ARM Mac coming in first half of 2021, says Ming-Chi Kuo

245678

Comments

  • Reply 21 of 149
    crowleycrowley Posts: 10,453member
    I wonder if Apple would ship an ARM laptop with the same size battery, and therefore two or three day battery life, or if everything would get even thinner and lighter.
    mwhite
  • Reply 22 of 149
    red oakred oak Posts: 1,089member
    tht said:
    lkrupp said:
    Any ideas on how Apple will handle the X86 code of current apps to run on ARM architecture? I am not educated on this. Is ARM close enough to X86 that the transition will be easy or will it require a Rosetta-like translation framework like the move from Moto 68000 to X86 did. Will we have universal binaries again or something else during the transition?
    Microsoft has an X86 emulation layer for ARM Windows. It isn't great, but it's a solution.

    A Rosetta-like framework is the most likely.
    It can be a lot like the last time.

    1. Apps using Apple’s frameworks and Xcode will mostly be a recompile
    2. Apps that mostly use Apple’s frameworks, but not Xcode, or use 3rd party libraries, will use a binary translator like Rosetta (switching x86 instructions with ARM instructions at runtime)
    3. Apps that only work with macOS/x86 with 3rd party x86 libraries will need to be run in a macOS/x86 VM a la classic. If Apple provides it, presumably, apps can be overlapped, and it won’t be macOS run inside a window.

    And once again, the single biggest impediment to success will be Microsoft and Adobe moving their apps to macOS/ARM. If that doesn’t happen, macOS/x86 won’t be successful. Presumably, Apple will just do the work to bring FCPX and LPX over. Plugins may need to use Rosetta for awhile before they are moved over.
    Microsoft and (to a lesser extent) Adobe have invested heavily in developing iOS apps.   I don't think there will be a lot of work to make them native on an ARM Mac 


  • Reply 23 of 149
    blastdoorblastdoor Posts: 3,296member
    melgross said:
    lkrupp said:
    Any ideas on how Apple will handle the X86 code of current apps to run on ARM architecture? I am not educated on this. Is ARM close enough to X86 that the transition will be easy or will it require a Rosetta-like translation framework like the move from Moto 68000 to X86 did. Will we have universal binaries again or something else during the transition?
    This is the problem I’ve been wondering about for some time. While some people dismiss this as an issue, or in most cases, don’t even think about it (aren’t aware it is an issue), it’s the biggest issue apple will need to deal with. In previous changeovers, even Apple was very lax in getting their own big apps out. It took a year for them. It took a long time for Adobe and Microsoft, with their massive software, to come over too.

    ARM is not close to x86. It’s optimized for battery life over performance. Apple and ARM have made significant advances on that front, but the instruction sets are different enough. We know from previous attempts at emulation, that a processor family needs to be 5 times as powerful in order to be able to run software at the same speed as the family they’re emulating. This hasn’t changed. Microsoft supposedly does it now, with their “universal” sdk. But they don’t, really. They require software to be rewritten, and recompiled for ARM. And there have still been issues with performance, specific features and bugs.

    im not saying it can’t be done, because obviously it can. But if Apple is really going to release a device next year, there will either be significant limitations, or they’ve figured out a way around them. My suggestion, which no one here has ever commented on, from my memory, is to add a dozen x86 instructions to the chip. It’s been found that 80% of the slowdown between chip families is from about a dozen instructions. The chip, or OS, could hand that over to those when native x86 software needs them. Individual instructions aren’t patented, or copyrighted, as far as I know. If true, that would give Apple a way around the problem.
    Keep in mind that Office is already available for ARM-based iOS, Android, and Windows systems. 
    randominternetpersoncornchip
  • Reply 24 of 149
    crowleycrowley Posts: 10,453member
    tht said:
    tmay said:
    tht said:
    lkrupp said:
    Any ideas on how Apple will handle the X86 code of current apps to run on ARM architecture? I am not educated on this. Is ARM close enough to X86 that the transition will be easy or will it require a Rosetta-like translation framework like the move from Moto 68000 to X86 did. Will we have universal binaries again or something else during the transition?
    Microsoft has an X86 emulation layer for ARM Windows. It isn't great, but it's a solution.

