Is the PowerPC architecture Dead?

[furryrabidbunny]

Apple just abandoned the powerpc architecture, and the only other mainstream PC company using it is IBM in its highend servers. Other than the next generation of video game consoles, it seems to have no future. Have we seen the last generation of the processor type?

 

[Manzana]

don't think so...

the console market is rather large...and well you may find that there are many people that think that the future for ppc architecture is much better than for x86. it's just that x86 seems to always win out, kinda like vhs over betamax!

i don't really know where intel is going with their pipeline as it seems stalled under 4 ghz right now, and to me it seems there was much more promise in the ppc/cell.

also i don't think a low volume pc manufacturer (apple) switching to intel will necessarily kill ppc.

 

[bosrs1]

Apple just abandoned the powerpc architecture, and the only other mainstream PC company using it is IBM in its highend servers. Other than the next generation of video game consoles, it seems to have no future. Have we seen the last generation of the processor type?

For computers probably, but that is why Apple is leaving. It's a dying architecture...

 

[Sun Baked]

PowerPC isn't dead.

But the PPC970 likely is, and the Power5-UL also (which is a shame since it would have been better than a pentium.)

The PPC970s problem was it's complexity -- the G3/G4 were as simple for a developer to use as a lawnmower engine, while the PPC970 was a bit more like a modern vehicle engine with scads of electronics.

Which meant that embedded companies were not very willing to swap out a G3/G4 for the PPC970.

It was not worth the effort to rebuild a Cisco switch to use the PPC970 when swapping in a faster G4 worked just as well, plus some of the G3's with their accelerated Ethernet engines weren't bad choices either.

Apple alone couldn't support the PPC970.

Edit:

Look at it this way -- three chipset vendors licensed the ElasticIO bus, nobody released a chipset.

Scads of computer makers looked at the PPC970, but only one company besides Apple ever delivered a PPC970 computer to market.

Basically the white box PowerPC died after the G4.

 

[bousozoku]

Apple just abandoned the powerpc architecture, and the only other mainstream PC company using it is IBM in its highend servers. Other than the next generation of video game consoles, it seems to have no future. Have we seen the last generation of the processor type?

IBM doesn't use PowerPC processors in their high end servers, they use Power processors.

The problem with the architecture has always been that it works better in multiples than at high clock rates. IBM often puts auxiliary processors in its machines to avoid the main processor becoming a bottleneck. This is not a problem when you're spending $40,000 - $2,000,000 on a machine but for something you'd put on the desktop, you can't afford to do it.

IBM saw the problem when the PPC604e hit 350 MHz as a maximum. It would have been appropriate to switch then. 5 years later and many teeth gnashed, Apple and we are finally making the switch.

 

[bosrs1]

IBM doesn't use PowerPC processors in their high end servers, they use Power processors.

The problem with the architecture has always been that it works better in multiples than at high clock rates. IBM often puts auxiliary processors in its machines to avoid the main processor becoming a bottleneck. This is not a problem when you're spending $40,000 - $2,000,000 on a machine but for something you'd put on the desktop, you can't afford to do it.

IBM saw the problem when the PPC604e hit 350 MHz as a maximum. It would have been appropriate to switch then. 5 years later and many teeth gnashed, Apple and we are finally making the switch.

Yeah, problem with switching then is it would have killed Apple. Steve had just come back and brought OSX. He deemed that a better and more needed change at the time correctly, and he made it. Now we're getting his other planned change. The fact OSX has been compliled for Intel since it's inception is proof this change was always coming... it just took years to arrive.

 

[mvc]

Yeah, problem with switching then is it would have killed Apple. Steve had just come back and brought OSX. He deemed that a better and more needed change at the time correctly, and he made it. Now we're getting his other planned change. The fact OSX has been compliled for Intel since it's inception is proof this change was always coming... it just took years to arrive.

Exactly, they couldn't change OS and Platform simultaneously, even Steve's RDF can only stretch so far.

So this change was probably inevitable, considering how little leverage Apple actually has on its chip makers at 3% marketshare, and much better done now than when we have stagnated for another two years. It's just the logical end of a long process of standardisation that began when Steve returned to Apple.

We won't miss PPC in 5 years any more than we miss Nubus slots, ADB Ports or SCSI conflicts.

 

[BornAgainMac]

PowerPC can be used in gaming systems, cars, TVs, and doorbells. It won't go away. In two years, it will be time to upgrade again and I look forward what Intel has in 2007. Another bonus is that if the PowerPC makes a huge development going forward then Apple can still use it on some Macs.

 

[jane doe]

PPC as you know it today is dead.... Development will slow over the next year at Apple. What development going on from that point will be to specialize it for certain console game units.

 

[mad jew]

If IBM gets its act together and whips out some crazy-cool G5s, will Apple maintain the relationship and continue to use them?

 

[feakbeak]

If IBM gets its act together and whips out some crazy-cool G5s, will Apple maintain the relationship and continue to use them?

No, you don't pull a publicity event like this and then change your mind. This has been planned for years. In fact, I highly doubt Apple's been compiling their OS for x86 for five years now as a contingency plan. This was deliberately executed over serveral years, but you don't go bashing the PPC roadmap until you've got your transition plan ready to unveil. Between creating OS X to run on x86, preparing easy developer support for fat binaries and licensing emulation software (Rosetta) it is obvious that this has been in the pipeline for years. This is not just a publicity stunt to threaten IBM to put more development into PPC.

It is too much work to maintain support for two architectures indefinitely. Over time, OS X apps will be optimized for SSE rather than Altivec and things will slowly become snappier on x86 than they are on PPC. Maintaining two platforms is inefficient and would ultimately weaken the OS X computing experience.

