Apple Plans to Ditch Intel and Use Custom Mac Chips Starting in 2020

[LiveWireless]

I think this rumor is most likely false, but not completely off base. I just don't think moving away from x86 would help them as much as it would hurt them.

However, i would suggest they they will do a hybrid system. They will add an additional processor to the Mac line .... in the same box. So look for it to have an x86 processor and an Apple processor.

OK the cost of this is maybe too high, but many people i see still using Mac desktops require x86. Apple needs to expand it's desktop base, not fracture it. Apple could be just fishing for feedback, but from a business view i see the addition of an Apple processor as the only viable path, but don't have the inside information to know if it is viable.

 

[cmaier]

I think this rumor is most likely false, but not completely off base. I just don't think moving away from x86 would help them as much as it would hurt them.

However, i would suggest they they will do a hybrid system. They will add an additional processor to the Mac line .... in the same box. So look for it to have an x86 processor and an Apple processor.

OK the cost of this is maybe too high, but many people i see still using Mac desktops require x86. Apple needs to expand it's desktop base, not fracture it. Apple could be just fishing for feedback, but from a business view i see the addition of an Apple processor as the only viable path, but don't have the inside information to know if it is viable.

Hardly anybody needs x86. Some people need software that currently is only available on x86. Big difference.

 

[Michael Goff]

Not true. There is almost certainly a change in control provision in the intel cross license. Such provisions are standard. In all likelihood if AMD is acquired it loses the license. Moreover such licenses frequently contain “have made” provisions, though this one may not.

Weirdly enough, though, AMD owns the license to x64. So, yes, Intel could pull the x86 license. But AMD could pull the x64. I’m sure Apple would be alright with an OS running on an x64 only chip more than Intel would be with having only x86.

 

[cmaier]

Weirdly enough, though, AMD owns the license to x64. So, yes, Intel could pull the x86 license. But AMD could pull the x64. I’m sure Apple would be alright with an OS running on an x64 only chip more than Intel would be with having only x86.

But there’s no such thing as supporting amd64 and not infringing x86 intellectual property, unfortunately. AMD64 is a modification of the x86 instruction set - a derivative work in copyright parlance. So you need a license from intel even if you don’t support 32-bit operations.

I designed the first cut of the integer oppcodes for AMD64. I designed them to be a natural extension of the 32 bit ops. (Luckily our architect decided eventually to stick with us and he finished the job and I went back to designing logic and circuits and such)

 

[Michael Goff]

But there’s no such thing as supporting amd64 and not infringing x86 intellectual property, unfortunately. AMD64 is a modification of the x86 instruction set - a derivative work in copyright parlance. So you need a license from intel even if you don’t support 32-bit operations.

I designed the first cut of the integer oppcodes for AMD64. I designed them to be a natural extension of the 32 bit ops. (Luckily our architect decided eventually to stick with us and he finished the job and I went back to designing logic and circuits and such)

So AMD couldn’t make a chip without the license. Interesting. Well, Apple could still hold the Sword of Damocles over Intel’s head with regards to x64. I doubt Intel wants to be unable to make 64-bit processors anymore.

 

[Altis]

Hardly anybody needs x86. Some people need software that currently is only available on x86. Big difference.

So the people who need to run x86 software don't need x86 to run it?

I'm not sure anyone really needs ARM any more.

 

[cmaier]

So the people who need to run x86 software don't need x86 to run it?

I'm not sure anyone really needs ARM any more.

When their software is ported to arm they won’t need it, and most software is very easy to port.
[doublepost=1523203192][/doublepost]

So AMD couldn’t make a chip without the license. Interesting. Well, Apple could still hold the Sword of Damocles over Intel’s head with regards to x64. I doubt Intel wants to be unable to make 64-bit processors anymore.

There’s always Itanium

 

[Michael Goff]

When their software is ported to arm they won’t need it, and most software is very easy to port.
[doublepost=1523203192][/doublepost]

There’s always Itanium

Thanks. I needed a good laugh this morning.

Also, I do kind of wish Itanium had succeeded.

 

[Altis]

When their software is ported to arm they won’t need it, and most software is very easy to port.

You're assuming the software is still supported and that the devs will be willing and able to port it.

So sure, people don't need x86 if all their x86 software manages to get ported. Otherwise, they need it. Kind of meaningless.

 

[cmaier]

You're assuming the software is still supported and that the devs will be willing and able to port it.

So sure, people don't need x86 if all their x86 software manages to get ported. Otherwise, they need it. Kind of meaningless.