    A Rosetta-like framework is the most likely.
    It can be a lot like the last time.

    1. Apps using Apple’s frameworks and Xcode will mostly be a recompile
    2. Apps that mostly use Apple’s frameworks, but not Xcode, or use 3rd party libraries, will use a binary translator like Rosetta (switching x86 instructions with ARM instructions at runtime)
    3. Apps that only work with macOS/x86 with 3rd party x86 libraries will need to be run in a macOS/x86 VM a la classic. If Apple provides it, presumably, apps can be overlapped, and it won’t be macOS run inside a window.

    And once again, the single biggest impediment to success will be Microsoft and Adobe moving their apps to macOS/ARM. If that doesn’t happen, macOS/x86 won’t be successful. Presumably, Apple will just do the work to bring FCPX and LPX over. Plugins may need to use Rosetta for awhile before they are moved over.
    Presumably, Apple will initially be delivering the "consumer" Mac's, a MacBook Air, and maybe a Mac Mini, analog with ARM, sans x86 compatibility, so users that require X86 will still have other Intel options. I would expect there to be overlap of ARM Mac and Mac Intel for quite some time, but also consider that AMD may replace Intel for some products.
    I think it will be desktops first. Start with the form factor with the least amount of sales for the customers who will want them first, namely developers. 
    I guess that works for a light launch, but that's the form factor that has the least to gain from such a move, as there's no power limitation on desktops.  Seems like Apple would need to hit with a bigger bang to get the momentum for such a migration.
    muthuk_vanalingam
  • Reply 25 of 149
    blastdoorblastdoor Posts: 3,296member

    red oak said:
    There is also (I believe) a substantial cost savings to Apple here 

    Assumptions: 

    -  Half of Macs use internally developed Apple CPU: 10M/yr    (the other half stays on Intel)
    -  Current cost Apple pays per Intel chip:  $200  (likely higher)
    -  Cost to Apple to manufacture new chip at TMSC:   $30 (likely lower)
    -  Dedicated Apple CPU chip employees and cost:   300 employees x $400K/yr fully weighted cost =  $120M 


    Total Intel Cost:  10M x $180 =  $2B
    Total internal Apple Cost:   10M x $30 =  $300M + $120M = $420M 


    Total Annual Savings:    $2B - $420M = $1.6 Billion 


    It would be $3.2 Billion if Apple were to move it all internal.   That would increase overall gross margin of the company by ~ 1%.    It would be a huge financial win.   But more importantly, it gets Apple untangled from the mess that is Intel

    Not included here is the massive R&D effort/spend to get to launch.   Maybe this is one of the driver's of Apple R&D spend exploding over the last 5 years 

    Agreed - -the savings are large. And the R&D effort is already being largely paid for by SOC development for iDevices. Just as OS/software R&D efforts cross-pollinate between iOS and macOS, so too could CPU/GPU R&D efforts cross-pollinate. We might see some innovations coming to the Mac SOC first -- things that require more power -- but then those could end up in an iPad or iPhone in a couple of years, after another die shrink. 
  • Reply 26 of 149
    red oakred oak Posts: 1,089member
    blastdoor said:
    red oak said:
    An "A15X" type 5 nm chip in a laptop is going to blow the doors off anything offered by Intel

    Batteries in the MBP are 2-2.5x the size of iPad Pros.   Plus the thermal envelope of the MBPs are much greater.    It will allow Apple to dramatically increase the number of cores plus boost the clock frequency.    It is going to be something to behold 

    This is the laptop I want 

    I'm not so sure. A TSMC 7nm A13X chip could probably blow the doors off Intel's 14nm Skylake-based lineup. But if Intel makes good on its roadmap of 7nm (which is more analogous to TSMC 5nm) in 2021, then it could be a much closer competition (especially with Sunny/Willow Cove cores and AVX512). 