 

[furryrabidbunny]

If IBM gets its act together and whips out some crazy-cool G5s, will Apple maintain the relationship and continue to use them?

I actually made a comment about this in another thread. Although it is extremely unlikely, what if IBM suddenly found a way to produce a ton of chips at high clocks speeds, and solved their heat issues. Then Apple might be screwed. They would have dropped a faster chip to be like everyone else. Although, its extremely unlikely.

 

[GFLPraxis]

Apple just abandoned the powerpc architecture, and the only other mainstream PC company using it is IBM in its highend servers. Other than the next generation of video game consoles, it seems to have no future. Have we seen the last generation of the processor type?

Nintendo is already using PowerPC and Microsoft and Sony WILL be using PowerPC. Video game consoles are a huge market.

Further, Apple won't abandon PowerPC for two years MINIMUM. They'll start introducing Intel Macs next year, and the transition will be "almost complete" in two years. If Intel stops speeding up or IBM has a breakthrough, Apple will keep selling PowerPC Macs for years to come. No reason to abandon them when you have fat binaries.

 

[GFLPraxis]

I actually made a comment about this in another thread. Although it is extremely unlikely, what if IBM suddenly found a way to produce a ton of chips at high clocks speeds, and solved their heat issues. Then Apple might be screwed. They would have dropped a faster chip to be like everyone else. Although, its extremely unlikely.

Not at all- for the next five years at LEAST developers will still be using fat binaries. Likely at least seven since Apple has already said they will be selling PowerPC Macs well into 2007 (add five years till the 2007 Macs are obsolete). Apple could very well reintroduce PowerPC's at ANY time if IBM makes a break through as long as developers are still using fat binaries.

 

[DakotaGuy]

If IBM gets its act together and whips out some crazy-cool G5s, will Apple maintain the relationship and continue to use them?

Not a chance, they can't continue to flip-flop between 2 platforms. It would require a lot of extra work in the future.

The x86 is a far more modern processor then the PowerPC and sometimes you have to admit you were wrong and move on. I still like my PPC systems, but I can see Apples thinking on this one.

 

[Nermal]

Steve said that they have some great PowerPC products lined up. I don't think those were his exact words, but that's the gist of it. So no, I don't think it's dead yet.

 

[MontyZ]

It's not dead, it's just wounded right now.

 

[mad jew]

The x86 is a far more modern processor then the PowerPC and sometimes you have to admit you were wrong and move on.

Okay, I'm not big on the CPU know-how, but I got the impression the PPC is more advanced but thanks to a relatively small developer base (IBM and Motorola), it isn't as fast at the moment. The x86 processor is further along in terms of development but the PPC has more potential. Is that the case or have I just eaten too much Apple Marketing Magic Pie?

 

[Tilmitt]

The x86 is a far more modern processor then the PowerPC and sometimes you have to admit you were wrong and move on. I still like my PPC systems, but I can see Apples thinking on this one.

x86 is not far more modern than PPC. x86 was developed in the 1970's. PPC was developed from POWER for the AIM Ailliance in 1991. Call this transition what it is: A step to a backwards, inefficient architecture that is faster by brute force and even then not faster in alot of areas.

 

[840quadra]

x86 is not far more modern than PPC. x86 was developed in the 1970's. PPC was developed from POWER for the AIM Ailliance in 1991. Call this transition what it is: A step to a backwards, inefficient architecture that is faster by brute force and even then not faster in alot of areas.

How much of a part in AIM does apple have? If power PC becomes faster in 2006 does apple have any rights to the PowerPC ?

or did AIM die in the 90's ?

 

[Platform]

Steve said that they have some great PowerPC products lined up. I don't think those were his exact words, but that's the gist of it. So no, I don't think it's dead yet.

But this does not really seem to be a good time to buy an PPC computer "a Mac"

 

[cube]

Not a chance, they can't continue to flip-flop between 2 platforms.

They don't need to flip-flop, they could develop multiple platforms like other vendors do.

 

[furryrabidbunny]

To comment on the x86 architecture I was reading somewhere (i wish i could find the article, it might be the the rss feed of wired) that apple may not be using x86 for long with intel. Intel is tired of x86 and they know it is about to bottleneck, and they want to develop new architecture, but Microsuck has been holding them back because they don't want to write a new OS. Intel has a lot more invested in this then it looks at face value. If intel enables Apple to get a larger marketshare, Intel will be able to develope their next generation architecture, because Apple would more than be behind them if it means faster, cooler, better Macs. Again, i really wish i could find the article.

 

[DakotaGuy]

To comment on the x86 architecture I was reading somewhere (i wish i could find the article, it might be the the rss feed of wired) that apple may not be using x86 for long with intel. Intel is tired of x86 and they know it is about to bottleneck, and they want to develop new architecture, but Microsuck has been holding them back because they don't want to write a new OS. Intel has a lot more invested in this then it looks at face value. If intel enables Apple to get a larger marketshare, Intel will be able to develope their next generation architecture, because Apple would more than be behind them if it means faster, cooler, better Macs. Again, i really wish i could find the article.

Well I don't know if that is a great development or not. Apple dropping PPC now for x86 and then turns around in a few years after that for something all new...it could get messy.

x86 is the future for as long as most businesses can project. Not to say that won't change way down the road, but I would hate to see Apple start platform hopping like no tomorrow...loosing your footing in the market is not good.

 

[MontyZ]

Too bad Apple can't just develop their own processor chips. I think that would be damn cool: Apple Inside