3 ms·

It didn't work out because Apple didn't have the in-house talent to pull it off. They had to rely on IBM and Motorola to do the actual chip design, while Apple

It didn't work out because Apple didn't have the in-house talent to pull it off. They had to rely on IBM and Motorola to do the actual chip design, while Apple focussed on creating the software for these chips. IBM and Motorola started to lose interest in promoting PowerPC in anything but the embedded space and server markets. This time around, Apple has brought all the talent they need to do all of the design in-house. No more design-by-committee. We're dealing with Steve Jobs after all ;)

Tamerlin 17y ago
Building a processor that can compete with the x86 juggernaut is expensive. Look at Intel's budget -- they have the production capacity and engineering resources to do pretty much whatever they feel like doing. For Apple to attempt to compete with Intel, it would need to be willing to invest billions of dollars just in fab technology with a minimum of 3-5 years before seeing any return on that investment. And Apple doesn't have the volume... so it wouldn't get the return on investment. So Apple worked with IBM and Moto, but weren't buying enough CPU's from either one to make much of a difference, so after Jobs ended the clone program, the PPC consortium focussed primarily on embedded computing, in particular networking. Apple wasn't big enough to merit a custom chip, so it had to make do with a chip designed for a network router. And so on... What they're doing with this in-house design probably differs in several ways. One is that it's probably aimed at a hand-held device, or something in a similar form factor. As such, it doesn't need to be able to compete with Intel's flagship x86, so Intel's technology edge isn't a factor. Two is that being aimed at a hand-held, it's going to be small -- which means a lower cost for manufacture. So it doesn't have to have Intel's manufacturing capacity to gain from economies of scale. Three is that it doesn't require that Apple builds its own fab; it can develop a design in-house and bid it out for fabrication to any of the big fabrication shops out there, like Chartered, UMC, TSMC, Samsung, etc. The fab process won't offer the same performance (i.e. clock speed ramping) as those of AMD and Intel, but in the hand-held market, that's a lot less important than minimizing power consumption. And if the first fab partner can't keep up, owning the design means that Apple can follow Microsoft's example and hire a 2nd fab partner to boost production. I guess that's a long winded way of agreeing with you, but there you go :)

Tamerlin 17y ago

Building a processor that can compete with the x86 juggernaut is expensive. Look at Intel's budget -- they have the production capacity and engineering resources to do pretty much whatever they feel like doing. For Apple to attempt to compete with Intel, it would need to be willing to invest billions of dollars just in fab technology with a minimum of 3-5 years before seeing any return on that investment. And Apple doesn't have the volume... so it wouldn't get the return on investment. So Apple worked with IBM and Moto, but weren't buying enough CPU's from either one to make much of a difference, so after Jobs ended the clone program, the PPC consortium focussed primarily on embedded computing, in particular networking. Apple wasn't big enough to merit a custom chip, so it had to make do with a chip designed for a network router. And so on... What they're doing with this in-house design probably differs in several ways. One is that it's probably aimed at a hand-held device, or something in a similar form factor. As such, it doesn't need to be able to compete with Intel's flagship x86, so Intel's technology edge isn't a factor. Two is that being aimed at a hand-held, it's going to be small -- which means a lower cost for manufacture. So it doesn't have to have Intel's manufacturing capacity to gain from economies of scale. Three is that it doesn't require that Apple builds its own fab; it can develop a design in-house and bid it out for fabrication to any of the big fabrication shops out there, like Chartered, UMC, TSMC, Samsung, etc. The fab process won't offer the same performance (i.e. clock speed ramping) as those of AMD and Intel, but in the hand-held market, that's a lot less important than minimizing power consumption. And if the first fab partner can't keep up, owning the design means that Apple can follow Microsoft's example and hire a 2nd fab partner to boost production. I guess that's a long winded way of agreeing with you, but there you go :)