    The big question is whether Intel will delay 7nm. They keep saying they're on track, but we've heard that before. 
    I'll be shocked if they deliver 7nm in volume in 2021 

    Even if performance is on par, battery life on an ARM Mac laptop is pointing to 20 hours (vs. 10 hours for current MBA and MBP).   The power management Apple would bring to this would be epic 
    razorpit
  • Reply 27 of 149
    crowleycrowley Posts: 10,453member
    What implications does an ARM CPU architecture have for the GPU?  Are we talking about an Apple-designed SOC for Mac laptops?
  • Reply 28 of 149
    red oakred oak Posts: 1,089member
    melgross said:

    lkrupp said:
    Any ideas on how Apple will handle the X86 code of current apps to run on ARM architecture? I am not educated on this. Is ARM close enough to X86 that the transition will be easy or will it require a Rosetta-like translation framework like the move from Moto 68000 to X86 did. Will we have universal binaries again or something else during the transition?
    Microsoft has an X86 emulation layer for ARM Windows. It isn't great, but it's a solution.

    A Rosetta-like framework is the most likely.
    Microsoft’s solution requires a software rewrite, and a recompile. It’s not a “real” solution. A Rosetta solution is just a quick, temporary hack, in the hope that developers will quickly move to new native apps, and a total rewrite of their old ones, and the expectation that newer hardware will be faster, and allow the Rosetta solution, which is half as fast, on an equivalent chip, to become less of a drag on performance. I remember all of Apple’s changeovers, and they were all pretty slow for about three years, when the new CPU’s caught up, and native software also took over.

    what’s been questioned about Apple making another changeover now, is whether major producers will again follow along. Anyone who says they will is just guessing, because they don’t know. I see moving to iOS, because of the number of devices out there, but Mac growth has stalled at just over 100 million installations around the world. Small developers might jump in, but large ones? It’s not a sure thing.

    what makes it more iffy is the way Apple will need to do this. They can just put their toe in the water, they can’t just jump in. So there will, at first, be no machines out there. Developers are reluctant to develop for what will be at first, a non existent market. They will wait. This is one of the reasons Windows phone died. Developers waited to see if it would sell in large numbers, and oeople would buy until the apps they wanted were there. Microsoft ended up paying developers to write for it, but it didn’t work.
    Mac installed base growth has not stalled.   Cook on last several earnings calls has consistently said that 50% of Macs are going to first time buyers.  And, that installed base of all major product groups are increasing.   So, roughly 9-10M new Mac users per year but we don't know how many are going from Mac to Windows 


    razorpit
  • Reply 29 of 149
    blastdoorblastdoor Posts: 3,296member
    red oak said:
    blastdoor said:
    red oak said:
    An "A15X" type 5 nm chip in a laptop is going to blow the doors off anything offered by Intel

    Batteries in the MBP are 2-2.5x the size of iPad Pros.   Plus the thermal envelope of the MBPs are much greater.    It will allow Apple to dramatically increase the number of cores plus boost the clock frequency.    It is going to be something to behold 

    This is the laptop I want 

    I'm not so sure. A TSMC 7nm A13X chip could probably blow the doors off Intel's 14nm Skylake-based lineup. But if Intel makes good on its roadmap of 7nm (which is more analogous to TSMC 5nm) in 2021, then it could be a much closer competition (especially with Sunny/Willow Cove cores and AVX512). 

    The big question is whether Intel will delay 7nm. They keep saying they're on track, but we've heard that before. 
    I'll be shocked if they deliver 7nm in volume in 2021 

    Even if performance is on par, battery life on an ARM Mac laptop is pointing to 20 hours (vs. 10 hours for current MBA and MBP).   The power management Apple would bring to this would be epic 
    Skepticism is warranted, given their past issues. 

