Rosetta 2 lacks support for x86 virtualization, Boot Camp not an Apple Silicon option [u]

1356

Comments

  • Reply 41 of 110
    What if emulation runs fast enough on Apple Silicon and Big Sur that you don't need virtualization? The Tomb Raider game demo was running in emulation.
    watto_cobra
  • Reply 42 of 110
    the comments were some seriously amusing reading.
    dewmeaderutterwatto_cobra
  • Reply 43 of 110
    elijahgelijahg Posts: 2,652member
    Well, I'm pretty sad to see this. Unless Apple changes its mind or there's some incredibly fast emulation software, when my 2019 iMac conks it, it will be the end of 25 years of Apple for me. I'm not buying two computers just to use x86 Windows at more than a snail's pace.

    Back in the days of yore, Macs being able to run Windows was a boon for switchers and removed a lot of apprehension as they could keep running Windows if they didn't like macOS. I used to run a Mac network at a school, and we used to use Windows in a VM for plenty of education programs that had no Mac equivalent, if we couldn't do that we likely wouldn't have had anywhere near as many Macs - as we'd need PCs to run essential software. Apple is hammering the Mac into a smaller and smaller niche, unfortunately. Speed is irrelevant if it won't run the programs you need.
    GeorgeBMaccurtis hannahprismatics
  • Reply 44 of 110
    elijahgelijahg Posts: 2,652member
    firelock said:
    I’m not sure that everyone understands the difference between virtualization and emulation. It was never likely that x86 virtualization software would run on ARM Macs even with Rosetta. But it is a virtual certainty that Windows emulators will be available. They won’t be as good as a virtual machine but they will probably suffice more most needs.
    Well it depends at which "level" Rosetta runs at. If it's low enough level that it literally converts x86 instructions to ARM (i.e. emulation), there's no reason virtualisation shouldn't be a feasible option. If it runs at a higher level and interprets API calls - which is mostly how how Rosetta 1 worked, then no it won't work.
  • Reply 45 of 110
    INTEL for Bootcamp and Virtualization was holding the Mac back, period. Why develop native software if they could just default to Windows on INTEL.

    AMR for Mac, Apple Silicon, will be twice the speed, developers will be compelled to develop native for Mac.

    Just look at the sorry state of Android. A $400 iPhone SE is faster than a $1300 Android phone. Vertical wins hands down.

    ARM servers will sweep the server space but they won't come from Apple or maybe they will.
    edited June 2020 aderutterfastasleepwatto_cobra
  • Reply 46 of 110
    elijahgelijahg Posts: 2,652member
    netrox said:
    godrifle said:
    Disappointing. Mac on intel was the greatest dev machine ever invented. I have recommended Macs based primarily on this feature itself in my software development classes, for years. Looks like those days are over. ☹️
    ARM is literally now in billions of devices already. We have Raspberry Pi with AMD which costs ONLY $100. A lot of operating systems are being ported to ARM. Even Ubuntu is now ported to ARM. Windows is already ported to ARM.  x86 has literally run out of all its juice while ARM has plenty of room for decades to come. In just a decade, x86 will be "dead" because it just won't be able to compete against ARM anymore. 

    We need to stop clinging to legacy hardware and move forward to the future. 
    There is no Raspberry Pi with AMD. All Raspberry Pis have ARM CPUs, and start off about $45. All major OSs are on ARM already. I agree x86 has run out of juice, but "decades to come" is somewhat of a baseless claim. All silicon CPU architectures will hit the same physics wall eventually. Apple's architecture was showing less generational speed increases by the iPhone 6, they will too hit the same wall as Intel, eventually. 

    Yeah "stop clinging to legacy hardware and move forward" is fine if your world is entirely Apple, and never interacts with anyone or anything else. For 90% of the world, this isn't the case. Businesses have hordes of usually ancient x86 software on Windows that they keep going, and as @rbnetengr said, most banks are still on mainframes using Cobol. Telling everyone to stop clinging to legacy hardware is like asking an entire country to change from speaking English to Japanese.
    GeorgeBMac
  • Reply 47 of 110
    elijahgelijahg Posts: 2,652member

    Apple doesn’t need to have Rosetta do this. 

    And boot camp camp isn’t a virtual machine. It simply sets up the Mac to load windows natively and event hunts down the windows drivers for the hardware in the Mac. 

    Its RS not a virtual machine. 

    If if you need that, parallels and VMware both will be up to the task. 