I’m assuming nothing. I spoke of most people and some people. There are a few people who really need x86. Most don’t.

Look, I still use illustrstor cs6 because it costs too much to buy a subscription. I will be out of luck. But that’s life. I knew that old software won’t be usable forever.
[doublepost=1523203982][/doublepost]

Thanks. I needed a good laugh this morning.

Also, I do kind of wish Itanium had succeeded.

It was a terrible design and a terrible idea. They thought they’d screwed us (AMD) because we had no license to it. That forced us to think creatively. We were going to do something horrendously complicated like Itanium, and we lost a lot of folks all at once (they quit for various reasons). So we had no choice but to start over and do so with something simple, and hence far more efficient. And designing it for multiple cores because it was easier to duplicate simple cores than design a single complicated core when you had only around 15 chip designers. Turns out that’s a much better solution for power efficiency too,

 

[Michael Goff]

I’m assuming nothing. I spoke of most people and some people. There are a few people who really need x86. Most don’t.

Look, I still use illustrstor cs6 because it costs too much to buy a subscription. I will be out of luck. But that’s life. I knew that old software won’t be usable forever.
[doublepost=1523203982][/doublepost]

It was a terrible design and a terrible idea. They thought they’d screwed us (AMD) because we had no license to it. That forced us to think creatively. We were going to do something horrendously complicated like Itanium, and we lost a lot of folks all at once (they quit for various reasons). So we had no choice but to start over and do so with something simple, and hence far more efficient. And designing it for multiple cores because it was easier to duplicate simple cores than design a single complicated core when you had only around 15 chip designers. Turns out that’s a much better solution for power efficiency too,

And yet, and no offense to anyone who works at Intel, they are not that great at power efficiency. Maybe it would have been way worse with Itanium, though.

 

[cmaier]

And yet, and no offense to anyone who works at Intel, they are not that great at power efficiency. Maybe it would have been way worse with Itanium, though.

Doubtless it would. It was at the time, for sure.

 

[Michael Goff]

Doubtless it would. It was at the time, for sure.

I’m still of the firm belief that we need to move beyond the current x86-64 at some point.

 

[Altis]

I’m assuming nothing. I spoke of most people and some people. There are a few people who really need x86. Most don’t.

Look, I still use illustrstor cs6 because it costs too much to buy a subscription. I will be out of luck. But that’s life. I knew that old software won’t be usable forever.

You assume here that their software will be ported by using "when" not "if":

When their software is ported to arm they won’t need it, and most software is very easy to port.

So back to the start, the people who need to run x86 software need x86. What's this "big difference"?

Hardly anybody needs x86. Some people need software that currently is only available on x86. Big difference.

And again, who actually 'needs' ARM, either? I mean sure, Apple seems to love taking stuff away so people have to spend more money -- but it's not a good thing for the consumer.

 

[Michael Goff]

You assume here that their software will be ported by using "when" not "if":



So back to the start, the people who need to run x86 software need x86. What's this "big difference"?



And again, who actually 'needs' ARM, either? I mean sure, Apple seems to love taking stuff away so people have to spend more money -- but it's not a good thing for the consumer.

Wait. Longer battery life, better standby time, faster wake from sleep, and at this time being the only way to have a built in cell modem aren’t good for consumers?

 

[cmaier]

You assume here that their software will be ported by using "when" not "if":



So back to the start, the people who need to run x86 software need x86. What's this "big difference"?



And again, who actually 'needs' ARM, either? I mean sure, Apple seems to love taking stuff away so people have to spend more money -- but it's not a good thing for the consumer.

The “big difference” is that other than certain very specialized users, people need particular SOFTWARE not a particular processor architecture.

Most people have no idea what the difference between arm and x86 is, and that difference makes no difference to them.

Most software will be ported or there will be alternatives available.

Some will not. (You keep telling me I assume things, but I never said all software will be ported. I said hardly anyone needs x86. I said some people need software that only runs on x86. And then I said the majority will not (because porting).)

But for the vast majority of people the change will end up making no difference at all in workflows.

 

[Altis]

Wait. Longer battery life, better standby time, faster wake from sleep, and at this time being the only way to have a built in cell modem aren’t good for consumers?

1/2: As if ARM is the only way to improve battery life. Apple could dramatically improve the battery life of Macs before needing ARM CPUs, but they can't seem to be bothered. Traditionally, Apple would just put ARM in and then put smaller batteries in so battery life would be unchanged.

