Intel dishes new details on Apple-bound Silverthorne chip

13

Comments

  • Reply 41 of 74
    wizard69wizard69 Posts: 13,377member
    Quote:
    Originally Posted by melgross View Post


    This is the first incarnation. It's a very good first step. Intel will quickly improve it.



    You never know with Intel. They completely lost interest in their ARM product and eventually dropped it.

    Quote:



    rememver that ARM has been this space for a long time. This is new. I hope that by now, no one here really doubts Intel's prowess.



    I also wouldn't go about second guessing what Intel is capable of. They have abandoned a lot of hardware over the years. ARM on the other hand has no choice.

    Quote:



    I also doubt very much that Stoke's "...is deal on arrival" (for now) will be true. There will be a lot of designs done around this, and products will arrive.



    That depends almost directly on what Stokes means by "MID". In some respects he could be right on. We all have this problem to some extent, I'm a strong advocate of tablets but every time they are brought up some idiot wants to talk about a 32" tablet. If Stokes takes the AIR to be a MID I would agree that it is not likely to end up there.



    In any event I believe you are right in that this chip will show up in all sorts of products. It could very well be one of the best embedded chips Intel has come up with in ages. I don't see them winning on the power front though.



    Lets face it ARM can be a process generation or two behind Intel and still win on power. Well power as in watts anyways.



    On thing that people seem to mis with Silverthorne is that it is a 64 bit chip. Frankly I see that as huge, I mean HUGE! I would be the first to agree that most new products to market won't leverage that feature but as a long term development path it is something ARM can't compete with. In a year or two Intel would very well have a SOC based on this product. That is a chip that integrates everything but memory. With improvements in memory (both Flash and RAM) it won't be long before the hand held space will be able to make use of that 64 bit address space.

    Quote:



    The problem for those products is that they will be the same old thing from the same old manufactures, and so THEY will be dead on arrival, as have all UMPC's been so far.



    Yep it is a product searching for a use. So is the AIR for that matter. I see the big potential for this sort of product to be in the sub UMPC arena, that is hand held devices.

    Quote:



    Apple might wait for the 32nm version next year (or, who knows, perhaps late this year). By then, it will be much more viable, and no other chipmaker can stay on the schedule Intel can.



    Oh I really hope not. Frankly I'm hoping Apple can get product out with this chip in the next few months. A lot of pressure for Apple but the potential here is huge. personally what I'm looking for is the slightly bigger that iPod Touch device. Something with maybe a 4 to 5.5 inch screen and a big ass battery.



    The big ass battery is important as I want the device to remain connected for up to we hours at a time. 12 hours is a good shift for me. Of course the battery is only part of the equation but that is to be expected. What I'm hoping for from Intel and Silverthorne is that they can truly address their side of the equation. Thus the interest in just how the power numbers are derived.



    Dave
  • Reply 42 of 74
    melgrossmelgross Posts: 33,510member
    Quote:
    Originally Posted by superkaratemonkeydeathcar View Post


    I understand your question completely.

    0-60 times & hp#s are still fun to know, even if you're comparing a Smart Two to a Gallardo Superleggera.



    melgross' problem is he has no life and the only way he can dhv or achieve any self-esteem is to be a gigantic fart-mongering condescension wonk here on AI, which is why he's not actually Listening to what you're saying.



    You should see his google images photo. --Hell I'd be a fart-wonking Mac tool if I looked like "The Librarian" from the original "Blade" movie, too.



    That was a brilliant post. I suppose you think that you are scintillating in your remarks, but really, you are not.
  • Reply 43 of 74
    melgrossmelgross Posts: 33,510member
    Quote:
    Originally Posted by JeffDM View Post


    That's not the same person.



    No, it's not.
  • Reply 44 of 74
    melgrossmelgross Posts: 33,510member
    Quote:
    Originally Posted by wizard69 View Post


    You never know with Intel. They completely lost interest in their ARM product and eventually dropped it.



    Yes, they did. For a good reason. Why produce more than one family? Now, everything will run the same codebase. That makes sense. It never did make sense for Intel to produce an ARM.



    Quote:

    I also wouldn't go about second guessing what Intel is capable of. They have abandoned a lot of hardware over the years. ARM on the other hand has no choice.



    A lot of what they did leave was hitting dead ends, or didn't fit into their main core of products. Apple does this all the time.



    Quote:

    That depends almost directly on what Stokes means by "MID". In some respects he could be right on. We all have this problem to some extent, I'm a strong advocate of tablets but every time they are brought up some idiot wants to talk about a 32" tablet. If Stokes takes the AIR to be a MID I would agree that it is not likely to end up there.



    It's what I think he's doing.



    Quote:

    In any event I believe you are right in that this chip will show up in all sorts of products. It could very well be one of the best embedded chips Intel has come up with in ages. I don't see them winning on the power front though.



    They don't have to win. That's what some people don't understand. They just have to get close. The advantages of using x86 will overcome any small disadvantage the chips will have.



