



    OpenCL will benefit Linux, OpenSolaris, FreeBSD, OpenBSD, NetBSD, et.al, as well. Those systems all now have Nvidia and ATi drivers.
  • Reply 48 of 72
    Quote:
    Originally Posted by hobBIT View Post


    I wonder how long it will really take. Are we too optimistic?



    If Imagination Technologies is only now hiring OpenCL compiler programmers how long until they are hired and get a useful compiler going? And then how long until programmers can start using it?

    And do we really know that Snow Leopard Mac and Snow Leopart iPhone are developed in sync? Or isn't it more likely that the Mac version will be out first and the iPhone version trailing a bit later?



    In a nutshell I doubt any of this will be finished for the iPhone by WWDC. They might have a prototype to show, but it won't be a product until September/October 2009 or even January 2010.





    Co-Processor

    In regards to PowerVR and PA Semi, I'm pretty certain this is also about a co-processor for the Mac not just for the iPhone.

    Think about it: by building their own low-power OSX accelerator Apple can achieve 2 things:

    - speed up aspects of OSX

    - ensure OSX only works on hardware with that chip installed, i.e. Apple hardware.



    The latter will help fight clones very effectively. Sure, OSX needs to be backwards compatible so they can't just have code that only works on the latest hardware with that co-processor. But I can imagine that new hardware will have new features which don't need to be backwards compatible - and won't be because it will require that co-processor.

    The point for Apple is to be compatible with Windows, but not offer Windows hardware compatibility with OSX.





    I don't think a June 2009 refresh for the iPhone is very long at all. The PowerVR SGX 540 (a very new design that's more powerful than what Apple will use in the iPhone) has been in silicon for almost a month.



    http://www.design-reuse.com/news/196...vr-sgx540.html



    It doesn't really matter if OpenCL is ready or not from Imagination it's software. The hardware play is likely well underway. Apple will most likely be using a PowerVR SGX 520 or 530 chip, ARM 11 processor and a new Wifi/BT chip in the next iPhone. You may not even see a lot of customization this time around. That may wait for 2010.



    I do not agree that the Snow Leopard will be ready for desktop/laptop faster than the iPhone. Count the different chipset and configurations for the Desktop/Notebook/servers and then compare that number with how many variations of iPhone need to be tested. The iPhone has the clear advantage here.



    http://buyersguide.macrumors.com/#iPhone



    average length between iPhone refreshes is 126. This cycle will likely be extended but no way are we waiting until 2010 for a new iPhone.





    Apple will beat the clones in court rather than spend the uncessary money modifying their hardware with a co-processor. They didn't do it when Trimedia was the hotness and they're not going to do it now.
  • Reply 49 of 72
    Quote:
    Originally Posted by hmurchison View Post


    I don't think a June 2009 refresh for the iPhone is very long at all. The PowerVR SGX 540 (a very new design that's more powerful than what Apple will use in the iPhone) has been in silicon for almost a month.



    http://www.design-reuse.com/news/196...vr-sgx540.html



    It doesn't really matter if OpenCL is ready or not from Imagination it's software. The hardware play is likely well underway. Apple will most likely be using a PowerVR SGX 520 or 530 chip, ARM 11 processor and a new Wifi/BT chip in the next iPhone. You may not even see a lot of customization this time around. That may wait for 2010.



    I do not agree that the Snow Leopard will be ready for desktop/laptop faster than the iPhone. Count the different chipset and configurations for the Desktop/Notebook/servers and then compare that number with how many variations of iPhone need to be tested. The iPhone has the clear advantage here.



    http://buyersguide.macrumors.com/#iPhone



    average length between iPhone refreshes is 126. This cycle will likely be extended but no way are we waiting until 2010 for a new iPhone.





    Apple will beat the clones in court rather than spend the uncessary money modifying their hardware with a co-processor. They didn't do it when Trimedia was the hotness and they're not going to do it now.



