4 ms·

> in-house designed ARM CPUs. What could possibly go wrong!

by rusk 8y ago

lostlogin 8y ago
A lot. But imagine what could go right. The thing I most want is battery life, but I hope that isn’t squandered by translating it to thinness.
- rusk 8y ago
  > imagine what could go right .
edbaskerville 8y ago
Many things could go wrong, but the Apple codebases and engineering processes should be pretty robust to architecture changes at this point. NeXT went from 68030 to Intel/SPARC/others, and then to PowerPC with Mac OS X. OS X went from PowerPC to Intel. And large chunks of it are shared with iOS on ARM already. (Fun fact: even classic Mac OS was running on Intel at one point before the 680x0-to-PowerPC transition: https://en.wikipedia.org/wiki/Star_Trek_project https://en.wikipedia.org/wiki/Star_Trek_project) So there's good reason to think that they could pull this off.
- rusk 8y ago
  That’s an awful lot of hope
- charlesism 8y ago
  I have very little confidence it will be a smooth transition this time. Apple now is not the same people, and not the same culture as it was then. I don't see them sweating the details over Mac hardware.
close04 8y ago
With the CPU design or the transition to a new ISA? Lots of things can go wrong at any time even by doing nothing. But Apple does have a team of pretty good engineers that managed to design what's probably the fastest (best?) ARM cores on the market, and with good enough perf/Watt numbers. So I doubt their biggest hurdle will be to design a CPU. It will be to split the "MacOS" ecosystem in 2 and maintain two branches for x86 and ARM. And while they also have a team of pretty good software engineers with x86 and ARM experience they are already having major quality control problems lately. Having to maintain a dual ISA MacOS will be a challenge. Then they have to provide the legacy compatibility in order to convince people to buy into something with little to no support for most of what's out there today. But you know... we could use some diversity in this space. I wish them all the best.
- 394549 8y ago
  > It will be to split the "MacOS" ecosystem in 2 and maintain two branches for x86 and ARM. And while they also have a team of pretty good software engineers with x86 and ARM experience they are already having major quality control problems lately. Having to maintain a dual ISA MacOS will be a challenge. Then they have to provide the legacy compatibility in order to convince people to buy into something with little to no support for most of what's out there today. They've done that twice before (68k -> PPC, PPC -> Intel), and IIRC, seem to be preparing to do it again (Bitcode). I think the biggest risk for Apple is that they may not be able to keep up with x86 on the desktop performance front. They have a much smaller market to amortize the R&D over than Intel does.
  - rusk 8y ago
    The biggest risk for Apple is they’ve become a shiny things company and don’t really have the engineering stones for doing this kind of thing any more
- close04 8y ago
  @394549, I don't doubt that they can do it, as I was also saying a couple of comments above this one. But at what price? The difference is that now they have big QA issues before having to maintain the 2 branches. Their software (and possibly the software team) a decade and a half ago when they were moving to Intel was in a lot better shape. Unless I'm looking at this through nostalgia goggles. :)
  - rusk 8y ago
    They were also moving to a well established architecture from their own bespoke one and now they’re going the other way ...
    - close04 8y ago
      ARM is a well established ISA. The customization Apple might bring to the CPU architecture will make no difference to a user as long as the compatibility is there. The problem is that as established as ARM is they really don't have a meaningful presence in the desktop/laptop ecosystem. You have probably tens of thousands of apps for the mobile space but even if Apple magically adapted them for a non-touchscreen, keyboard and touchpad world the vast majority still won't cater to the regular or professional user wanting to buy a laptop, not a glorified tablet. And that's a hard sell. Why am I buying a MacBook if all I get is an iPad with a keyboard? And don't get me wrong, having used both iPads and CoreM ultraportables (Apple or otherwise) the iPad offered the better experience more often than not. But if I compare it to the regular CPU it just can't compete. The productivity with mobile apps is not where it should be to convince a normal buyer to switch, let alone a more demanding user.