    Quote:

    Lets face it ARM can be a process generation or two behind Intel and still win on power. Well power as in watts anyways.



    I don't know about that. Certainly two generations would put Intel way in front. One could be enough as well. Right now, this first gen chip is ok, but not fantastic. I think of it as a test generation. The next ones will be better, and on 32nm, will beat whatever ARM will be. I really don't doubt that. We can look at what Intel has done to AMD. Once they got off the idea that process alone would win, they came out with new designs that, when process was also considered, blew the pants off AMD. I can see the same thing happening here as well.



    Quote:

    On thing that people seem to mis with Silverthorne is that it is a 64 bit chip. Frankly I see that as huge, I mean HUGE! I would be the first to agree that most new products to market won't leverage that feature but as a long term development path it is something ARM can't compete with. In a year or two Intel would very well have a SOC based on this product. That is a chip that integrates everything but memory. With improvements in memory (both Flash and RAM) it won't be long before the hand held space will be able to make use of that 64 bit address space.



    Now you're getting the idea. And with Hyperthreading, it it the equiv of two cores, even though it's just a "look forward" design. At smaller sizes, Intel might go to two cores. And who knows, at some time, the may use out of order as well.



    Quote:

    Yep it is a product searching for a use. So is the AIR for that matter. I see the big potential for this sort of product to be in the sub UMPC arena, that is hand held devices.



    Well, that's what I'm saying.



    Quote:

    Oh I really hope not. Frankly I'm hoping Apple can get product out with this chip in the next few months. A lot of pressure for Apple but the potential here is huge. personally what I'm looking for is the slightly bigger that iPod Touch device. Something with maybe a 4 to 5.5 inch screen and a big ass battery.



    Maybe. by the way. That device you're talking about is the one that I was pushing as well.



    Quote:

    The big ass battery is important as I want the device to remain connected for up to we hours at a time. 12 hours is a good shift for me. Of course the battery is only part of the equation but that is to be expected. What I'm hoping for from Intel and Silverthorne is that they can truly address their side of the equation. Thus the interest in just how the power numbers are derived.



    Dave



    What I find interesting about these chips is something that Stokes mentioned, but that seems to have gone past people.



    While Intel is giving .5 to 2 watts as the TCP, the actual average draw may be far smaller. If so, its power usage may compete much more favorably with the ARM than people think.
  • Reply 45 of 74
    thttht Posts: 5,421member
    Quote:
    Originally Posted by backtomac View Post


    It's interesting that Silverthorne appears ill suited for MIDs (according to Stokes) and would seem underpowered for something like the MBA.



    What's the market for this chip?



    UMPC and UMPC-sized MIDs. I take an MID to be a device with 4 to 5 inch size screen (pocket mobile), while UMPCs are on the order of 5 to 8 inch screens (bag mobile), but Intel and Stokes probably have different definitions. So, an Asus Eee PC would be a good target. It certainly would be a good upgrade on the sub 1 GHz Celeron it uses. Then there are the untold number of candy bars, clamshell and slider form factor devices floating around that currently use the sub 800 MHz Intel A100/110 chips (ULV Pentium-M) which has 5 W TDP.



    So, it's a nice performance/watt upgrade for a lot of these UMPCs. Could be a 4x increase in performance per Watt.



    As for the MBA, well, it's an interesting proposition. A 2 GHz Silverthorn is 2 Watts TDP. The 1.8 GHz Merom special in the MBA is somewhere in the neighborhood of 20 Watts. The interesting question is, if a 2 GHz Silverthorne performs like a 1.6-1.8 GHz Core 2 Duo, at least in single-threaded apps, then Silverthorne in the MBA is quite interesting. Factor of 10 decrease in power consumption may mean an extra hour or two of battery life.



    Odds are that Silverthorne will not perform like a Core 2 Duo, but that's why the questions and answers have to be asked. Also, since Silverthorne has something like a 17 stage execution pipeline, hmm, Intel can probably crank this thing up to 4 GHz with 50 Watt TDP. The performance question would be interesting at those levels too.
  • Reply 46 of 74
    Quote:
    Originally Posted by wizard69 View Post


    You never know with Intel. They completely lost interest in their ARM product and eventually dropped it.



    Because it was (A) crap and (B) wasn't x86. The market cared about the first, and Intel the second. The XScale is a bloated implementation of an ARM without much potential for licensing... kind of goes against the whole point.



    Furthermore x86 is only tolerable when a million or so extra transistors can be added to perform microcode generation, but that's not really of concern for Intel, who have yet to find success in the embedded market. I am, simply put, not bullish on Silverthorne. It's a no-man's-land market, much like the Xscale's was, except that today there are better, faster, and more flexible ARM solutions which are capable of being spun into SoCs that can meet the needs of nearly all of the mobile computing market. The Silverthorne looks poised to cripple Transmeta's already dying [ed: dead] marketshare.
  • Reply 47 of 74
    wizard69wizard69 Posts: 13,377member
    Quote:
    Originally Posted by melgross View Post


    Yes, they did. For a good reason. Why produce more than one family? Now, everything will run the same codebase. That makes sense. It never did make sense for Intel to produce an ARM.