    I think he meant 2010 for an iPhone built using custom chips rather than off-the-shelf parts, though I find it unlikely that Apple will add significant new features every year like they do with their iPod lines.
  • Reply 50 of 72
    Quote:
    Originally Posted by dagamer34 View Post


    I think he meant 2010 for an iPhone built using custom chips rather than off-the-shelf parts



    Yes, that's what I meant. Sorry if this wasn't clear.



    I expect Apple to broaden the iPhone models before using these new custom chips and OpenCL features. Perhaps by March we will see an iPhone nano and an iPhone 3G with a slider keyboard along with a new mainstream 3G model that has a few extra features (higher-res camera?, flash?, movie capabilities?) and more memory.

    Only 6-12 months later will we see a new iPhone using a PA Semi designed core using OpenCL.



    Apple would be prudent to ensure existence of a few tried and tested models based on existing technology - just in case Apple's first SoC chip design has flaws. Afterall this will be a big step forward and they don't want to bet the whole iPhone success on a single (possible) failure.





    I could even see that new chip shows up in a completely new product (tablet?) first and in an iPhone only once all the 'version 1' kinks are sorted out.
  • Reply 51 of 72
    mactelmactel Posts: 1,275member
    Quote:
    Originally Posted by hobBIT View Post


    I wonder how long it will really take. Are we too optimistic?



    If Imagination Technologies is only now hiring OpenCL compiler programmers how long until they are hired and get a useful compiler going? And then how long until programmers can start using it?

    And do we really know that Snow Leopard Mac and Snow Leopart iPhone are developed in sync? Or isn't it more likely that the Mac version will be out first and the iPhone version trailing a bit later?



    In a nutshell I doubt any of this will be finished for the iPhone by WWDC. They might have a prototype to show, but it won't be a product until September/October 2009 or even January 2010.



    That would be assuming Apple's programmers are not doing the work that a dedicated Imagination Technologies programmer would be doing. Apple own 3.6% so sending a team to work within the company isn't out of the question since they have a stake in the company they have some say in its operations (some).





    Quote:
    Originally Posted by hobBIT View Post


    Co-Processor

    In regards to PowerVR and PA Semi, I'm pretty certain this is also about a co-processor for the Mac not just for the iPhone.

    Think about it: by building their own low-power OSX accelerator Apple can achieve 2 things:

    - speed up aspects of OSX

    - ensure OSX only works on hardware with that chip installed, i.e. Apple hardware.



    The latter will help fight clones very effectively. Sure, OSX needs to be backwards compatible so they can't just have code that only works on the latest hardware with that co-processor. But I can imagine that new hardware will have new features which don't need to be backwards compatible - and won't be because it will require that co-processor.

    The point for Apple is to be compatible with Windows, but not offer Windows hardware compatibility with OSX.



    I could see the next iPhone have a separate PowerVR and PA Semi and would be surprised if they came on the same silicon by next June, but you never know.
  • Reply 52 of 72
    hmurchisonhmurchison Posts: 12,259member
    Quote:
    Originally Posted by hobBIT View Post


    Yes, that's what I meant. Sorry if this wasn't clear.



    I expect Apple to broaden the iPhone models before using these new custom chips and OpenCL features. Perhaps by March we will see an iPhone nano and an iPhone 3G with a slider keyboard along with a new mainstream 3G model that has a few extra features (higher-res camera?, flash?, movie capabilities?) and more memory.

    Only 6-12 months later will we see a new iPhone using a PA Semi designed core using OpenCL.



    Apple would be prudent to ensure existence of a few tried and tested models based on existing technology - just in case Apple's first SoC chip design has flaws. Afterall this will be a big step forward and they don't want to bet the whole iPhone success on a single (possible) failure.





    I could even see that new chip shows up in a completely new product (tablet?) first and in an iPhone only once all the 'version 1' kinks are sorted out.



    Oh ok I totally agree with that. In fact that really appears to make sense.



    Snow Leopard- optimizations but no drastic changes mates well with off the shelf parts.