    But... part of the problem with their 10nm process is that they were very aggressive in their scaling goals for a process that relied entirely on DUV, no EUV. Intel's 7nm process will be using EUV, so complexity will be reduced in some ways. Granted, EUV is new and complicated in its own right, but Intel won't be the first to use EUV -- TSMC will be using it a bit this year. So Intel might benefit from industry experience with EUV (specifically, ASML experience). It's no guarantee, but worth noting that the specific issues that drastically delayed 10nm aren't in play with 7nm. 
  • Reply 30 of 149
    thttht Posts: 5,450member
    crowley said:
    What implications does an ARM CPU architecture have for the GPU?  Are we talking about an Apple-designed SOC for Mac laptops?
    The vast majority of PC laptops uses on-die GPUs (processor graphics in Intel vernacular). Well, probably desktops too. This is the same arrangement that Apple has with their SoCs: on-chip GPUs. At minimum it will be performance competitive. For the high end pro machines, Apple will have to use discrete GPUs as they are now.

    GPU performance is basically a power game. Power as in Watts, and how many transistors they are willing to devote for the GPU. They are going to use their custom GPU with Imagination Tech IP as they are now. How performant they are will depend on what their transistor and power budgets are. Since they don’t need to sell the chip for profit, this theoretically lets them use 30 to 50% more transistors at the same cost they would buy one for from a 3rd party vendor. So, it could be more performant than on-die GPUs from Intel or AMD they choose to take advantage of this margin, or use less power for the same performance.
  • Reply 31 of 149
    melgrossmelgross Posts: 33,510member
    blastdoor said:
    melgross said:
    lkrupp said:
    Any ideas on how Apple will handle the X86 code of current apps to run on ARM architecture? I am not educated on this. Is ARM close enough to X86 that the transition will be easy or will it require a Rosetta-like translation framework like the move from Moto 68000 to X86 did. Will we have universal binaries again or something else during the transition?
    This is the problem I’ve been wondering about for some time. While some people dismiss this as an issue, or in most cases, don’t even think about it (aren’t aware it is an issue), it’s the biggest issue apple will need to deal with. In previous changeovers, even Apple was very lax in getting their own big apps out. It took a year for them. It took a long time for Adobe and Microsoft, with their massive software, to come over too.

    ARM is not close to x86. It’s optimized for battery life over performance. Apple and ARM have made significant advances on that front, but the instruction sets are different enough. We know from previous attempts at emulation, that a processor family needs to be 5 times as powerful in order to be able to run software at the same speed as the family they’re emulating. This hasn’t changed. Microsoft supposedly does it now, with their “universal” sdk. But they don’t, really. They require software to be rewritten, and recompiled for ARM. And there have still been issues with performance, specific features and bugs.

    im not saying it can’t be done, because obviously it can. But if Apple is really going to release a device next year, there will either be significant limitations, or they’ve figured out a way around them. My suggestion, which no one here has ever commented on, from my memory, is to add a dozen x86 instructions to the chip. It’s been found that 80% of the slowdown between chip families is from about a dozen instructions. The chip, or OS, could hand that over to those when native x86 software needs them. Individual instructions aren’t patented, or copyrighted, as far as I know. If true, that would give Apple a way around the problem.
    Keep in mind that Office is already available for ARM-based iOS, Android, and Windows systems. 
    It’s still a truncated version, and the iOS version is indeed native. But it’s for iOS, and not macOS. Therefor, it needs a lot of rewriting before it will work on macOS on ARM.
    edited February 2020 caladanian
  • Reply 32 of 149
    red oak said:
    There is also (I believe) a substantial cost savings to Apple here 

    Assumptions: 

    -  Half of Macs use internally developed Apple CPU: 10M/yr    (the other half stays on Intel)
    -  Current cost Apple pays per Intel chip:  $200  (likely higher)
    -  Cost to Apple to manufacture new chip at TMSC:   $30 (likely lower)
    -  Dedicated Apple CPU chip employees and cost:   300 employees x $400K/yr fully weighted cost =  $120M 


    Total Intel Cost:  10M x $180 =  $2B
    Total internal Apple Cost:   10M x $30 =  $300M + $120M = $420M 


    Total Annual Savings:    $2B - $420M = $1.6 Billion 


    It would be $3.2 Billion if Apple were to move it all internal.   That would increase overall gross margin of the company by ~ 1%.    It would be a huge financial win.   But more importantly, it gets Apple untangled from the mess that is Intel

    Not included here is the massive R&D effort/spend to get to launch.   Maybe this is one of the driver's of Apple R&D spend exploding over the last 5 years 

    Lots of assumptions/guesses there but I think your point is an under appreciated motivation for this.  The marginal cost will go from something like 200 a computer to 50.  All the other costs you describe are sunk costs that as another reader mentioned are being counted for already in development for iPhone iPad.