    I would argue against that but I don't think it makes much difference a this point.

    Quote:



    ......

    They don't have to win. That's what some people don't understand. They just have to get close. The advantages of using x86 will overcome any small disadvantage the chips will have.



    In the embedded space there is only a marginal advantage to X86. Smart phones and tablets though I consider computers. So I agree there is some advantages to the x86 in Apple products. When you consider the wide range of products that ARM slips into though x86 really has to work hard for acceptance because x86 doesn't mean squat.



    So what does all this have to do with Silverthorne? A lot actually because I don't see a few hand held devices from Apple supporting an entire product line. Which means that Silverthorne and the rest of the family needs to go into a lot of embedded devices besides Apple products. The problem is Intel will need to work real hard to gain acceptance in this market.

    Quote:







    I don't know about that. Certainly two generations would put Intel way in front. One could be enough as well. Right now, this first gen chip is ok, but not fantastic. I think of it as a test generation. The next ones will be better, and on 32nm, will beat whatever ARM will be. I really don't doubt that. We can look at what Intel has done to AMD. Once they got off the idea that process alone would win, they came out with new designs that, when process was also considered, blew the pants off AMD. I can see the same thing happening here as well.



    Ok this is something that needs to be worked out. The problem is of course to much or Intel playing the good info close to its chest. Thus we can't have a wise discussion. The only thing that can be pointed out is that there are a lot fo ARM SOC available off the shelf these days that run remarkably cool. That a process generation or two behind Intel. As far as I can see Intel hasn't even discussed a SOC that is equal in functionality to the SoC available from ARM suppliers. Silverthorne does real good but they still expect the use of an I/O support chip.

    Quote:





    Now you're getting the idea. And with Hyperthreading, it it the equiv of two cores, even though it's just a "look forward" design. At smaller sizes, Intel might go to two cores. And who knows, at some time, the may use out of order as well.



    I'd rather have two cores. Especially on something like a smart phone where one core could be effectively dedicated to a real time OS for baseband processing duties.

    Quote:



    Maybe. by the way. That device you're talking about is the one that I was pushing as well.



    Yes I do wish that Apple would stop screwing around and build the types of devices that many of us want. What is really frustrating is that things like Touch are so damn close it hurts. At least for a first gen device.

    Quote:



    What I find interesting about these chips is something that Stokes mentioned, but that seems to have gone past people.



    While Intel is giving .5 to 2 watts as the TCP, the actual average draw may be far smaller. If so, its power usage may compete much more favorably with the ARM than people think.



    The chips certainly are interesting but I'm not sure how quickly they will be adapted by Apple. I know that what I want in a hand held can be produced today on ARM and would do just about every thing I want. Silverthorne looks to be months out or maybe longer. Then there is the battery life concern, while you maybe convinced that it will be there with Silverthorne I'm not to sure.



    It might not sound like much but one watt difference for the same general performance would kill the intel processor. If people don't grasp this just have them grab onto a one watt flashlight bulb. If is still waste heat that has to be supplied via a battery.



    Don't get me wrong I do believe that Intel has something on their hands here. The question is how long until we can get our hands on product. I won't be happy if I have to wait until next year for an Apple tablet.



    Dave
  • Reply 48 of 74
    wizard69wizard69 Posts: 13,377member
    Quote:
    Originally Posted by Splinemodel View Post


    Because it was (A) crap and (B) wasn't x86. The market cared about the first, and Intel the second. The XScale is a bloated implementation of an ARM without much potential for licensing... kind of goes against the whole point.



    Yes but who's fault was it for that? Ok so it wasn't all Intels fault, but they never invested nor really tried to understand the market for embedded processors. They lost interest and direction with XScale and can do so with Silverthorne.

    Quote:



    Furthermore x86 is only tolerable when a million or so extra transistors can be added to perform microcode generation, but that's not really of concern for Intel, who have yet to find success in the embedded market.



    This is my concern, that is Silverthorne needs success in the embedded space if they expect to succeed. Apple producing a few iPhone or other tablets with this device won't justify its existence.

    Quote:

    I am, simply put, not bullish on Silverthorne. It's a no-man's-land market, much like the Xscale's was, except that today there are better, faster, and more flexible ARM solutions which are capable of being spun into SoCs that can meet the needs of nearly all of the mobile computing market.



    Intel's lack of ability to join the embedded space with Xscale needs to be addressed if they expect Silverhorne to win. They need to understand the need to be able to build SOC and other wise customize their processors. I continue to hear rumors about Apple and Intel working on a SOC of some sort maybe it will be a Silverthorne product. Trouble is Apple is big time money, ARM oon the other hand has all sorts of products that are on the market ready to roll for specific applications.

    Quote:

    The Silverthorne looks poised to cripple Transmeta's already dying [ed: dead] marketshare.



