Apple looking to develop custom ARM chips for future Macs, cutting out Intel - report

123468

Comments

  • Reply 101 of 159
    sog35 said:
    Here. We. Go.

    I think we will be seeing an iOS desktop very soon..

    Hopefully in 5 years Apple is almost 100% vertically integraded and does not have to rely on any vendor to provide parts.
    Would that be a touch screen laptop?
    ... They already nixed that idea.

    Well, until they added a keyboard to the iPad.  
    Then they added a file system to the iPad.
    ...  Can you guess what's coming next?
  • Reply 102 of 159
    nhtnht Posts: 4,214member
    asdasd said:


    karmadave said:
    Here we go again. Apple is NOT abandoning Intel on the Mac! I wouldn't be surprised however to see an AMD processor show up, in the Mac, at some point.

    MacOS = Intel/x86
    iOS = ARM

    Any questions?
    I'd be surprised.  Apple gains nothing by going with AMD.   I don't think Apple abandons Intel entirely but they are likely going to shift a majority of their high volume hardware running Mac OS to ARM within a couple of years. 


    It depends on how many dual boot or use VMWare or parallels. The Mac book is becoming increasingly the dev machine of choice as you can use it to build iOS, Android and Windows products. Lots of people in fact like the hardware and run Windows primarily. Intel is not going away on the high end, and at the moment it is all high end. 

    Will they re-introduce the Mac mini with ARM? Possibly. 
      
    VMware and Parallels are passé running VM on your local machine still happens but all the movement in development is in containers which often are accessible in the cloud (Docker, Kubernetes etc).  The Cloud has never really been the driver it's the fast broadband access to the cloud that is propelling new ways of development.   Where I live they've been rolling out Gigabit internet speeds which makes accessing containers fast and flexible. 
    Its not passed for enterprise users who need to run excel "apps", ms project and Visio locally.  If I wasn't able to open Windows Excel to do all the earned value, costing and other stuff that's embedded into Excel workbooks that don't run on the Mac I wouldn't be able to get a Mac for work.  If I'm locked into always logging into a windows VM to run these apps I might as well not own a Mac either.
    cornchipasdasd
  • Reply 103 of 159
    nhtnht Posts: 4,214member

    I don't think there's much doubt that Windows still rules the biz sector but I agree with 
    Sog35 it's simply  not that important to Apple's future that they ship computers with 
    Intel Hardware. 

    70% of their Rev comes from iOS. 
    So what? We're talking Macs.  Macs are huge in enterprise now.  Removing x86 business apps would crater Mac enterprise sales which is a growth area for Macs:

    For the increase in Mac and iOS adoption now at 91 and 99 percent, Jamf notes that 74% of organizations saw an increase in Mac adoption and 76% an increase in iPhone and iPad adoption in 2016 versus the year before. IBM has now almost reached its goal of deploying 100,000 Macs, the report confirms, making it the largest company Mac deployment. IBM, which uses Jamf software to manage its deployment, first announced the goal during the JAMF Nation User Conference (JNUC) last October. At the time, IBM said it was saving on average a minimum of $265 per Mac versus a comparable PC due the cost of device itself, OS, support, resale value and deployment.

    https://daringfireball.net/linked/2017/03/07/jamf-apple-enterprise

    Especially as Apple positions iPad Pros for lighter workloads.
    tmaywilliamlondon
  • Reply 104 of 159
    nhtnht Posts: 4,214member


    As far as performance ARM has beefed up their architecture.  You can now have up to 8 CPU in a cluster and you can cluster 32 
    of these Octoclusters together.   Want more speed?  Add more CPU. 
    Doesn't work that way for many apps which are constrained by single threaded performance because the expensive computations aren't easily parallellizable.

    Also multiple CPUs has architecture costs vs multiple cores.  There's no free lunch.

    I have no idea why folks will look at the A series of processors and then claim that by adding more cores you can outperform the faster processor when the older quad core A series chips regularly beat the pants off the hexacore and octocore chips that Samsung and Qualcomm made and the A11 will demolish everything Samsung or Qualcomm will have because its single core performance is so much better.  Samsung could make a dual processor octacore phone (16 cores) and still lose to the A11 in usable performance.

    Those performance advantages don't simply disappear because it's Intel desktop quad core i7 vs any ARM octacore, even Apple's.  Intel has been working extremely hard to improve performance per watt to the point we have highly performant 4.5W Intel processors.  The dual core Intel Core i7-7Y75 MacBook scores 4020 single core and 7999 multicore on 4.5W TDP.  That's on par with the A11 but only using 2 cores and not 4+2.  All those folks that claim Intel has done nothing in the last few years is ignoring that Intel has been doing EXACTLY what Apple has wanted them to do:  concentrate on performance per watt over top end speed.