    Parllels was even even doing this when Apple was using RISC PowerPC chips. 

    ‘’Apple is distancing itself from everyone else. They are leveraging what made the iPhone and iPad the best and most popular high end products in their fields and applying it to the Mac. The Mac will be faster, more efficient, and tie into the largest and most secure ecosystem there is. 

    Basically, Apple Silicon Macs are obsoleting the competition. So no need to support them directly. Third parties can tackle that. And they are already working on it with their A12Z dev kits. 
    1 - Yes it does.
    2 - Apple provides drivers for the hardware in the Mac, most of which is standard PC hardware anyway.
    3 - ?
    4 - As of right now, no they won't.
    5 - No it wasn't. Parallels didn't exist on PPC machines.
    6 - Seems they may be distancing themselves from their own customers too, business especially.
    GeorgeBMac
  • Reply 48 of 110
    elijahgelijahg Posts: 2,652member

    cropr said:
    For the software company I own, the main question is: can the new Macs run Docker images at (almost) native speed.  The documents are not 100% clear, but point in the wrong direction.  Which basically means that  2020 will be the year we stop buying Macs,  (With the exception of the single developer who develops an iOS front end for our cloud services)

    Since Docker does no emulation, only a form of virtualisation; if the images are compiled for Apple Silicon then yes, they will be native speed and in all likelihood faster than Intel. If they are not compiled for AS then no, they will not run at all.
  • Reply 49 of 110
    elijahgelijahg Posts: 2,652member
    My hope is that this will prove to devs both Windows and Mac that ARM is a viable, fast platform for general purpose computing. If that's the case, Windows on ARM will probably improve a lot, and programs will slowly be recompiled for ARM on Windows, and x86 will die. With Apple's support, it may even be possible to boot ARM Windows on an ARM Mac. However, since Apple's instruction set is no longer that similar to ARM, I think that's unfortunately unlikely.
    GeorgeBMaccurtis hannah
  • Reply 50 of 110
    MarvinMarvin Posts: 14,836moderator
    There is already Windows 10 for ARM64. For example, it's available on Microsoft Surface Pro X. Getting it running on Mac should be no problem as long as the drivers are available.
    Yes someone managed to get Windows ARM to run on the Rasberry Pi here, there's a disk image online for it:



    There are some compatibility issues due to missing drivers. Apple would probably need to make Windows drivers for the hardware (CPU/GPU/wifi/peripherals) like they do for Bootcamp. That should also be able to run in the native ports of Parallels and VMWare.

    If Microsoft's x86/64 translation works under that, Parallels/VMWare would run software much like it does now but graphics-intensive apps would probably need some kind of hardware virtualization. Apple could make some GPU virtualization tech.

    The GPU is different from the CPU for translation. When they showed Tomb Raider running, they mentioned that the GPU recompiles what it needs; shaders and things are compiled dynamically. It's just the CPU code that is translated. That includes calls to the GPU so there is some overhead but the actual processing of the graphics and rendering would be native speed.

    If Apple can virtualize the GPU and Parallels/VMWare can get full compatibility with ARM Windows, including x86/x64 translation, that should come close to running native Bootcamp and saves having to partition/reboot.

    The safe option for now if people depend on good Windows performance and compatibility is to plan for buying a secondary Windows PC or stick with Intel Macs as long as possible (e.g just keeping hold of current Macs). Windows PC are cheap enough and they can sit in the corner somewhere hooked up to a network.

    I reckon there will be a lot of experimenting going on over the next couple of years to see what's possible. I expect Microsoft will want to get Windows ARM running on Macs in some form, even if Apple doesn't go the route of allowing dual-boot. Parallels/VMWare certainly will as they have over 7 million users:

    https://www.parallels.com/pd/virtual-pc-on-mac/

    If you had 7 million customers, you'd do everything possible to stop that dropping to near 0 and Apple working with the virtualization companies is a positive sign.
    muthuk_vanalingamwatto_cobra
  • Reply 51 of 110
    MacProMacPro Posts: 19,505member
    anome said:
    Rayz2016 said:
    I suppose you could run Windows in the Cloud

    https://docs.microsoft.com/en-us/azure/virtual-desktop/overview

    That can get pricy, to the point where it's probably cheaper to just buy a Windows box and be done with it.