    I see Sliverthorne as possibly Intels last attempt to get into the embedded space. It won't be successful otherwise, if it is not adopted strongly outside of Apple. The good thing is that I see potential here, as Silverthorne does have some attractive qualities for embedded. Even those qualities don't mean a design in as intel hasn't demonstrated at all that they know how to support and sell to this market.



    Dave
  • Reply 49 of 74
    melgrossmelgross Posts: 33,510member
    Quote:
    Originally Posted by wizard69 View Post


    I

    In the embedded space there is only a marginal advantage to X86. Smart phones and tablets though I consider computers. So I agree there is some advantages to the x86 in Apple products. When you consider the wide range of products that ARM slips into though x86 really has to work hard for acceptance because x86 doesn't mean squat.



    The embedded space is automobiles, dishwashers, etc.



    That is an entirely different market than we are interested in here.



    Our market of UMPC's, phones, iPods, Ultra lights, etc, is not the embedded market. It is the computer market, in one way or the other, with the posibility of installing purchased programs from companies who are moving these programs from one form factor to others.



    The x86 does have great importance there.



    Quote:

    So what does all this have to do with Silverthorne? A lot actually because I don't see a few hand held devices from Apple supporting an entire product line. Which means that Silverthorne and the rest of the family needs to go into a lot of embedded devices besides Apple products. The problem is Intel will need to work real hard to gain acceptance in this market.



    I don't understand you point here. The entire purpose of Silverthorn and its successors is to be code compatable with x86. Intel made that very clear. Companies using it will do so because of that.



    If it also ends up in embedded products, that's more wins for Intel, but has nothing to do with our interest in it.



    Quote:

    Ok this is something that needs to be worked out. The problem is of course to much or Intel playing the good info close to its chest. Thus we can't have a wise discussion. The only thing that can be pointed out is that there are a lot fo ARM SOC available off the shelf these days that run remarkably cool. That a process generation or two behind Intel. As far as I can see Intel hasn't even discussed a SOC that is equal in functionality to the SoC available from ARM suppliers. Silverthorne does real good but they still expect the use of an I/O support chip.



    As I say, this is first generation, and it may run cooler than it seems to, from Intel's specs. We'll see when it arrives.



    Quote:

    I'd rather have two cores. Especially on something like a smart phone where one core could be effectively dedicated to a real time OS for baseband processing duties.



    Effectively, there might be little difference. Right now, the iPhone has three processors.



    Quote:

    Yes I do wish that Apple would stop screwing around and build the types of devices that many of us want. What is really frustrating is that things like Touch are so damn close it hurts. At least for a first gen device.



    I'd like to see it as well. But I don't want it rushed. I'd rather Apple spent some time perfecting it. Lately, they seem to be releasing products that are, in some way, not completed. That's a bad trend.



    Quote:

    The chips certainly are interesting but I'm not sure how quickly they will be adapted by Apple. I know that what I want in a hand held can be produced today on ARM and would do just about every thing I want. Silverthorne looks to be months out or maybe longer. Then there is the battery life concern, while you maybe convinced that it will be there with Silverthorne I'm not to sure.



    It might not sound like much but one watt difference for the same general performance would kill the intel processor. If people don't grasp this just have them grab onto a one watt flashlight bulb. If is still waste heat that has to be supplied via a battery.



    Don't get me wrong I do believe that Intel has something on their hands here. The question is how long until we can get our hands on product. I won't be happy if I have to wait until next year for an Apple tablet.



    Dave



    We don't know what Apple is planning. I don't want them to come out with a product that has an ARM today, and then switch to x86 later in the year. That would confuse, and frustrate, developers. I'd rather they waited until everything was just right. I'm sure they knew about this well before the announcement, as they must have known about Yonah and Core when Jobs made the announcement about moving to x86 before Apple switched.
  • Reply 50 of 74
    Quote:
    Originally Posted by melgross View Post


    The embedded space is automobiles, dishwashers, etc.



    That is an entirely different market than we are interested in here.



    Our market of UMPC's, phones, iPods, Ultra lights, etc, is not the embedded market. It is the computer market, in one way or the other, with the posibility of installing purchased programs from companies who are moving these programs from one form factor to others.



    Don't get hung up on semantics. We are taking about the same thing. Moreover, the phone & device industries are still, generally, regarded as "embedded," some systems more-so than others. Basically, if the device is not intended to have the user to install software not provided by the developer, it falls under the "embedded" umbrella. On the cusp are devices like the iPhone or iPod Touch, which were obviously designed with extensibility, but haven't yet taken advantage of that. I would argue, though, that the iPod Nano is very much an embedded device.





    Quote:
    Originally Posted by melgross View Post


    The x86 does have great importance there. ...



    Effectively, there might be little difference. Right now, the iPhone has three processors.



    In regard to code compatibility, the problem I see with Silverthorne is that its architecture is very different than that of the x86's in PCs. I expect that most people interested it developing for Silverthorne UMPCs will STILL have to come out with alternate versions of their software in order to meet the ergonomic factors of the smaller devices -- no successful mobile device in history has used desktop apps as-is. Accordingly, these developers aren't terribly disadvantaged by using an ARM cross compiler rather than a Silverthorne cross compiler.





