Mac OS X 10.6 to Drop PowerPC Support?

[christian_k]

Support for first gen Intel systems?

Some of you said, it makes sense to stop support for PPC's if they are unable to handle features of a new OS release. I agree.

But, it would be hard to justify this if they continue to support all Intel based systems. An owner of a high end quad-G5 with 8 GB Ram, GF 6800 or whatever that was exluded from 10.6 would be very angry to see 10.6 run on a lame and cheap first generation core solo Mac Mini with integrated gfx, 512 MB and combo drive....

There is overlap between the architectures. High end G5's do outperform a lot of entry level or even mid-range Intel based systems.

But I don't think they will do it this way. 10.5 even runs on a 2001(?) Powerbook G4 with 400 MHz with no CoreImage, no QuartzExtreme, no 64 Bit support....

Christian

 

[JFreak]

Leopard is not supporting G3 based on what I heard, iMovie does not support G3 or G4

Two different things.

OSX itself may not be supported on non-Altivec PPC and that's fine and it's called progress. But iLife is not a part of the operating system but an additional bundle. If you can run the operating system but cannot run some apps, that's fine too. My father's G3 iBook never ran any audio apps that required Altivec and I never complained about it because such apps clearly require Altivec unit on the CPU. And if the OSX now uses Altivec somewhere, it makes perfec sense to drop support for non-Altivec PPC hardware.

Dropping support just for the sake of dropping something is just not wise and therefore not going to happen.

The only fundamental difference between G3 and G4 is Altivec unit and only fundamental difference between G4 and G5 is 32/64 bitness. It would make sense to only support G5 or not support PPC at all if the next big cat (10.6) would be 64-bit only — but WHY ON EARTH would they make it 64-bit only as they now tout the wonderful feature of letting 32bit and 64bit hardware live nicely together under the same roof? No, that's not going to happen.

There is no technical reason why G4/G5 support would have to be dropped so I can see Apple supporting PPC hardware for some time and I think this speculation is nonsense. It will happen eventually, but not just yet. We had Classic support on OSX for years and I bet Rosetta will be here for some time also. If you don't need to use it, it does no harm. It will not go away just yet as there are tons of software still not converted to UB and some never will as the development has been stopped and such apps are still being widely used.

But it's always nice to speculate, isn't it

10.6 won't come until 2009 or 2010... can PPC users still reasonably expect to have updates at that point?

Yes. We're not living in a Windows world where everything becomes obsolete the same day you buy it. We expect being able to update our Macs for at least 5-7 years. If they come up with something that is so demanding to the hardware that Quad PowerMac cannot handle it, then I'm cool about not supporting PPC hardware. Until that happens, I'm going to assume there is support for it.

That said, I think I may be upgrading my Quad PowerMac before 2012 but that's not the point. I expect that thing to still be quite usable at that time.

10.5 should drop PPC entirely and be Intel-only. this is what is truly holding back technological progress in this country---the constant need for backward compatibility. it's why windows is such a mess.

No. Windows is a mess, because it is coded to be a gigantic mess. It is not a mess because it supports plenty of hardware, but because the software itself is a mess.

OSX will not become any better if they just dropped the PPC support making it Intel-only. It would become few hundred megabytes smaller, but that would be the only benefit. And because storage space tends to grow, I'd say the benefits [of dropping PPC code from the release] are close to zero.

 

[christian_k]

...and only fundamental difference between G4 and G5 is 32/64 bitness. It would make sense to only support G5 or not support PPC at all if the next big cat (10.6) would be 64-bit only — but WHY ON EARTH would they make it 64-bit only as they now tout the wonderful feature of letting 32bit and 64bit hardware live nicely together under the same roof? No, that's not going to happen.

They won't do that.
There are a lot of Intel based Minis, MacBooks, MBPs and iMacs that don't have 64 bit CPUs. The Mini got the C2D only 2 months ago. When 10.6 comes out in 2009, they won't exclude systems that are only two years old.

Christian

 

[JFreak]

Given that Intel Macs started in Tiger, I doubt Leopard will be the last PPC supported system, The base of G4 & G5 PowerMacs is HUGE and will continue to dominate the overall population of Macs by 2009. Quad G5 is not going to be dead by then unless they all start leaking bloody murder. It is only a little slower than the Quad 2.66 Mac Pro.