    I agree and already did that since Boot Camp itself has so many limitations other than for Office type use. Of course, we solved this on Apple ][s and on early Macs with an add-in card that ran the Microsoft flavors de jour so perhaps again, just for Office type usage a similar external add-on could be created that has an x86 in a dongle attached via the fastest connection perhaps USBc or better.
    edited June 2020 dewmeelijahgwatto_cobra
  • Reply 52 of 110
    beeble42beeble42 Posts: 32member
    mikeinca said:
    I so don’t care about this. Maybe this is an issue for what... 1% of users?

    im not a developer.  I’m in education.  At most we need MS Office.   Most apps are web apps. 

    Wtf you guys complaining about?   You, all of you whom so boldly proclaim you are the hardcore and future users, don’t know squat!   All the apps I need are web-based.  

    All the apps I need are web-based!!!!

    read that over and over again!!! 

    You hardcore users are so out of touch it’s actually amusing.  

    Get over yourselves.  
    So why are you buying a Mac instead of something like a Chromebook or iPad? Same web browser based experience but far less money. You declare that you aren't actually using OSX for anything but running a web browser so what is it to you that other people actually do use OSX and applications that depend on it. No one is disputing that many people don't need the "truck", as Steve Jobs put it. What people are complaining about is that Apple appears to be taking away some of the important truck features that many of the truck users need.

    The Mac will stop being the best and most cost effective development platform available because it could do and run EVERYTHING. Without the ability to run critical Windows based business applications that most businesses still rely on, it will simply disappear in the corporate market where it's had massive growth in the last 5 years. For those who do desktop gaming (flight simulation, high end fps, complex strategy - games that require more than a simplistic d-pad controller) the Mac has always struggled but had gained some momentum in recent years in spite of poor GPU support from Apple. That's all dead. I can't imagine game developers investing effort into porting these games when none of the community plugins that bring much of the value to these games will work. But simpler games from iOS will be everywhere. Except why play them on a Mac instead of an iPad or iPhone?

    So many small market segments are losing the things that have made them choose Apple. They will have little alternative other than to switch to Windows. All those little segments add up to a very large number of customers. Having been an Apple customer for over 20 years, developer and a former Apple employee, for the first time I'm not sure what OS will run on my next desktop. For the first time in well over 2 decades I'm actually finding myself questioning whether I can justify the risk of buying another Mac. I know that a Windows PC will do the things I need it to do. I know a Mac will be well built and be far more pleasurable to use. I know that it specifically won't do a number of things I need. And there are many, many things I don't know that could be non-issues or deal breakers. I don't have to make a decision right now. Time will tell and things may change as more information comes to light. But none of those questions exist when I consider switching to Windows. In fact, there are certain things with 3rd party software I could do on Windows I can't do on my Mac.

    When a tool stops doing the thing you need it to do, it doesn't matter how much you love the tool, or how good it looks, how satisfying it is to use. It becomes nothing more than a fancy paperweight. For the time being the Mac is undeniably, by far the most versatile technology tool in existence. But it looks like the clock is ticking before it loses that claim.
    elijahgGeorgeBMacinTIMidator
  • Reply 53 of 110
    avon b7avon b7 Posts: 6,434member
    INTEL for Bootcamp and Virtualization was holding the Mac back, period. Why develop native software if they could just default to Windows on INTEL.

    AMR for Mac, Apple Silicon, will be twice the speed, developers will be compelled to develop native for Mac.

    Just look at the sorry state of Android. A $400 iPhone SE is faster than a $1300 Android phone. Vertical wins hands down.

    ARM servers will sweep the server space but they won't come from Apple or maybe they will.
    I'm not sure developers will be compelled to develop. We'll have to wait and see. After these kinds of announcements there is always a lot of technical stuff that seeps out over the following months. The devil is going to be in the details. I'm still not sure if I want in on this move or not. I'll have to wait and see how things play out. 

    That $1,300 Android price tag has very little to do with the SoC. That price represents a lot of tech that is not available on the SE and that is why the SE costs what it does in the first place.

    ARM servers are out there, along with absolutely huge clusters for HPC, AI etc. Apple could move into those areas, and I believe that would be a good move, but commercially Apple doesn't yet have the base it requires to make a splash nor the confidence of the market. It would have to make some big promises to convince potential buyers that their commitment to any such initiatives would not wane and end up leaving buyers high and dry. At some point I think it can be pulled off though in spite of some huge competitors that are already making a name for themselves. 