    Essentially Intel has caught up with ARM on power consumption for performance per watt.  The primary difference is at the top end where ARM doesn't compete and price where ARM is much cheaper than Intel and Intel doesn't compete.
    cgWerks
  • Reply 105 of 159
    cornchipcornchip Posts: 1,087member
    netrox said:
    I wonder if it is possible to have a dual processors system - say have an Atom-like Intel for system support and ARM for dedicated application use? 
    Aren't they already going that direction with the MBPs with touch bar?
  • Reply 106 of 159
    cgWerkscgWerks Posts: 1,609member
    GeorgeBMac said:
    Why would they be "working on it"?   
    They already worked on it and completed it.   They called the result "iOS".
    I think that's what Apple is thinking too... but if you've ever actually tried it, iOS doesn't cut it for a pretty big segment of creators making the stuff the masses need. No creators, no masses.

    GeorgeBMac said:
    Would that be a touch screen laptop?
    ... They already nixed that idea.

    Well, until they added a keyboard to the iPad.  
    Then they added a file system to the iPad.
    ...  Can you guess what's coming next?
    Heh, no doubt. But, those things are needed for iOS to become more productive.
    On the other hand, IMO, a touch-screen laptop is just a poor UI/UX. But, that doesn't mean Apple won't do it, as they aren't exactly the UX/UI experts any longer. :(
  • Reply 107 of 159
    Does this mean going back to the days of most programs being incompatible with Mac like with the Power PC? No bootcamp anymore? 
    I'm curious what percentage of the Mac market is still running Windows. While Bootcamp helped me make the jump from PC to Mac, I haven't felt the slightest temptation to run Windows in many years. The BYOD revolution has made workplaces more platform-agnostic, and enterprise software is increasingly transitioning from Windows to SaaS.

    Bottom line, I don't think the lack of Windows support is going to be a significant deterrent to buying a future ARM-powered MacBook—especially if it blows its Intel-powered competition out of the water in both performance and battery life.
    edited September 2017 tmay
  • Reply 108 of 159
    sog35 said:
    Here. We. Go.

    I think we will be seeing an iOS desktop very soon..

    Hopefully in 5 years Apple is almost 100% vertically integraded and does not have to rely on any vendor to provide parts.
    Would that be a touch screen laptop?
    ... They already nixed that idea.

    Well, until they added a keyboard to the iPad.  
    Then they added a file system to the iPad.
    ...  Can you guess what's coming next?
    "Can you guess what's coming next?"

    No.
  • Reply 109 of 159
    asdasdasdasd Posts: 5,102member
    rcfa said:
    IOS and macOS are basically the same OS with different UI libraries. ARM Macs are a no brainer.
    Most people know it’s technically possible (although they seem to think an ARM chip is somehow linked to ios - as if Apple needs to create a surface type hybrid to run ARM. High Sierra runs as is on arm chips with a recompile). 

    The question is why? Leverage against Intel I get. However I doubt Apple are going back to the Mac mini model. That didn’t work. It shifted no needle - either in profits, sales or market share. Apple abandons non profitable lines. 

    There seems to be a strange correlation between posters here who think Apple should go cheap on Macs and think Apple should not have a cheaper iPhone line. In fact going cheaper on the latter makes sense, not least because it hasn’t been tried yet. There are other reasons. OS loyalty is less sticky on mobile devices. Sunk software costs are lower on mobile devices, particularly androids, so moving is easy.

    If Apple competes for the lower end mobile markets in developing countries they can lock people in and get a life long customer in countries where people’s income is increasing. Sell a cheap iphone now to get more expensive sale later. And to gain service revenue. 

    In the desktop/laptop world this largely doesn’t apply. People are either locked in because they spent significant money on software in higher end machines or they are going the cheap no software route of chrome books. 

    The Mac mini failed. Arm won’t change that. Cheaper iPhones are necessary though 
    edited September 2017
  • Reply 110 of 159
    tht said:
    MacPro said:
    Does this mean going back to the days of most programs being incompatible with Mac like with the Power PC? No bootcamp anymore? 
    My guess would be there will be two types of Macs for a few years, with and without Intel.

    There are a few ways they can go. 