Only slightly slower on UB apps but very much faster on PPC-only apps that MacPro would have to run with Rosetta and INFINITELY faster on PPC-only that don't work with Rosetta.

 

[bimmzy]

As we approach the release of Apple's Mac OS X 10.5 (Leopard), Appleinsider suggests that this may be the last major Mac OS revision that supports the PowerPC architecture.

Cited as speculation, Apple may not yet have made a decision on this matter. Apple has made no announcements about Mac OS X 10.6. Based on the most recent Mac OS development cycles, Mac OS X 10.6 would not likely ship until 2009.

Article Link

There will always be a PPC/RISC port of the Mac OS.
Apple will always hedge its bets, and who knows one day they may transition back to PPC.
Lets face it the fastest processors around currently are the Power 6 series running at speeds of 4GHz+ faster than Intel offerings today.
With future 'Power' technology tantalizingly offering 100GHz+ who knows what apple will do.
What you can be sure of is that apple will have the latest PPC port, its just that Apple may not ever release it! haha

 

[Manic Mouse]

There will always be a PPC/RISC port of the Mac OS.
Apple will always hedge its bets, and who knows one day they may transition back to PPC.
Lets face it the fastest processors around currently are the Power 6 series running at speeds of 4GHz+ faster than Intel offerings today.
With future 'Power' technology tantalizingly offering 100GHz+ who knows what apple will do.
What you can be sure of is that apple will have the latest PPC port, its just that Apple may not ever release it! haha

And how hot do Power 6 processors run? It's not merely about speed and power but also about price and heat. With all factors considered the Intel chips are by far the best at the moment, which is why Apple are using them. They also give the added advantage of being able to run windows natively should it be required.

 

[opeter]

And how hot do Power 6 processors run? It's not merely about speed and power but also about price and heat. With all factors considered the Intel chips are by far the best at the moment, which is why Apple are using them. They also give the added advantage of being able to run windows natively should it be required.

The power consumtion and heat is not the problem of Power 6... the G5 was a lite version of Power 4 CPUs anyway. The problem of Power 6 is that it was not constructed to be used as a desktop or mobile procesor.

It was meant for supercomputing and servers.

 

[Masquerade]

AltiVec still has superior permute operations among a few others. SSE4 brings some additional instructions that close this gap.

simple as the e=mc^2

 

[EDL]

Obsolete

I would hope that it would be 10.7 that would end PPC support. Think about all the PowerMac G5's out there in the industry because people couldn't wait for Adobe to release the Native CS3 apps.

That would mean that all the G5's out there sold last year would be obsolete in 2 years if 10.6 took away PPC support all together? No way. Especially when some of the last model PowerMacs still rival the current Intel line in performace and handling of applications.

Come on Apple give the G5's at least 4 more years. That's not asking a lot.

Otherwise you could easily see another lawsuit similar to the Beige PowerMac G3 suit over OS X.

Not really, all the working PMG5 won't be obsolete since they're still running up to and including Leopard. No one forces you to buy an operating system, all your choice. And then, why would you want to have the newest high-tech operating system and run it on an almost 5 year old machine??? Paradox! That's like running Windoze Vista on a Pentium 4 2GHz... well, good luck with that (sorry, I don't mean to call Vista high-tech but you get the idea...).
Should there be a CS4 or CS5 that requires a new machine, then you have to upgrade. If the costs are too high for you, then you're not using the machine/software to make money, otherwise you would gain the money back through more efficiency... hence, you don't need either CS4/5, nor do you need a faster machine and OS, you just want it. Well, I want a lot of things too, doesn't mean I can have it...

And yes, it is asking quite a lot from Apple to develop an operating system for two different platforms. You might want to consider that Apple is developing a bunch of other industry-leading software as well plus a very ambitious line of hardware products... and of course, everyone expects Apple to have the most innovative product on the market with the next following update. There's no other company in the industry or even on the whole planet that comes close to fulfilling such high expectations like Apple does. Instead of always demanding more and more, just say thank you for what they're doing and appreciate it.

 

[guzhogi]

Whics apps you mean?

I work in an elementary school & we use a bunch of phonics, math & game apps that are classic apps only. Examples are Phonics Express, Math Keys, Where in the World is Carmen Sandiego. Great apps, just never upgraded for Mac OS X. Plus, since we're a public school, we don't have the money to upgrade every time there's a new version of the software. And, it takes awhile to test how new software affects the rest of the system.

