Intel switch from PPC questions

[dark knight]

i know this has been argued to death but despite reading quite a few articles i am still confused. keen amateur here so sorry for my ignorance

when apple shifted away from PPC i had it in my mind that one of the key reasons was that the chips were getting too hot to put into a nice slim laptop. Though everything i have read suggests that the RISC PPC was much more power efficient than the increasingly bloated and amended CISC architecture of the x86. Also, Pentiums around this time were famed for being hot.

Also, why are ARM's RISC based processors so much more suite to mobile efficiency than say, PPC. You would have thought the the much more mature PPC could have been the basis of a great mobile line.

Basically, where did PPC go wrong. I have heard of many advantages of the architecture, such as quantity of registers, simplicity, but in the end it lost the war.
thanks.

 

[roadbloc]

Heat, size and performance were the main factors I think. The first step Apple took from the Pro market.

 

[Hellhammer]

IBM concentrated mainly on server and enterprise markets, and still do. AFAIK, Apple was the only consumer-level client. It's not easy to make a chip (or even architecture) that's suitable for small laptops and power hungry servers. Apple wasn't big enough to make the development worth it, so IBM couldn't really provide Apple with the chips they wanted.

Plus the switch to Intel allowed Macs to run Windows, which has been a substantial selling point.

 

[Abstract]

You're pretty much right.

Motorola and IBM were making PPC chips. IBM only made the G3s, and were going to release G5s. Moto was making G4s.

From memory, Apple couldn't offer Mac users that were competitive with laptops that ran Intel processors (Centrino?). The G4s were actually fine in laptops, but the G4 was hitting the limit in terms of speed, and Moto couldn't offer anything faster than a ~1.42 GHz G4, or something. Then IBM released the G5, which couldn't be put into anything but desktops due to heat and power consumption, and a laptop-capable version that was supposed to come from IBM never came. Perhaps it was coming, but Apple jumped ship to Intel before we ever saw its release. I don't remember the details.

Back then, there was a loooong period where they were barely able to offer Apple users new, upgraded laptop models. If I remember correctly, updates to the 15" PowerBook almost didn't happen for 12 months. They were stuck on 1.42 GHz G4s from Moto for far too long, and that was after a long period where G4s were at 1.25 GHz. So basically, in 1.5 to 2 years, Mac users went from 1.25 GHz G4 laptops to 1.42 GHz processors, and got stuck.


PowerPC chips were supposed to be good because they were efficient. But that's like saying that the Honda S2000 used to be fast for its small (engine) size. Problem is that in reality, nobody gives a toss whether it's fast for "what it is".

 

[dark knight]

thanks for the history lesson
its bizarre that IBM didn't want to stir intel up with some great processors. Think how good it would be for the marketplace if IBM were really challenging Intel/AMD.
It then seems equally bizarre then that they would go and spend all that time and money developing the CELL (PPC based) for consumer electronics like the PS3 (though i know it has other applications).
Oh well x86 has proven beneficial in other areas as pointed out. Waiting for Ivy Bridge MBP update now! and my C2D MBP has proven itself over and over.
Just can't imagine how a company as big as IBM could have taken an inherently more efficient processor design and produced a model that was unsuitable for anything mobile

 

[maflynn]

The thing with IBM they're priorities were changing, they enticed apple to use the G5 when they were about to switch to intel. They promised apple both speed, regular speed increments and a mobile processor and the quantities they needed. Motorola was unable to provide quantities and was not really interested in improving the G4 (IBM had made its own proprietary improvements to the G4 and called it the G5)

When they signed on, IBM was unable to deliver on any of the promises, they were unable to ratchet up the speed, provide a mobile CPU and was unable to provide the quantities apple needed.

 

[interrobang]

Though everything i have read suggests that the RISC PPC was much more power efficient than the increasingly bloated and amended CISC architecture of the x86. Also, Pentiums around this time were famed for being hot.

It really depends on what timeframe you're looking at. In the mid-1990s, the PowerPC was way ahead of the Intel processors of the time. It was both faster and more power-efficient.

That didn't last.

 

[steve2112]

IBM concentrated mainly on server and enterprise markets, and still do. AFAIK, Apple was the only consumer-level client. It's not easy to make a chip (or even architecture) that's suitable for small laptops and power hungry servers. Apple wasn't big enough to make the development worth it, so IBM couldn't really provide Apple with the chips they wanted.

Plus the switch to Intel allowed Macs to run Windows, which has been a substantial selling point.

The ability to run Windows is a big selling point, and I think it has led to the dramatic growth of Mac sales since the switch. Oddly enough, theXbox 360 runs on the PowerPC.

 

[maflynn]

The ability to run Windows is a big selling point, and I think it has led to the dramatic growth of Mac sales since the switch. Oddly enough, theXbox 360 runs on the PowerPC.

Indeed, that really revitalized the rebirth of the Mac. I remember the sites that popped up that worked on creating hackintoshes and running windows on the Mac (prior to apple's release of bootcamp).

 

[Hellhammer]

Heat, size and performance were the main factors I think..

Oddly enough, theXbox 360 runs on the PowerPC.

Hmmm... I'm smelling a coincidence...

 

[steve2112]

Hmmm... I'm smelling a coincidence...

Image

Hah! I never really thought about it that way. I guess that's enough proof as to why Apple dropped the PowerPC.

 

[strider42]

thanks for the history lesson
its bizarre that IBM didn't want to stir intel up with some great processors. Think how good it would be for the marketplace if IBM were really challenging Intel/AMD.
It then seems equally bizarre then that they would go and spend all that time and money developing the CELL (PPC based) for consumer electronics like the PS3 (though i know it has other applications).
Oh well x86 has proven beneficial in other areas as pointed out. Waiting for Ivy Bridge MBP update now! and my C2D MBP has proven itself over and over.
Just can't imagine how a company as big as IBM could have taken an inherently more efficient processor design and produced a model that was unsuitable for anything mobile

Intel couldnt' compete against intel because their chips couldn't run windows. They were in a compeltey different market and still are. That was the problem.

They developed cell so it could be used by consoles and servers. There was overlap in what it could do. These applications dwarf mac sales, and don't have the same requirements for fitting it into a laptop and slim designs, etc.

"inherently more efficient" is really a misnomer. That may have been how it started, but the realities were they just needed to keep adding to it. I've read that cisc has become more risc like, and vice versa. I'm no engineer so I can't speak to the specifics or veracity of that though.

I have no doubt that if PPC processors were used by more of the computer market, they could have gotten them more efficient; they just had not reason to do so to the extent needed by apple when balanced against other priorities.