    1. Competition: offer macOS/ARM and macOS/x86 machines concurrently. If Apple is good with ARM, x86 machines sales will dry up and people won't notice their retirement. Long term commitment with basically 2 teams.

    2. Forced Migration: basically what Apple did with PPC to Intel. This only works if Apple can provide clearly faster, and cheaper, ARM machines than x86 ones. 

    3. Asymmetric: offer iOS laptops and desktops. iOS will have to be expanding to have overlapping windows, shell access, same feature set as macOS. Then apps need to be ported.

    Number 3 is interesting as it means AppKit can be retired and most apps will be on the more modern UIKit or UXKit, or a new superset developer kit. All this iOS apps and games could be used. Xcode, FCPX, Logic, etc have to be ported, and there would be automated tools for it. 


    1.  Agreed - FYI - Apple does not use x86 - they use x86-64 only

    2. Apple will move if it sees an advantage - e.g. that hardware is in its ascendancy - to a specific technology.  Whether that is forced migration is debatable as individuals do not have to buy Apple hardware.  If enough do not then Apple will change.

    3. Disagree - there will always be distinction between iOS devices and macOS devices.  Apple will work to bring parity in the API[s] and then decide if it makes sense to expose a particular API in the same way on both iOS and macOS - e.g. Metal 2.  This will be on a low level and not on a user interface level as iOS is geared to touch and macOS is geared to mechanical pointer[s].
  • Reply 111 of 159
    asdasdasdasd Posts: 5,102member
    What difference is there between the metal api on Mac and iOS anyway?
  • Reply 112 of 159
    cgWerkscgWerks Posts: 1,609member
    freediverx said:
    I'm curious what percentage of the Mac market is still running Windows. While Bootcamp helped me make the jump from PC to Mac, I haven't felt the slightest temptation to run Windows in many years. The BYOD revolution has made workplaces more platform-agnostic, and enterprise software is increasingly transitioning from Windows to SaaS.
    Depends... it isn't just for switchers. For example, a web designer might run Windows either in BootCamp or Parallels, etc. to test websites via browsers on the Windows platform. Or, a home user might use macOS most of the time, but fire up BootCamp to run a few Windows based games. Some business users buy Apple laptops and then just load them up with Windows.

    But, the bigger question is about development of apps. While I'm not a developer, I'd imagine it's easier to port an app from the Windows platform to Intel based Macs than it is to ARM based Macs. That said, Software is no longer so Windows-centric, so it's much less important today than back when Apple switched to Intel.

    asdasd said:
    The Mac mini failed. Arm won’t change that. Cheaper iPhones are necessary though 
    The Mac mini didn't fail. Apple failed by nerf'ing it and not updating it. It was fairly popular back when it was a reasonably priced, competent machine. Also, iMacs have continued to gain popularity, but not everyone wants an all-in-one.

    Also, while Macs have never been cheap, after Jobs returned to Apple, they started to become fairly competitive given the quality and components. Apple has been slowing moving the prices up and the tech down on the Mac line. Yes, they need to make a good profit, but at some point, the straw starts to break the camel back.
    GeorgeBMac
  • Reply 113 of 159
    nhtnht Posts: 4,214member
    cgWerks said:
    freediverx said:
    I'm curious what percentage of the Mac market is still running Windows. While Bootcamp helped me make the jump from PC to Mac, I haven't felt the slightest temptation to run Windows in many years. The BYOD revolution has made workplaces more platform-agnostic, and enterprise software is increasingly transitioning from Windows to SaaS.
    Depends... it isn't just for switchers. For example, a web designer might run Windows either in BootCamp or Parallels, etc. to test websites via browsers on the Windows platform. Or, a home user might use macOS most of the time, but fire up BootCamp to run a few Windows based games. Some business users buy Apple laptops and then just load them up with Windows.

    But, the bigger question is about development of apps. While I'm not a developer, I'd imagine it's easier to port an app from the Windows platform to Intel based Macs than it is to ARM based Macs. That said, Software is no longer so Windows-centric, so it's much less important today than back when Apple switched to Intel.

    asdasd said:
    The Mac mini failed. Arm won’t change that. Cheaper iPhones are necessary though 
    The Mac mini didn't fail. Apple failed by nerf'ing it and not updating it. It was fairly popular back when it was a reasonably priced, competent machine. Also, iMacs have continued to gain popularity, but not everyone wants an all-in-one.

    Also, while Macs have never been cheap, after Jobs returned to Apple, they started to become fairly competitive given the quality and components. Apple has been slowing moving the prices up and the tech down on the Mac line. Yes, they need to make a good profit, but at some point, the straw starts to break the camel back.
    Given that Mac sales are doing very well I wouldn't hold my breath about Apple changing Mac pricing strategy...