As for the people who are saying that G5s are more powerful than even the entry level Intel Macs, good point. Power Mac G5s did have a lot of oomph. On the other hand, it may not necesarily be about the power of the processor, maybe it's just about too many architectures to support. Who knows, maybe Apple will keep a PowerPC version of Mac OS X in hiding like they did before the Intel transition. Only time will tell.

 

[-hh]

Apple will always hedge its bets, and who knows one day they may transition back to PPC.
Lets face it the fastest processors around currently are the Power 6 series running at speeds of 4GHz+ faster than Intel offerings today.
With future 'Power' technology tantalizingly offering 100GHz+ who knows what apple will do...

And how hot do Power 6 processors run? It's not merely about speed and power but also about price and heat. With all factors considered the Intel chips are by far the best at the moment, which is why Apple are using them. They also give the added advantage of being able to run windows natively should it be required.

Sure, "but..."

For one thing, how much do I really need to care about power/heat when my application is serious desktop iron? Or do you really think that it was pure coincidence that the first Intel Macs were in the laptops, and that the tower came utterly dead last?

Next, how is it that we're able to predict with 100% infallability that Intel will continue to be the best "desktop application" chip designer for the next 10-20 years? There's a ton of really interesting 'stuff' going on out there that's not merely evolutionary (smaller), but revolutionary, and there's a lot of players other than Intel that are getting DARPA funding and so forth. As such, the potential for a 'break out' is IMO increasing. For example, there's ground work being done on optical chips as well as diamond semiconductors which demonstrated 80+GHz three years ago...its only a matter of time until the manufacturing technology and other aspects catch up and make these currently highly exotic efforts down into the realm of affordable.

Finally, if Apple's strategic objective is to develop hardware-independent code (see 'hedging bets'), as someone awhile back said, maintaining support & testing on multiple hardware incarnations serves as a highly effective form of Quality Control to make sure that the code successfully stays abstracted. This means that maintaining of PPC support alligns with their strategic goals.

FWIW, I'm of the general opinion that OS Support will (hopefully) continue, but this will not obviate specific software packages from requiring higher hardware minimums. That's hardly anything new or profound, and we've just seen it occur in the latest revision to iMovie (iLife '08).


-hh

 

[johnnyjibbs]

My 12" PB G4 (1GHz) is getting towards the lower end of spec to run Leopard, but it will run it fine, just as it runs things really well at the moment. It's 4 years old this month (but you could still buy my model 3 and a half years ago) and, more than anything, it's the hard disk space that I'm going to be struggling with.

I upgraded my hard drive at purchase through Apple so as to be the max 80GB at the time. With music and photos taking up nearly 40GB at present alone, I'm really struggling with space. I've cleared it to about 5GB left, but I'm still struggling to delete anything more, without deleting any apps. (Classic comes to so little that I don't want to delete it.) My main concern is that Leopard is weighing in at 9GB. But that is the only limitation I'm up against. My machine could continue running, I envisage, perfectly fine for years to come. So I'd prefer that it would not be excluded in the future, even though I know it makes sense.

But 10.5 isn't even out yet, so let's not worry about 10.6.

 

[flopticalcube]

Can you give me a source? I can't find information to back your statement.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2748&p=3

In earlier designs such as the P6 (Pentium Pro, PII, PIII) architecture, these instruction would have been broken up into two or even three micro-ops. Remember that the whole philosophy behind all modern x86 CPUs, since the P6, is to decode x86 instructions into RISC-y micro-ops which are then fed to a fast RISC backend; the backend then schedules, issues, executes and retires the instructions in a smooth RISC way.

<Mr Mackey voice>
Posting old information as new is bad, m'kay?
</Mr. Mackey voice>

 

[scorp888]

Yup. Seems much more likely to me. It gives G5'ers plenty of time to
consider upgrading.

Except everyone seems to forget that the Powerbook was available at the same time as the Quad G5, indeed, Apple ended the life of its 15" PowerBook G4 on February 22, 2006.

For a lot of people portable Photoshop = G4, as until CS3, it was quicker on a G4 than on intel.

I would assume that 1GZ and below G4 will be excluded 10.6, or perhaps limited functionality as per other os's.

You get base functionality G4 1ghz + 1.5 ghz most of the glitz.

"full experiance" Intel Core solo or above.