    So while the revenue difference is not huge it could allow Apple to price Macs closer to PCs while still maintaining the same per unit profit.  
    randominternetperson
  • Reply 33 of 149
    crowleycrowley Posts: 10,453member
    tht said:
    crowley said:
    What implications does an ARM CPU architecture have for the GPU?  Are we talking about an Apple-designed SOC for Mac laptops?
    The vast majority of PC laptops uses on-die GPUs (processor graphics in Intel vernacular). Well, probably desktops too. This is the same arrangement that Apple has with their SoCs: on-chip GPUs. At minimum it will be performance competitive. For the high end pro machines, Apple will have to use discrete GPUs as they are now.

    GPU performance is basically a power game. Power as in Watts, and how many transistors they are willing to devote for the GPU. They are going to use their custom GPU with Imagination Tech IP as they are now. How performant they are will depend on what their transistor and power budgets are. Since they don’t need to sell the chip for profit, this theoretically lets them use 30 to 50% more transistors at the same cost they would buy one for from a 3rd party vendor. So, it could be more performant than on-die GPUs from Intel or AMD they choose to take advantage of this margin, or use less power for the same performance.
    Really?  I was under the impression that Intel GPUs are integrated into the motherboard, but not part of the CPU?  Happy to be corrected if I'm wrong.

    But I guess my question was more about the discrete element, and whether moving to ARM would place restrictions on what discrete GPUs are available in Macs.  Are there impediments to getting drivers for an AMD X850 Megaballs or an nVidia Titania XL5000 on a version of macOS that runs on ARM, or is it something that'll happen pretty easily?
  • Reply 34 of 149
    asdasdasdasd Posts: 5,686member
    JinTech said:
    I imagine when someone "ports" their iPad over to the macOS via Catalyst it already has the code to run on ARM so it would not surprise me if Apple comes out and says "and if you develop for both iOS and macOS your apps will work right out of the bag!" With that being said, WWDC should be an interesting one this year. Maybe Apple will start encouraging developers who do not have an iOS app to bring their apps to that platform so it will be a seamless task to bring it back to the Mac once they have ARM processors?
    There’s no need for catalyst. All iOS apps now run on intel. In reverse a Mac app could easily run on arm. That’s just a recompile. 