> in-house designed ARM CPUs. What could possibly go wrong!

by rusk 8y ago

> in-house designed ARM CPUs. What could possibly go wrong!

A lot. But imagine what could go right. The thing I most want is battery life, but I hope that isn’t squandered by translating it to thinness.

rusk 8y ago
> imagine what could go right .

edbaskerville 8y ago

Many things could go wrong, but the Apple codebases and engineering processes should be pretty robust to architecture changes at this point. NeXT went from 68030 to Intel/SPARC/others, and then to PowerPC with Mac OS X. OS X went from PowerPC to Intel. And large chunks of it are shared with iOS on ARM already. (Fun fact: even classic Mac OS was running on Intel at one point before the 680x0-to-PowerPC transition: https://en.wikipedia.org/wiki/Star_Trek_project https://en.wikipedia.org/wiki/Star_Trek_project) So there's good reason to think that they could pull this off.

rusk 8y ago
That’s an awful lot of hope
charlesism 8y ago
I have very little confidence it will be a smooth transition this time. Apple now is not the same people, and not the same culture as it was then. I don't see them sweating the details over Mac hardware.

close04 8y ago

With the CPU design or the transition to a new ISA? Lots of things can go wrong at any time even by doing nothing. But Apple does have a team of pretty good engineers that managed to design what's probably the fastest (best?) ARM cores on the market, and with good enough perf/Watt numbers. So I doubt their biggest hurdle will be to design a CPU. It will be to split the "MacOS" ecosystem in 2 and maintain two branches for x86 and ARM. And while they also have a team of pretty good software engineers with x86 and ARM experience they are already having major quality control problems lately. Having to maintain a dual ISA MacOS will be a challenge. Then they have to provide the legacy compatibility in order to convince people to buy into something with little to no support for most of what's out there today. But you know... we could use some diversity in this space. I wish them all the best.

394549 8y ago
> It will be to split the "MacOS" ecosystem in 2 and maintain two branches for x86 and ARM. And while they also have a team of pretty good software engineers with x86 and ARM experience they are already having major quality control problems lately. Having to maintain a dual ISA MacOS will be a challenge. Then they have to provide the legacy compatibility in order to convince people to buy into something with little to no support for most of what's out there today. They've done that twice before (68k -> PPC, PPC -> Intel), and IIRC, seem to be preparing to do it again (Bitcode). I think the biggest risk for Apple is that they may not be able to keep up with x86 on the desktop performance front. They have a much smaller market to amortize the R&D over than Intel does.
- rusk 8y ago
  The biggest risk for Apple is they’ve become a shiny things company and don’t really have the engineering stones for doing this kind of thing any more
close04 8y ago
@394549, I don't doubt that they can do it, as I was also saying a couple of comments above this one. But at what price? The difference is that now they have big QA issues before having to maintain the 2 branches. Their software (and possibly the software team) a decade and a half ago when they were moving to Intel was in a lot better shape. Unless I'm looking at this through nostalgia goggles. :)
- rusk 8y ago
  They were also moving to a well established architecture from their own bespoke one and now they’re going the other way ...
  - close04 8y ago
    ARM is a well established ISA. The customization Apple might bring to the CPU architecture will make no difference to a user as long as the compatibility is there. The problem is that as established as ARM is they really don't have a meaningful presence in the desktop/laptop ecosystem. You have probably tens of thousands of apps for the mobile space but even if Apple magically adapted them for a non-touchscreen, keyboard and touchpad world the vast majority still won't cater to the regular or professional user wanting to buy a laptop, not a glorified tablet. And that's a hard sell. Why am I buying a MacBook if all I get is an iPad with a keyboard? And don't get me wrong, having used both iPads and CoreM ultraportables (Apple or otherwise) the iPad offered the better experience more often than not. But if I compare it to the regular CPU it just can't compete. The productivity with mobile apps is not where it should be to convince a normal buyer to switch, let alone a more demanding user.