    Apple killed the mini's bang for the buck to maintain ASPs.  I doubt we'll see a quad core mini again...especially with full eGPUs support on the horizon.  

    If we do it'll be priced the same as a 21" iMac with the same processor...and not because the dropped the price of the iMac...
  • Reply 114 of 159
    nht said:


    As far as performance ARM has beefed up their architecture.  You can now have up to 8 CPU in a cluster and you can cluster 32 
    of these Octoclusters together.   Want more speed?  Add more CPU. 
    Doesn't work that way for many apps which are constrained by single threaded performance because the expensive computations aren't easily parallellizable.

    Also multiple CPUs has architecture costs vs multiple cores.  There's no free lunch.

    I have no idea why folks will look at the A series of processors and then claim that by adding more cores you can outperform the faster processor when the older quad core A series chips regularly beat the pants off the hexacore and octocore chips that Samsung and Qualcomm made and the A11 will demolish everything Samsung or Qualcomm will have because its single core performance is so much better.  Samsung could make a dual processor octacore phone (16 cores) and still lose to the A11 in usable performance.

    Those performance advantages don't simply disappear because it's Intel desktop quad core i7 vs any ARM octacore, even Apple's.  Intel has been working extremely hard to improve performance per watt to the point we have highly performant 4.5W Intel processors.  The dual core Intel Core i7-7Y75 MacBook scores 4020 single core and 7999 multicore on 4.5W TDP.  That's on par with the A11 but only using 2 cores and not 4+2.  All those folks that claim Intel has done nothing in the last few years is ignoring that Intel has been doing EXACTLY what Apple has wanted them to do:  concentrate on performance per watt over top end speed.

    Essentially Intel has caught up with ARM on power consumption for performance per watt.  The primary difference is at the top end where ARM doesn't compete and price where ARM is much cheaper than Intel and Intel doesn't compete.
    The highest end Intel processors (which is what you are comparing to) account for about 10-20% of the market.   That leaves a whole lot of room for the "A" series to step in.
  • Reply 115 of 159
    nhtnht Posts: 4,214member
    nht said:


    As far as performance ARM has beefed up their architecture.  You can now have up to 8 CPU in a cluster and you can cluster 32 
    of these Octoclusters together.   Want more speed?  Add more CPU. 
    Doesn't work that way for many apps which are constrained by single threaded performance because the expensive computations aren't easily parallellizable.

    Also multiple CPUs has architecture costs vs multiple cores.  There's no free lunch.

    I have no idea why folks will look at the A series of processors and then claim that by adding more cores you can outperform the faster processor when the older quad core A series chips regularly beat the pants off the hexacore and octocore chips that Samsung and Qualcomm made and the A11 will demolish everything Samsung or Qualcomm will have because its single core performance is so much better.  Samsung could make a dual processor octacore phone (16 cores) and still lose to the A11 in usable performance.

    Those performance advantages don't simply disappear because it's Intel desktop quad core i7 vs any ARM octacore, even Apple's.  Intel has been working extremely hard to improve performance per watt to the point we have highly performant 4.5W Intel processors.  The dual core Intel Core i7-7Y75 MacBook scores 4020 single core and 7999 multicore on 4.5W TDP.  That's on par with the A11 but only using 2 cores and not 4+2.  All those folks that claim Intel has done nothing in the last few years is ignoring that Intel has been doing EXACTLY what Apple has wanted them to do:  concentrate on performance per watt over top end speed.

    Essentially Intel has caught up with ARM on power consumption for performance per watt.  The primary difference is at the top end where ARM doesn't compete and price where ARM is much cheaper than Intel and Intel doesn't compete.
    The highest end Intel processors (which is what you are comparing to) account for about 10-20% of the market.   That leaves a whole lot of room for the "A" series to step in.
    So what?  The majority of that 80% is better off with an iPad Pro vs an Arm MacBook.  The remainder are largely business users with a need to run windows apps natively. 

    Apple doesnt compete in most of that 80-90% market with the Mac anyway but with the iPad. 
    edited September 2017
  • Reply 116 of 159
    hmurchisonhmurchison Posts: 12,125member
    asdasd said:
    rcfa said:
    IOS and macOS are basically the same OS with different UI libraries. ARM Macs are a no brainer.