    On the iPhone: it probably has more than three. I think I just bought a pair of socks with ARMs in them.
  • Reply 51 of 74
    wizard69wizard69 Posts: 13,377member
    Quote:
    Originally Posted by melgross View Post


    The embedded space is automobiles, dishwashers, etc.



    That is an entirely different market than we are interested in here.



    Yes it is different but is important for the Silverthorne family in the long run.

    Quote:



    Our market of UMPC's, phones, iPods, Ultra lights, etc, is not the embedded market. It is the computer market, in one way or the other, with the posibility of installing purchased programs from companies who are moving these programs from one form factor to others.



    Yes that I understand, what I'm saying is that that market is not large enough to sustain a product family like Silverthorne. It is the same issue we had with PPC, not enough usage in industry to justify its development at a rapid pace. Silverthorne is enough of a departure from Intels main line up that development will have to be justified on its ability to sell in the market place. Thus Intel really needs to have a plan in place for this family in the embedded space, something they have little skill with right now.



    I don't see the idea of moving programs from one form factor to another to be a big deal. The programs that work well on hand held devices will be those written for the platform.



    Quote:



    The x86 does have great importance there.



    Well yes and no. It is only important if the hardware can maintain its self against the competition and sell enough that the whole works doesn't get canceled. The getting canceled part should worry anybody that has worked with Intel as they expect volumes that most embedded users can never achieve.

    Quote:





    I don't understand you point here. The entire purpose of Silverthorn and its successors is to be code compatable with x86. Intel made that very clear. Companies using it will do so because of that.



    I thought I was pretty clear here in that I don't believe Apple will ever have the volume to allow Intel to justify development of Silverthorne long term. Intel needs to address multiple markets with this device and give potential users some confidence that the family will be around for awhile.

    Quote:



    If it also ends up in embedded products, that's more wins for Intel, but has nothing to do with our interest in it.



    See this is where we disagree as I believe it has everything to do with our interests. Intel needs to sell a lot ore of these than Apple can ever move. To do so they need to look all the potential markets and address concerns specific to those markets. Embedded is a good place for this chip but the question is will Intel strive to offer up the assurances that the embedded world demands?



    Now you can argue that there are any number of hand held PC and cell phone applications for this processor and that Apple will not be sole user. That is very possible but I don't see this family as being the easy design in that some would like it to be.

    Quote:



    As I say, this is first generation, and it may run cooler than it seems to, from Intel's specs. We'll see when it arrives.



    Yeah I'm hoping so myself but I don't want Apple or anybody else for that matter trading off x86 compatibility for battery life. Especially when the ARM world is just getting better and better at the low power jig.

    Quote:





    Effectively, there might be little difference. Right now, the iPhone has three processors.







    I'd like to see it as well. But I don't want it rushed. I'd rather Apple spent some time perfecting it. Lately, they seem to be releasing products that are, in some way, not completed. That's a bad trend.



    This I very much agree with. I took them almost half a year to get the Mobile X apps running in user space. That is very sad indeed. All that time and we still don't have a SDK, a full Bluetooth stack nor a really stable suite of apps.

    Quote:



    We don't know what Apple is planning. I don't want them to come out with a product that has an ARM today, and then switch to x86 later in the year. That would confuse, and frustrate, developers.



    It would to some extent but then again Apple has been doing that for years. At one time I thought the delay in the SDK was to make sure the x86 based devices where on the market. Now I'm not to sure.



    I look at it this way, if there is money to be made, the developers will write code for whatever is out there.

    Quote:

    I'd rather they waited until everything was just right. I'm sure they knew about this well before the announcement, as they must have known about Yonah and Core when Jobs made the announcement about moving to x86 before Apple switched.



    I very much sure they know what is up at Intel. There are Intel spokes people indicating that Apple gave them a lot of insight into what is needed in these low power processors. Plus there are rumors afloat that Intel was working on custom hardware for Apple. The question in my mind was that the AIR's processor or a custom Silverthorne chip?



    All I know is that Apple needs new products quick as their stock is getting hammered hard and frankly justifiably so. MWSF simply did not have the types of hardware debuts that even justified Apples time at the event.



    DAve
  • Reply 52 of 74
    wizard69wizard69 Posts: 13,377member
    Quote:
    Originally Posted by Splinemodel View Post


    Don't get hung up on semantics. We are taking about the same thing. Moreover, the phone & device industries are still, generally, regarded as "embedded," some systems more-so than others.



    For the purposes of this discussion I think the distinction is OK. The reasoning is that I do not see one class of devices form Apple supporting the development of a Family like Silverthorne is a member of. So Intel needs to obtain market share outside of the sorts of devices that Apple will put the processor into.

    Quote:

    Basically, if the device is not intended to have the user to install software not provided by the developer, it falls under the "embedded" umbrella. On the cusp are devices like the iPhone or iPod Touch, which were obviously designed with extensibility, but haven't yet taken advantage of that.