At the moment, if they are supporting PPC on servers they'll keep on supporting it on Desktop.

10.7 Bengal, different story.

 

[scorp888]

(I don't care if a diesel engine is technically superior to a gasoline engine, the fact is that gasoline engine manufacturers have overcompensated to the point that diesel engines just don't compete, except in a few very tightly controlled circumstances. Although in this comparison, diesel is MUCH closer to matching and surpassing gasoline than PPC is to matching x86.)

Bad analagy given that Audi just won the Lemans with a Diesel car.

It's all about investment, intel invested a whole heap more money in x-86 that anyone else did in anything else, they could afford to, they sold more, so they invested more...circle of fullfillment.

Diesel, lots of investment shared across multiple manufacturers, means progress is being made on diesels that hasn't been made in years as focus was on the Petrol ICE. Although saying that, petrol engines are also getting far more efficient.

However, if you look at Diesel bikes, you can see how far we can go.

200 mpg anyone...?

 

[guzhogi]

(I don't care if a diesel engine is technically superior to a gasoline engine, the fact is that gasoline engine manufacturers have overcompensated to the point that diesel engines just don't compete, except in a few very tightly controlled circumstances. Although in this comparison, diesel is MUCH closer to matching and surpassing gasoline than PPC is to matching x86.)

Bad analagy given that Audi just won the Lemans with a Diesel car.

It's all about investment, intel invested a whole heap more money in x-86 that anyone else did in anything else, they could afford to, they sold more, so they invested more...circle of fullfillment.

Diesel, lots of investment shared across multiple manufacturers, means progress is being made on diesels that hasn't been made in years as focus was on the Petrol ICE. Although saying that, petrol engines are also getting far more efficient.

However, if you look at Diesel bikes, you can see how far we can go.

200 mpg anyone...?

My family had a diesel-powered Volvo a few years ago. Great car, built like a tank. The only problem w/ it is in the winter, the diesel actually gelled up and often and to get towed or something. Don't know if there's any comparison w/ Intel/PowerPC.

 

[CmdrLaForge]

Even if its final update happened in the fifth year of its life, it would still be a valuable computer for a good three years after that. I think that eight years out of a computer is pretty darn good. By that time, you'd want the better specs anyways, so buying a new one wouldn't be such a bad thing

I always want the new specs after 2 years. Hmmm.

 

[CmdrLaForge]

But I don't think they will do it this way. 10.5 even runs on a 2001(?) Powerbook G4 with 400 MHz with no CoreImage, no QuartzExtreme, no 64 Bit support....

Christian

Well - no it doesn't. 10.5 needs a higher CPU speed then 800 MHz and a G4.

But you are right when you say it would be lame not to support G5 systems with 10.6.

 

[leekohler]

... your PowerMac will be just old enough to go to that special farm...

You mean Server-land?

 

[Erasmus]

Well - no it doesn't. 10.5 needs a higher CPU speed then 800 MHz and a G4.

But you are right when you say it would be lame not to support G5 systems with 10.6.

Well, I intend to try installing Leopard on my 450MHz G4 Cube, so I guess I'll get back to you on that. I would expect it to run, just without all the pretty eye candy, just like Tiger.

Next, how is it that we're able to predict with 100% infallability that Intel will continue to be the best "desktop application" chip designer for the next 10-20 years? There's a ton of really interesting 'stuff' going on out there that's not merely evolutionary (smaller), but revolutionary, and there's a lot of players other than Intel that are getting DARPA funding and so forth. As such, the potential for a 'break out' is IMO increasing. For example, there's ground work being done on optical chips as well as diamond semiconductors which demonstrated 80+GHz three years ago...its only a matter of time until the manufacturing technology and other aspects catch up and make these currently highly exotic efforts down into the realm of affordable.

Sure, but just as a thought, out of all the CPU manufacturing companies, who has the funds to firstly do all this research, and then start producing? I would have thought Intel would have the most money to just drop on a project like this.

Also, I would have thought that Intel has the best CPU designs available at the moment. And their future lineup seems to just get better and better. Most likely, whatever a liquid cooled Power 6 CPU could do, an air cooled Xeon could do. What does this mean for you, who doesn't care how hot the chip runs? It means this: Stick a liquid cooling block on the Xeon, and you can overclock it way past the Power 6.