    The Mac mini failed. Arm won’t change that. Cheaper iPhones are necessary though 
    The Mac mini never failed.  They stilll sell despite not being upgraded in 4 yeard
  • Reply 117 of 159
    tallest skiltallest skil Posts: 43,399member
    The Mac mini never failed.  They stilll sell despite not being upgraded in 4 yeard
    Sort of sad that the Mac mini apparently sells better than the Mac Pro, despite the dearth of updates for both.
  • Reply 118 of 159
    This is Apple running its MacOS from the A chips and having an intel processor for 3rd party apps.  If you think two processors in one system is bonkers, remember the touchbar chip in the new macbook pros, or think back 20 (25?) years to the old mac desktops with the cardslot for an x86.  Anybody remember those?

    Benefits:
    - Apple stuff runs on its own stuff.  Cheaper for Apple to develop.
    - the intel chips are still "inside" for running all the enterprise stuff, excel, etc.  And computer games.  And because the intel chips aren't also handling system processes there should be a performance improvement or the potential to use lower-wattage intel chips with less of a performance hit.
    - Apple would have to figure out how to use multiple chips/cores at the same time, to manage efficiency/performance question marks... new iPhones seem to have put a lot of work into this.
    - An operating system running on a dedicated chip will feel "snappier" (sorry, couldn't help it)

    Probably late 2018-mid 2019. Two years. I agree with assessment, because the current iPhone chips "could" boot a mac desktop OS, but right now every A11 they produce is probably needed for phones. This article is probably from a leaker who saw that, with the new A11s, Apple is now really seriously kicking off R&D into this whereas before it was a side project, "wouldn't it be nice if we only compiled for ARM?".
  • Reply 119 of 159
    GeorgeBMacGeorgeBMac Posts: 2,760member
    nht said:
    nht said:


    As far as performance ARM has beefed up their architecture.  You can now have up to 8 CPU in a cluster and you can cluster 32 
    of these Octoclusters together.   Want more speed?  Add more CPU. 
    Doesn't work that way for many apps which are constrained by single threaded performance because the expensive computations aren't easily parallellizable.

    Also multiple CPUs has architecture costs vs multiple cores.  There's no free lunch.

    I have no idea why folks will look at the A series of processors and then claim that by adding more cores you can outperform the faster processor when the older quad core A series chips regularly beat the pants off the hexacore and octocore chips that Samsung and Qualcomm made and the A11 will demolish everything Samsung or Qualcomm will have because its single core performance is so much better.  Samsung could make a dual processor octacore phone (16 cores) and still lose to the A11 in usable performance.

    Those performance advantages don't simply disappear because it's Intel desktop quad core i7 vs any ARM octacore, even Apple's.  Intel has been working extremely hard to improve performance per watt to the point we have highly performant 4.5W Intel processors.  The dual core Intel Core i7-7Y75 MacBook scores 4020 single core and 7999 multicore on 4.5W TDP.  That's on par with the A11 but only using 2 cores and not 4+2.  All those folks that claim Intel has done nothing in the last few years is ignoring that Intel has been doing EXACTLY what Apple has wanted them to do:  concentrate on performance per watt over top end speed.

    Essentially Intel has caught up with ARM on power consumption for performance per watt.  The primary difference is at the top end where ARM doesn't compete and price where ARM is much cheaper than Intel and Intel doesn't compete.
    The highest end Intel processors (which is what you are comparing to) account for about 10-20% of the market.   That leaves a whole lot of room for the "A" series to step in.
    So what?  The majority of that 80% is better off with an iPad Pro vs an Arm MacBook.  The remainder are largely business users with a need to run windows apps natively. 

    Apple doesnt compete in most of that 80-90% market with the Mac anyway but with the iPad. 
    80% of users should buy a MacBook Pro instead of a MacBook?   Really?   Where did you come up with that?  And, where are they going to come up with the money for that?   There's a reason why Apple named it a "Pro" line.

    And, once the iPad has a cursor & touchpad on its keyboard, it will basically be a MacBook since it will be able to do the things that a MacBook does.  Or it can be a tablet.   Whichever you need at the time.
  • Reply 120 of 159
    asdasdasdasd Posts: 5,102member
    asdasd said:
    rcfa said:
    IOS and macOS are basically the same OS with different UI libraries. ARM Macs are a no brainer.



    The Mac mini failed. Arm won’t change that. Cheaper iPhones are necessary though 
    The Mac mini never failed.  They stilll sell despite not being upgraded in 4 yeard
    Everything will sell at some limited amount. Apple hasn't updated in what - 3 years. It is not a market they are interested in. 
Sign In or Register to comment.