    I would simply classify those devices as computers. Even if Apple is not up to so classifying them.

    Quote:

    I would argue, though, that the iPod Nano is very much an embedded device.



    I have to wonder if there are others that see Silverthorne in the same manner I do. That is that there is a risk of it following the same foot steps that PPC took. That is slow development and eventual lack of interest upon the part of the suppliers. I guess the good thing in this respect is that people seem to be switching over to smart phones rather quickly. Maybe Intel has got the timing right on this processor and the demand it will see in the consumer space.



    Dave
  • Reply 53 of 74
    melgrossmelgross Posts: 33,510member
    Quote:
    Originally Posted by Splinemodel View Post


    Don't get hung up on semantics. We are taking about the same thing. Moreover, the phone & device industries are still, generally, regarded as "embedded," some systems more-so than others. Basically, if the device is not intended to have the user to install software not provided by the developer, it falls under the "embedded" umbrella. On the cusp are devices like the iPhone or iPod Touch, which were obviously designed with extensibility, but haven't yet taken advantage of that. I would argue, though, that the iPod Nano is very much an embedded device.



    It's not semantics. An embedded device is one that comes programmed. You can't do anything to it, such as adding programs. What we are talking about are NOT embedded devices.



    Phones that are not open to other programs can be thought of as embedded, but not smartphones.



    Otherwise, you micht as well consider laptop cpu's as being embedded as well. I don't think people will do that.



    Quote:

    In regard to code compatibility, the problem I see with Silverthorne is that its architecture is very different than that of the x86's in PCs. I expect that most people interested it developing for Silverthorne UMPCs will STILL have to come out with alternate versions of their software in order to meet the ergonomic factors of the smaller devices -- no successful mobile device in history has used desktop apps as-is. Accordingly, these developers aren't terribly disadvantaged by using an ARM cross compiler rather than a Silverthorne cross compiler.



    It really isn't. As Intel says, it's code compatible. Other writers on tech sites have said the same after seeing Intel's presentations. I see no point in arguing that.



    If an individual device limits the programs, that's not because of the chips, it's because of the device the chips are in. That's very different.



    Quote:

    On the iPhone: it probably has more than three. I think I just bought a pair of socks with ARMs in them.



    I get the joke, but three is what's been seen upon teardown.
  • Reply 54 of 74
    melgrossmelgross Posts: 33,510member
    Quote:
    Originally Posted by wizard69 View Post


    Yes it is different but is important for the Silverthorne family in the long run.



    You are right, but it doesn't affect us, except to make the chips more likely to succeed.



    Quote:

    Yes that I understand, what I'm saying is that that market is not large enough to sustain a product family like Silverthorne. It is the same issue we had with PPC, not enough usage in industry to justify its development at a rapid pace. Silverthorne is enough of a departure from Intels main line up that development will have to be justified on its ability to sell in the market place. Thus Intel really needs to have a plan in place for this family in the embedded space, something they have little skill with right now.



    AsI said above, it will crtainly help. But I don't think it will be required. We will see, in a very few years, tens of millions of devices using these chips other than those in the embedded space. That would be enough. Don't forget that Intel's market consists of a number of different chips, each of which doesn't sell more that several tens of millions a year.



    Quote:

    I don't see the idea of moving programs from one form factor to another to be a big deal. The programs that work well on hand held devices will be those written for the platform.



    It matters. the less work is involved in writing programs, the more likely we will see them on other products. Code can be reused for several different platforms based on the same chips. The entire program doesn't have to go, just parts.



    Quote:

    Well yes and no. It is only important if the hardware can maintain its self against the competition and sell enough that the whole works doesn't get canceled. The getting canceled part should worry anybody that has worked with Intel as they expect volumes that most embedded users can never achieve.



    It's because it's an x86 design that it won't get cancelled.



    Quote:

    I thought I was pretty clear here in that I don't believe Apple will ever have the volume to allow Intel to justify development of Silverthorne long term. Intel needs to address multiple markets with this device and give potential users some confidence that the family will be around for awhile.



    I never said "Apple". I said "companies". That includes Apple, but doesn't exclude others. I'm pretty sure that every company now making a UMPC is looking very closely at this. We may see the slightly larger devices that we both want come out on this.



    Quote:

    See this is where we disagree as I believe it has everything to do with our interests. Intel needs to sell a lot ore of these than Apple can ever move. To do so they need to look all the potential markets and address concerns specific to those markets. Embedded is a good place for this chip but the question is will Intel strive to offer up the assurances that the embedded world demands?



    We're not really disagreeing. We're stressing different areas, but we agree on principal. You can see by my above replies. I think that embedded parts will swell Intel's sales, and help development, but I think that it may not be required.



    But remember that embedded needs are somewhat different from those of the devices we want to see. Look at some of the problems Apple has in its last years of using the G4, which was by then, being designed for the embedded market, and wasn't entirely suitable for computers.



    Quote:

    Now you can argue that there are any number of hand held PC and cell phone applications for this processor and that Apple will not be sole user. That is very possible but I don't see this family as being the easy design in that some would like it to be.