    A good place to start would be with its own data centres and internal projects. The car project is going to require onboard data centres and enormous capacity to sync with cloud services. 
    elijahgmuthuk_vanalingam
  • Reply 54 of 110
    aderutteraderutter Posts: 539member
    So maybe later this year I will have the choice of buying a new Intel Mac or a new ARM Mac.
    Every Mac I have ever bought from the 0x0 processor versions onwards have lasted me at least 7 years as useable machines.
    This is way longer than it would take me to write them off against tax (depreciation) for my business.

    So if I buy an Intel Mac either this year or next, or maybe even in 2022 I will likely not have a problem until 2030 if I still want to run Windows for those few apps scenarios in 10 years time.

    So basically in my opinion nobody has a sensible or viable reason to stop buying Macs (yet) as far as I can tell.

    Personally I expect to buy an ARM Mac as my current 2014 and 2019 iMacs will be usable for years ahead.
    Heck I can install Big Sur or Catalina (or windows via bootcamp) on even my old 2014 iMac if I want.
    I currently use Parallels and Windows7 on my 2014 iMac for some testing purposes and don’t see that changing.

    My 2014 IMac is still really great and this to me highlights how slow Intel have been to upgrade/improve their chips. In the last 6 years single core performance advance from Intel has been so limited. If I was to buy another new Intel iMac I would realistically only be buying significantly better multi-core performance and that is not enough reason to buy Intel (for me) after 6 years! If Intel had advanced at a decent rate I would have bought Intel again.

    I think that the first ARM Macs (apart from the dev kit) will have an A13 or A14 chip variant and be blazing fast compared to their Intel equivalent when running native apps like Adobe Creative Cloud suite or MS Office or Maya or FCP or...

    I think the reason the iPad Pro still has an A12 variant is precisely so the new Macs can have better processors than the iPad Pro.
    I wonder if the iPhone12 will still use A13 and only Macs will get A14 this year...




    GG1fastasleepwatto_cobra
  • Reply 55 of 110
    nhtnht Posts: 4,522member
    godrifle said:
    Disappointing. Mac on intel was the greatest dev machine ever invented. I have recommended Macs based primarily on this feature itself in my software development classes, for years. Looks like those days are over. ☹️
    I’ve been seeing some transition from MacOS to Linux since mostly folks develop in Docker...this will simply accelerate that.  

    Multi-arch docker development depends on qemu and it struck me as a little kludgy.




    elijahgprismatics
  • Reply 56 of 110
    nhtnht Posts: 4,522member
    cropr said:
    For the software company I own, the main question is: can the new Macs run Docker images at (almost) native speed.  The documents are not 100% clear, but point in the wrong direction.  Which basically means that  2020 will be the year we stop buying Macs,  (With the exception of the single developer who develops an iOS front end for our cloud services)

    No, or at least multi-arch docker uses qemu to support ARM so I would assume the inverse is true.  Whether that’s fast enough for your folks will be debatable.

    For docker devs targeting intel hardware it strikes me that Ubuntu running on cheaper intel hardware is a viable alternative.

    What might happen is more ARM servers but that’s been REAL slow coming.  
  • Reply 57 of 110
    nhtnht Posts: 4,522member
    Ditching Bootcamp?
    So why would someone dump a 4 digit investment and lock themselves into something with 10% market share -- and no recourse to the dominant OS with a 90% share -- particularly when equivalent hardware can be had far cheaper in the WIndows World.?

    In other words:
    Can MacOS isolate Mac hardware and be successful?  A person buying a Mac will be locked into a limited edition system -- unless they buy a 2nd PC.
    Or, can integration with iOS carry it?
    For consumer or edu users the world is 90% web based and not Windows or MacOS based.

    For MacOS users most native apps will get moved over to Arm.  The biggest question is whether MS will port Visio and Project to ARM.

    Given where MS is positioned these days they may not care about protecting their Windows ecosystem as much today especially if Apple provides a sufficiently large check.
  • Reply 58 of 110
    nhtnht Posts: 4,522member
    elijahg said:
    Well, I'm pretty sad to see this. Unless Apple changes its mind or there's some incredibly fast emulation software, when my 2019 iMac conks it, it will be the end of 25 years of Apple for me. I'm not buying two computers just to use x86 Windows at more than a snail's pace.