    10.7 Lion - New feature release along with new custom iPhone/iBook (hehe the return) chips.



    Hmmmm lets see what features I'd like to see in a iPhone/Netbook circa 2010-2011



    1. The ratified version of 802.11n

    2. Digital Output (if we don't already have them by next refresh)

    3. Bluetooth 3.0 and Wimax support

    4. Voice to Text

    5. Dual cameras. 5MP photo and iSight for AV Chat.

    6. Improved location based services

    7. Specific Mac integration that blows our minds.



    That's all.
  • Reply 53 of 72
    mjtomlinmjtomlin Posts: 1,833member
    There's a reason Apple wanted this and spent so much time developing it. DirectX is a Windows-only tech. Why wouldn't they want to stop development of it? People always say how games run beter under Windows vs. Mac and yet they never mention that graphics cards manufacturers design and build their chips around DirectX, which only benefits Widows users.



    I always found it funny how Macs were always considered the "Graphics" computers even though the GPU acceleration was never there. A lot of people don't understand that Apple actually writes the graphics drivers for their computers and under Windows the GPU manufacturers write the drivers themselves, making sure to take advantage of all the features present on the cards.



    With OpenCL, this all takes a backseat... it allows them to write once and apply to just about any GPU, with very little overhead or programming modifications. However OpenCL is much more than just made to take advantage of GPUs, it designed to take advantage of every available processing unit... be it CPU, GPU, or DSP.
  • Reply 54 of 72
    mjtomlinmjtomlin Posts: 1,833member
    Quote:
    Originally Posted by MacTel View Post


    That would be assuming Apple's programmers are not doing the work that a dedicated Imagination Technologies programmer would be doing. Apple own 3.6% so sending a team to work within the company isn't out of the question since they have a stake in the company they have some say in its operations (some).



    Well the two times Apple was really interested in chips, they co-developed them...





    1. the ARM in 1988, along with Acorn (base design), and VLSI, they developed the modern versoin of the ARm chip for use in the Newton. This design is still in use today.



    2. the PowerPC; A subset of the IBM POWER CPU, Apple developed and designed the Alti-Vec processor, which was a 128-bit unit for vector processing to enhance a lot of multi-media instructions. Apple went with Motorola for the G4 because IBM refused to support Alti-Vec instructions until the G5. From there, we all know happened. I still prefer the PowerPC over the crappy Intel CPUs - multi-tasking was so much more fluid and smooth.
  • Reply 55 of 72
    mjtomlinmjtomlin Posts: 1,833member
    Quote:
    Originally Posted by mdriftmeyer View Post


    OpenCL will benefit Linux, OpenSolaris, FreeBSD, OpenBSD, NetBSD, et.al, as well. Those systems all now have Nvidia and ATi drivers.



    Well, not automatically. Programs and/or systems will have to be re-written to take advantage of it.
  • Reply 56 of 72
    Quote:



    = Marketing





    Wikipedia - there everyone can write. Yes you can use PSGL but the Developer are using normally the native API from the PS3.





    Marketing



    I can post also things from a forum:

    "Some of the consoles support mainstream APIs -- the Dreamcast could handle Direct3D (6?) and the PS3 can handle a modified OpenGL. However, nobody doing serious work on those consoles uses those APIs, but instead uses the console's native API."

    "It supports a modified OpenGL|ES calles PSGL. Nobody doing serious PS3 development uses it."

    http://209.85.129.132/search?q=cache...&client=safari





    Quote:

    From what I could find Nintendo's API is proprietary



    There are some things like in OpenGL



    Quote:
    Originally Posted by wobegon View Post


    Alright, I finally understand your point: While OpenGL is supposed to be platform independent, rather than proprietary and system-locked like DirectX is to Windows, thanks to the vendor-specific extensions required when writing OpenGL, you believe these vendor-specific extensions nullify OpenGL's "platform independent" promise. Am I right?



    Yes - I am sorry for my bad english.

    But that is the direction about I am speaking. And there was much hope for OpenGL3.



    Quote:

    If I'm right, that's an interesting point, but DirectX has only seen any major adoption on the Windows PC, not mobile-optimized devices, which is the focus of the article:



    Thats true. And on mobile devices OpenGL ES will be big. But don't forget OpenGL ES != OpenGL.