If IBM makes an 80GHz diamond semiconductor Power 6 CPU on a 90 nm process (or whatever they use atm), and Intel makes an 80 GHz diamond semiconductor Xeon on a 65 nm process, I'm guessing that they would both have similar power and cooling requirements, but that the Xeon would blow the Power 6 away. After all, high GHz and hot chips are just compensations for inferior chip design.

 

[CalBoy]

I always want the new specs after 2 years. Hmmm.

No. You have to finish up those G5s on your plate, and then dessert: a juicy new Mac Pro.

You mean Server-land?

Shhhhh! Don't let him hear you say that.

 

[leekohler]

Shhhhh! Don't let him hear you say that.

Oh no! Too late!

 

[Gav]

Well, it'll definitely be gone in OS 11.

 

[-hh]

analogies - - trivia.

My family had a diesel-powered Volvo a few years ago. Great car, built like a tank. The only problem w/ it is in the winter, the diesel actually gelled up and often and to get towed or something. Don't know if there's any comparison w/ Intel/PowerPC.

Probably not, but...

If "diesel" fuel were fundimentally incompatible with cold temperatures, jet aircraft couldn't fly (since JP8 is closer to diesel than it is to gasoline).

The gelling of automotive diesel fuel is frequently because of a too-high parafin ("wax") content in the fuel. Diesel is quite similar to #2 home heating oil, and the latter typically includes wax as a cheap source of hydrocarbons. The fuel suppliers are supposed to lower the wax content in the winter to prevent any gelling problems ("winter mix"), but some suppliers load up on the last of the cheaper "summer mix" and its higher wax content ends up messing up their customers this way (both home and automotive) when it gels during a cold snap. It can also take some time to get wax out of your fuel tank too.

In Europe, some diesel vehicles have heated fuel filters (some OEM, some Aftermarket), as it is on the paper element pass-thru where the clog normally occurs. A friend's father was an aircraft mechanic who owned a diesel Rabbit: he got fed up with the low-grade diesel here in the USA, so he modified a heated fuel filter element from a Boeng 737 to use on his Rabbit...problem solved!


-hh

 

[-hh]

Originally Posted by -hh
Next, how is it that we're able to predict with 100% infallability that Intel will continue to be the best "desktop application" chip designer for the next 10-20 years? There's a ton of really interesting 'stuff' going on out there...

Sure, but just as a thought, out of all the CPU manufacturing companies, who has the funds to firstly do all this research, and then start producing?
I would have thought Intel would have the most money to just drop on a project like this.

I'd agree, if it wasn't for "outside forces" with money, such as DARPA.

For example, its not the Intels, IBMs or AMDs who's currently paying the big bucks for chips to do hyperspectral imaging and analysis: its the Government. As such, whoever DARPA selects to do this research will invariably have a leg up on their competition when it comes to its commercialization spin-offs. And the simple facts are that Intel does not have 100% of these research contracts.

Also, I would have thought that Intel has the best CPU designs available at the moment.

They may, particularly in certain specialized fields, such as General Purpose CPUs. However, this doesn't mean that they're leading in every field across the board. An example of this very well may be exotic material semiconductors:

If IBM makes an 80GHz diamond semiconductor Power 6 CPU on a 90 nm process (or whatever they use atm), and Intel ...

We can stop right there: upon what basis are we making the assumption that Intel has made an independent foundry investment that's equal to what IBM (or others) have done, either on their own, or with Government sponsorship?

My point is that with the research that is occuring, I expect that we're going to see a change in base semiconductor material coming out of the foundries within the next few years. As such, there's going to be early adopters who have the potential for a "transformational" lead, at least until the other players are able to make the parallel investment.

As such, the implication is that if Company A has a 80GHz chip, and no matter how "inefficient" its design, simple brute force is going to trump Company B's highly elegant ... but only 8GHz ... product, at least for some applications.

After all, high GHz and hot chips are just compensations for inferior chip design.

By today's paradigm standards, agreed.

But some technologies are paradigm-shifts, and therein lies the rub.

For a (poor) analogy, consider an aircraft jet engine vs a piston engine: the jet runs hotter and has a higher RPM, all of which make it a poor candidate for flying low and slow. But for a transcontinental flight, do you want to go low and slow? Probably not. Since piston aircraft have a fairly difficult time exceeding 25,000ft and 300mph, which technology would you rather have for your next intercontinental flight?


-hh