  • Reply 35 of 149
    knowitallknowitall Posts: 1,648member
    lkrupp said:
    Any ideas on how Apple will handle the X86 code of current apps to run on ARM architecture? I am not educated on this. Is ARM close enough to X86 that the transition will be easy or will it require a Rosetta-like translation framework like the move from Moto 68000 to X86 did. Will we have universal binaries again or something else during the transition?
    Not such a big deal, it is possible to translate the binaries ahead of time and run them straight away.
    This doesn't require a recompilation of the source code (and doesn't require source code at all), people could just use the apps they have and use a translator (note that you translate once and run endlessly).
    It is likely Apple is aiming at that with its push to 64 bit Arm and Intel. This simplifies translation because 64 bit processors are a lot more similar, so apps can be translated easily.
    Universal binaries is an option but not a necessity for new apps (or new releases of apps) the compilation is for multiple platforms and the Appstore can deliver for the required platform (so no bloated apps).
  • Reply 36 of 149
    crowleycrowley Posts: 10,453member
    asdasd said:
    JinTech said:
    I imagine when someone "ports" their iPad over to the macOS via Catalyst it already has the code to run on ARM so it would not surprise me if Apple comes out and says "and if you develop for both iOS and macOS your apps will work right out of the bag!" With that being said, WWDC should be an interesting one this year. Maybe Apple will start encouraging developers who do not have an iOS app to bring their apps to that platform so it will be a seamless task to bring it back to the Mac once they have ARM processors?
    There’s no need for catalyst. All iOS apps now run on intel. In reverse a Mac app could easily run on arm. That’s just a recompile. 
    Porting from iOS to Mac would still require a UI transformation, which is what Catalyst offers (with mixed results).
    netmage
  • Reply 37 of 149
    knowitallknowitall Posts: 1,648member
    I expect Arm Macs this year.
    I bought my Arm desktop computer already.
    It’s a $44 rock64, add a 27inch 4K ips monitor, touch mouse (bluetooth) and bluetooth keyboard and it will top out at $400.
    I expect Macs to be priced similarly.
  • Reply 38 of 149
    asdasdasdasd Posts: 5,686member
    melgross said:
    lkrupp said:
    Any ideas on how Apple will handle the X86 code of current apps to run on ARM architecture? I am not educated on this. Is ARM close enough to X86 that the transition will be easy or will it require a Rosetta-like translation framework like the move from Moto 68000 to X86 did. Will we have universal binaries again or something else during the transition?
    This is the problem I’ve been wondering about for some time. While some people dismiss this as an issue, or in most cases, don’t even think about it (aren’t aware it is an issue), it’s the biggest issue apple will need to deal with. In previous changeovers, even Apple was very lax in getting their own big apps out. It took a year for them. It took a long time for Adobe and Microsoft, with their massive software, to come over too.

    ARM is not close to x86. It’s optimized for battery life over performance. Apple and ARM have made significant advances on that front, but the instruction sets are different enough. We know from previous attempts at emulation, that a processor family needs to be 5 times as powerful in order to be able to run software at the same speed as the family they’re emulating. This hasn’t changed. Microsoft supposedly does it now, with their “universal” sdk. But they don’t, really. They require software to be rewritten, and recompiled for ARM. And there have still been issues with performance, specific features and bugs.

    im not saying it can’t be done, because obviously it can. But if Apple is really going to release a device next year, there will either be significant limitations, or they’ve figured out a way around them. My suggestion, which no one here has ever commented on, from my memory, is to add a dozen x86 instructions to the chip. It’s been found that 80% of the slowdown between chip families is from about a dozen instructions. The chip, or OS, could hand that over to those when native x86 software needs them. Individual instructions aren’t patented, or copyrighted, as far as I know. If true, that would give Apple a way around the problem.
    I can’t emphasise enough that the vast majority of apps produced for the Mac right now will be a recompile and some won’t even need that. The compiler is doing the work if you use the Apple tool chain. 

    Dont confuse the compiled machine code with the higher level frameworks that might be used. 
    netmage
  • Reply 39 of 149
    Mike WuertheleMike Wuerthele Posts: 6,861administrator
    knowitall said:
    I expect Arm Macs this year.
    I bought my Arm desktop computer already.
    It’s a $44 rock64, add a 27inch 4K ips monitor, touch mouse (bluetooth) and bluetooth keyboard and it will top out at $400.
    I expect Macs to be priced similarly.
    There is right about zero percent chance of this. The RK3328 in there supports a whopping 4GB of RAM too.
    netmagemuthuk_vanalingamcornchip
  • Reply 40 of 149
    Rev2LivRev2Liv Posts: 7unconfirmed, member
    Seems quite ambitious given that Apple has very little desktop GPU experience relying on AMD For silicon for METAL3. 

    If I recall correctly, Sony had to enlist nVidia late into PS3 development after realizing the “Cell Broadband Engine” wasn’t capable of doing graphics.

    Moreover, Intel recruited Raj Khoduri of AMD and Apple fame to supercharge Intel’s graphics effort both in silicon and drivers.

    Why jump ship now unless this is a power play to extract concessions out of Intel.?
    edited February 2020
Sign In or Register to comment.