    I think we can both argue that Apple won't be the sole user.



    Why do you think it won't be fairly easy?



    Quote:

    Yeah I'm hoping so myself but I don't want Apple or anybody else for that matter trading off x86 compatibility for battery life. Especially when the ARM world is just getting better and better at the low power jig.



    They'll all get better. but Intel will have an increasing advantage, fueled by this being an x86. I never though I'd be saying that, but it's true.



    Quote:

    This I very much agree with. I took them almost half a year to get the Mobile X apps running in user space. That is very sad indeed. All that time and we still don't have a SDK, a full Bluetooth stack nor a really stable suite of apps.



    It seems slow, but it's happening. I have bright thought for the future.



    Quote:

    It would to some extent but then again Apple has been doing that for years. At one time I thought the delay in the SDK was to make sure the x86 based devices where on the market. Now I'm not to sure.



    I said here, from the very beginning, that Apple would have an SDK. I was sure of it. I hoped it would be sooner, but at least it's arriving in a couple of weeks (unless it's delayed).



    Quote:

    I look at it this way, if there is money to be made, the developers will write code for whatever is out there.



    You're absolutely right. But making money reflects the total costs involved. Writing code from scratch, if that program is specialized, and therefore won't sell many copies, and must be sold for a low price, might be the last straw on the camel's back, so to speak.



    Quote:

    I very much sure they know what is up at Intel. There are Intel spokes people indicating that Apple gave them a lot of insight into what is needed in these low power processors. Plus there are rumors afloat that Intel was working on custom hardware for Apple. The question in my mind was that the AIR's processor or a custom Silverthorne chip?



    Of course. Not a Silverthorn. It's a low power Core 2 Duo.



    Quote:

    All I know is that Apple needs new products quick as their stock is getting hammered hard and frankly justifiably so. MWSF simply did not have the types of hardware debuts that even justified Apples time at the event.



    DAve



    Yeah. My daughter is still waiting for a new 24" iMac, and I'm waiting to see new monitors.
  • Reply 55 of 74
    wizard69wizard69 Posts: 13,377member
    Quote:
    Originally Posted by melgross View Post


    It's not semantics. An embedded device is one that comes programmed. You can't do anything to it, such as adding programs. What we are talking about are NOT embedded devices.



    The embedded space is actually much bigger than what you describe. Many an embedded device can have is program changed or updated.

    Quote:

    Phones that are not open to other programs can be thought of as embedded, but not smartphones.



    I like ot think of smart phones as small computers. If we could get this across to manufactures we might be able to get the sorts of open devices we want in the way of hand held devices.

    Quote:



    Otherwise, you micht as well consider laptop cpu's as being embedded as well. I don't think people will do that.



    It really isn't. As Intel says, it's code compatible. Other writers on tech sites have said the same after seeing Intel's presentations. I see no point in arguing that.



    Err yes there is. The Silverthorne will execute the same code as the rest of Intels hardware but that is not the same thing as saying it will be optimal. Considering the limitations of the execution environment it might take a lot of work to a code base to get acceptable performance. While this isn't a case of running a completely different instruction set it none the less highlights that the core of Silverthorne is very different than modern Intel processors.

    Quote:



    If an individual device limits the programs, that's not because of the chips, it's because of the device the chips are in. That's very different.



    I get the joke, but three is what's been seen upon teardown.



    Dave
  • Reply 56 of 74
    wizard69wizard69 Posts: 13,377member
    Quote:
    Originally Posted by melgross View Post


    It matters. the less work is involved in writing programs, the more likely we will see them on other products. Code can be reused for several different platforms based on the same chips. The entire program doesn't have to go, just parts.



    Well first most software is written in C/C++ or a high level language these days so that kinda nixes issues with code compatibility. Second machine specific code will execute very differently on Silverthorne so what is optimized for one processor might suck on Silverthorne.

    Quote:



    It's because it's an x86 design that it won't get cancelled.



    I don't see that as even a remote argument. Intel has canceled a lot of x86 processors over the years. Netburst was dropped like a rock and that is a recent example. Silverthorne will stay around if it is profitable that is about it. To be profitable it needs to demonstrate that it is better in some ways than ARM.



    I don't think that will be difficult for Intel to do for certain classes of devices. I'm not convinced that it will be a success in the smallest hand held devices though.

    Quote:



    I never said "Apple". I said "companies". That includes Apple, but doesn't exclude others. I'm pretty sure that every company now making a UMPC is looking very closely at this. We may see the slightly larger devices that we both want come out on this.



    This would almost be perfect of ASUS's Eee PC. There are a number of low power platforms this could power in the future. A Newton 2 should be high on Apples wish list. Ok so that is my wish list but you get the idea.

    Quote:





    We're not really disagreeing. We're stressing different areas, but we agree on principal. You can see by my above replies. I think that embedded parts will swell Intel's sales, and help development, but I think that it may not be required.



    I think it is a question of how strongly the processor gets adopted outside of Apple. The problem is Intel will have to work hard for that market, it isn't a shoe in.