3: Is anyone bothered by the < 1s wakeup time as it is? Even my 2010 is woken up by the time the lid is opened.

4: There are already LTE laptops (and pretty much nobody seems to bother with them). I'd wager more people benefit from x86 support than would opt for LTE anyways.

 

[cmaier]

Wait. Longer battery life, better standby time, faster wake from sleep, and at this time being the only way to have a built in cell modem aren’t good for consumers?

Not to mention better security.

 

[Michael Goff]

1/2: As if ARM is the only way to improve battery life. Apple could dramatically improve the battery life of Macs before needing ARM CPUs, but they can't seem to be bothered. Traditionally, Apple would just put ARM in and then put smaller batteries in so battery life would be unchanged.

3: Is anyone bothered by the < 1s wakeup time as it is? Even my 2010 is woken up by the time the lid is opened.

4: There are already LTE laptops (and pretty much nobody seems to bother with them). I'd wager more people benefit from x86 support than would opt for LTE anyways.

I’m not just referring to Apple. Look at the huge battery life increases on the Windows side on the few Qualcomm laptops out there. And that’s with software that probably isn’t fully optimized.
[doublepost=1523205661][/doublepost]

Not to mention better security.

Yep. Most big problems that aren’t tied to x86-64 are phishing. And hardware can’t stop that.

 

[Altis]

Most software will be ported or there will be alternatives available.

Some will not. (You keep telling me I assume things, but I never said all software will be ported. I said hardly anyone needs x86. I said some people need software that only runs on x86. And then I said the majority will not (because porting).)

I've highlighted your assumptions.

Sure, currently supported software will likely offer a new version that will be ARM-compatible. But there will be lots that either won't get the update (not supported) or will require buying a new version (or like you pointed out with Adobe, subscribing to new software).

I heard the same thing about USB-C, how everything would be USB-C not that the MB/MBP only have that port. Yet even years later, almost everything still comes in USB-A.

The “big difference” is that other than certain very specialized users, people need particular SOFTWARE not a particular processor architecture.

Most people have no idea what the difference between arm and x86 is, and that difference makes no difference to them.
....[snipped]....
But for the vast majority of people the change will end up making no difference at all in workflows.

Sure, people don't generally care about the inner workings. But they'll absolutely care when things don't work.

Not being able to run x86/x64 Windows, for instance, is a pretty big deal. Every single person I know with a Mac in engineering runs Windows via bootcamp.

 

[Michael Goff]

I've highlighted your assumptions.

Sure, currently supported software will likely offer a new version that will be ARM-compatible. But there will be lots that either won't get the update (not supported) or will require buying a new version (or like you pointed out with Adobe, subscribing to new software).

I heard the same thing about USB-C, how everything would be USB-C not that the MB/MBP only have that port. Yet even years later, almost everything still comes in USB-A.



Sure, people don't generally care about the inner workings. But they'll absolutely care when things don't work.

Not being able to run x86/x64 Windows, for instance, is a pretty big deal. Every single person I know with a Mac in engineering runs Windows via bootcamp.

If you take every person with a Mac in engineering and they then stopped using the Mac, Apple would probably still sell 15m+ a year.

 

[Altis]

If you take every person with a Mac in engineering and they then stopped using the Mac, Apple would probably still sell 15m+ a year.

Perhaps. It certainly isn't the only industry where Macs are only viable at all because they can run full Windows, just one example.

Would Apple be willing to throw them all away so they can what, offer LTE laptops with smaller batteries? Certainly possible considering how they've shown to only be interested in certain users.

 

[Michael Goff]

Perhaps. It certainly isn't the only industry where Macs are only viable at all because they can run full Windows, just one example.

Would Apple be willing to throw them all away so they can what, offer LTE laptops with smaller batteries? Certainly possible considering how they've shown to only be interested in certain users.

If a Mac is only viable because it runs Windows, you’re wasting your money on the Mac.

 

[Altis]

If a Mac is only viable because it runs Windows, you’re wasting your money on the Mac.

It's a matter of flexibility. Some people also prefer the hardware itself.

It could be argued that by only running Chrome and Office, most Mac users are wasting their money on the Mac.

 

[Michael Goff]

It's a matter of flexibility. Some people also prefer the hardware itself.

It could be argued that by only running Chrome and Office, most Mac users are wasting their money on the Mac.

Maybe. A Chromebook would be cheaper and can run Chrome and Office.

Edit: Then again, some people prefer the Mac Operating System for other reasons.