    Back in the days of yore, Macs being able to run Windows was a boon for switchers and removed a lot of apprehension as they could keep running Windows if they didn't like macOS. I used to run a Mac network at a school, and we used to use Windows in a VM for plenty of education programs that had no Mac equivalent, if we couldn't do that we likely wouldn't have had anywhere near as many Macs - as we'd need PCs to run essential software. Apple is hammering the Mac into a smaller and smaller niche, unfortunately. Speed is irrelevant if it won't run the programs you need.
    I can’t think of any student facing edu software today that doesn’t run in the cloud. 
  • Reply 59 of 110
    elijahgelijahg Posts: 2,652member
    nht said:
    elijahg said:
    Well, I'm pretty sad to see this. Unless Apple changes its mind or there's some incredibly fast emulation software, when my 2019 iMac conks it, it will be the end of 25 years of Apple for me. I'm not buying two computers just to use x86 Windows at more than a snail's pace.

    Back in the days of yore, Macs being able to run Windows was a boon for switchers and removed a lot of apprehension as they could keep running Windows if they didn't like macOS. I used to run a Mac network at a school, and we used to use Windows in a VM for plenty of education programs that had no Mac equivalent, if we couldn't do that we likely wouldn't have had anywhere near as many Macs - as we'd need PCs to run essential software. Apple is hammering the Mac into a smaller and smaller niche, unfortunately. Speed is irrelevant if it won't run the programs you need.
    I can’t think of any student facing edu software today that doesn’t run in the cloud. 
    Ours were mainly for media stuff, Photoshop, Illustrator, iMovie and Final Cut. Unfortunately due to typical resistance to upgrades there's plenty of fairly bespoke education stuff that's still just about kept running, but needs Windows.
    GeorgeBMac
  • Reply 60 of 110
    dewmedewme Posts: 4,547member
    As I mentioned prior to WWDC, if Apple were to fully sanction the running of macOS virtual machines in VMWare Workstation or Player (rather than artificially blocking them so they can only run in VMWare Fusion) they could certainly beat down some of the bellyaching about the lack of virtualization support in Apple Silicon Macs. This would allow developers to still develop in XCode in a macOS VM on a Windows box in addition to still running Mac apps.

    I hate to say it, but if I could run macOS VMs on a Windows box I'd probably be buying a beefy new Windows machine this year. The  Apple Silicon Mac transition is going to take a few years to get all of the kinks ironed out, like virtualization. Yes, Apple is going to be releasing new Intel Macs over the next coupe of years. The problem is that Apple does not support older machines with OS updates nearly as long as Microsoft does. I still have 2005 era PCs that run Windows 10 fine because I maxed out the hardware when I bought it. I now have a maxed out late-2012 iMac 27" that absolutely kicks the sh** out of my 2014 Mac mini, but with Big Sur Apple pulled the plug on the iMac and kept it on the mini. Go figure.

    BTW, I've been able to drastically improve the dog-slow 2014 HDD-only Mac mini tremendously without disemboweling the mini at all using this very cool kit (https://www.amazon.com/Sintech-NGFF-NVMe-Upgrade-A1347/dp/B07Q5FBNVG/ref=sr_1_8?dchild=1&keywords=mac+mini+ssd&qid=1593009709&sr=8-8) and installing a 500 GB Samsung 970 EVO. I left the HDD in place and turned it all into a single Fusion drive using Apple's very slick fusion creation command line utility. It's still not close to my older iMac in overall performance because the iMac has 32 GB RAM, more cores, i7 vs i5, discrete graphics, and faster CPU but it's no longer tortuous to use (primarily as a home theatre computer) as it was in HDD-only configuration.

    So yeah, I'm grateful to get 8 years of macOS upgrades for my big iMac but Microsoft's try-it-at-your-own-risk model is much more open ended. I'm not sure why Apple dropped the late 2012 iMac from Big Sur, as I'm sure everyone who got booted from Big Sur feels, but I'm more puzzled about why my now-somewhat-less-pathetic 2014 Mac mini made the cut. So from a developer-machine perspective I currently don't have a beefy Mac to develop for Big Sur and use XCode 12.

    To stay current during the Apple Silicon Mac (ASM) transition I guess I'll just have to suck it up and buy whatever new iMac Apple rolls out next, which is likely to be Intel based, or get a super-stuffed newest Mac mini and a couple of 4K monitors. Sure, this is a one-percenter problem, just like the virtualization question, but it does further illustrate that the ASM transition is a serious cheese-move by Apple and we'll all have to strap ourselves in for the ride and expect a little turbulence along the way. 


    elijahgGG1GeorgeBMac
Sign In or Register to comment.