    Quote:



    But remember that embedded needs are somewhat different from those of the devices we want to see. Look at some of the problems Apple has in its last years of using the G4, which was by then, being designed for the embedded market, and wasn't entirely suitable for computers.



    The G4 was perfectly suitable for computers. The problem was that the market wasn't large enough to justify the R & D investment. So it quickly fell behind the rapid development on i86. I'm hoping that Silverthorne doesn't fall into this abyss.



    By the way in many ways Silverthorne is very similar to older processors. The in order execution being one feature that is a throw back to the past. In fact I suspect that an old G4 would perform on a par with Silverthorne.

    Quote:



    I think we can both argue that Apple won't be the sole user.



    I would hope so just to make the product a success.

    Quote:



    Why do you think it won't be fairly easy?



    Well for phones every single milliwatt makes a difference. It is still not clear how this processor stacks up against ARM power (as in watts) wise. There is a big difference between 500 milliwatts from Silverthornes processor and 450 milliwatts from a ARM SOC. Remember that Silverthorne apparently still needs a support chip so that is additional power.



    At this point in time the processor is still a two chip deal where ARM has very notable single chip solutions. Board space is a big deal for cell phones.



    On bigger devices Silverthorne quickly gains an advantage. Simply because there is enough board space and power available to make up for Silverthornes current fat.

    Quote:

    They'll all get better. but Intel will have an increasing advantage, fueled by this being an x86. I never though I'd be saying that, but it's true.



    To an extent it is true sure. I think the big factor for forward looking managers is that it is a 64 bit platform. We won't be running out of address space with this guy at all.

    Quote:



    It seems slow, but it's happening. I have bright thought for the future.



    The big problem if you want to buy into this platform is that there is nothing from Apple to indicate where they are going with the platform. For example I mentioned Bluetooth which needs to support more profiles, there is no indication form Apple at all that this will happen. With the looming SDK this really limits what one can design for the platform. Apple should simply come clean with a development road map so that we can determine if the Touch series devices will be able to support our hoped for apps.

    Quote:



    I said here, from the very beginning, that Apple would have an SDK. I was sure of it. I hoped it would be sooner, but at least it's arriving in a couple of weeks (unless it's delayed).



    The SDK is an important element but we also need to know where they are going with the OS. I mentioned Bluetooth above but there is a lot of other work that should be done. All of these things being resolved would turn the iPhone into a real platform to deliver apps.

    Quote:





    You're absolutely right. But making money reflects the total costs involved. Writing code from scratch, if that program is specialized, and therefore won't sell many copies, and must be sold for a low price, might be the last straw on the camel's back, so to speak.



    Of course. Not a Silverthorn. It's a low power Core 2 Duo.



    What I was trying to say and apparently missed the boat is that the rumors about Intel involved in custom chips for Apple, did they mean the processor in AIR or something else. Is that something else an Apple specific Silverthorn chipset?



    I've heard the rumors but of course there is little information. The rumor may simply have been about the custom work done for AIR or not.

    Quote:



    Yeah. My daughter is still waiting for a new 24" iMac, and I'm waiting to see new monitors.



    Monitors to me are not that big of a deal. At least not today. I might be needing a bigger monitor in the future but that is old age. What I want has a monitor that fits into the palm of my hand. It should also do Cell or WiMax communications and be cheap. The last thing Apple needs to do is to introduce another over priced machine like AIR or the 32 GB Touch.



    Dave
  • Reply 57 of 74
    jeffdmjeffdm Posts: 12,951member
    Quote:
    Originally Posted by wizard69 View Post


    I don't see that as even a remote argument. Intel has canceled a lot of x86 processors over the years. Netburst was dropped like a rock and that is a recent example.



    I don't think that's comparable. Intel has cancelled a lot of different x86 projects, they seem to have too many anyway. Still, Netburst was used over four process generations (180, 130, 90 & 65nm) and was produced for about six years. I really don't see these very low power chips going away soon.
  • Reply 58 of 74
    wizard69wizard69 Posts: 13,377member
    Quote:
    Originally Posted by JeffDM View Post


    I don't think that's comparable. Intel has cancelled a lot of different x86 projects, they seem to have too many anyway. Still, Netburst was used over four process generations (180, 130, 90 & 65nm) and was produced for about six years. I really don't see these very low power chips going away soon.



    Well I don't think anybody wants to see them go away. The problem is does intel have a plan in place to sell enough of them to make production likely into the far future? Or to look at it another way how many of these does intel have to sell to be considered successful and how many do they have to sell to keep the R & D machine for them going.



    That is in a nutshell the concern. Some people see a huge adoption rate but frankly I just don't know. Intel probably will know in three years.



    Dave
  • Reply 59 of 74
    Only win mobile!
  • Reply 60 of 74
    UP2U - Magazine and eBook Online Borrowe Website,FHM,Penthouse,PCmagazine,Vogue,Men health:

    http://www.up2u.in/
Sign In or Register to comment.