    Don't understand me wrong. I would like to see more OpenGL instead of Direct-X. But unfortunately in Games I don't believe it will happen, also because they failed with OpenGL3. Windows is the biggest market on the Desktop and DirectX > OpenGL. I hope it will change (but there was an opportunity and they didn't use it) and that maybe Apple is doing some pressure in the future to make a cleanup in OpenGL.
  • Reply 57 of 72
    jmmxjmmx Posts: 341member
    Quote:
    Originally Posted by Marvin View Post


    That's assuming that OpenCL can be used to deliver what OpenGL lacks vs DirectX.

    ...



    OpenCL doesn't look very easy to develop with compared to DirectX and OpenGL code though. It's specialized code and it's not something developers will be able to hit the ground running with..



    Marvin, I believe you have mossed the point here.



    In my [rather limited] understanding, openCL is NOT designed to be a development language for graphics apps. That is the task of the openGL. The C in openCL is for "Compute." The objective is to open up the GPU to tun other compute-intensive tasks - specifically linear-algebra and other mathematical tasks on the GPU's highly parallel math engine. Thus, the GPU can be harnessed to do physics computations for games, or for specific math or science computations. I am not sure where graphics tasks for say photoshop get done now, but they would certainly be able to be moved to the GPU if they are not there now. If you want to do graphic programming, then you use the openGL graphics API.



    The end of this all is that an next gen. iPhone/Touch will have amazing compute capabilities. I am sure it will absolutely blow anything else out of the water. It will be near desktop compute power in your hand. Also add in that one of the main objectives of Snow Leopard is to reduce the memory Paw Print... WOW!
  • Reply 58 of 72
    Quote:
    Originally Posted by jmmx View Post


    Marvin, I believe you have mossed the point here.



    In my [rather limited] understanding, openCL is NOT designed to be a development language for graphics apps. That is the task of the openGL. The C in openCL is for "Compute." The objective is to open up the GPU to tun other compute-intensive tasks - specifically linear-algebra and other mathematical tasks on the GPU's highly parallel math engine.



    Exactly right...



    ... but there is a limited class of tasks for which this is useful. OpenCL is far more useful on the desktop than on the phone.



    C.
  • Reply 59 of 72
    nceencee Posts: 836member
    - good or nor

    - better or not

    - the right move or not

    - for better or worst ?



    maybe it's just the next step to the NEXT best thing, which is only a step, to, you guessed it, the Next best thing. Which if I'm not mistaken, is the NEXT step to the next best thing.



    Technology changes by the second, minutes, hours, daily, monthly, yearly (that should cover most times for folks) ? oops, nano seconds.



    In any case, bring on the changes, however, whenever you see them as a good thing, and I, like many others, will line up to buy what at the time, WILL be the latest and greatest, next best ? phone, laptop, desktop, iPod, tablet and so on.



    Skip
  • Reply 60 of 72
    The article is talking about embedded, real-time operating systems, but in fact the mobile OS X is neither.



    It's not embedded, because it has become an open computing platform. It's not much different from the desktop Mac OS X anymore. The non-OS X iPods (Nano, Classic, Shuffle) do run embedded OSes but the Touch and the iPhone have full grown, desktop class operating systems.



    Second, the mobile OS X is not a real-time OS. It might be surprising for some, especially because Windows CE and Symbian OS are both real-time. However, the non-real-timeness of the mobile OS X is a real sign of a multi-purpose system. Real-time systems are typically built for a few very specific tasks, that are time critical (control systems, robot control, networking components, etc.). They are tricky to program properly, so that they don't break the real-timeness of the system, and they usually run few applications (or only one) at a time. Moreover, they have a larger processing overhead, so they consume more CPU time (and thus drain the battery faster).



    In case of Windows CE, the real-time design comes from the original purpose of the OS, which was embedded systems. In a PDA, CE's real-timeness has littleor no advantage.

    As for Symbian, I guess the reason why it is real-time is that originally the OS had to control the network traffic and handle the data flow which requires great time accuracy.

    However modern mobile devices have fast CPUs and some auxiliary special purpose hardware to handle real-time tasks under the control of a non-real-time OS. This makes the system much more versatile and efficient.
Sign In or Register to comment.