This is a cut and paste of a post I made on another topic, but it is strangely relavent:
Motorola have been planning to "kill" the PPC series for a while now (by Kill I mean give over exclusive rights to Apple). It's been no secret in the industry. It was, IMHO, a very unwise decision to choose the PPC simply for this reason.
Compared to AMD, for example, Motorola will always be a few steps behind, they don't have market forces on their side. ARM have stolen the embeded market and the x86 has stolen the desktop market.
It doesn't matter if a PPC is more powerful, at the same clock speed, than an x86. AMD and Intel can always make their chips run faster and can produce them inquantity. The x86 gets constant revision throughout it's life cycle (pushing performance up daily), the PPC only gets an upgrade once every 6 months or so.
It is important to remember that when a company like AMD release a chip, they keep redesiging it!! if you buy an Athlon one month and then you buy another one a few months down the line, you get a quite different chip with new technology!
This can be seen in the Intel P4, the first chips sucked ass, but now they have seen where improvemnts need to be made and implemented them to make the chip better (the P4 still sucks though :-) ). Motorola simple don't have a big enough market to keep this sort of constant development going. The x86 chips get faster because of newer technology, not because someone puts a faster crystal on. The 600Mhz Athlon and the 1.8Ghz Athlon run at the same themperature!!!! Run a G4 700Mhz at
1Ghz and you need some serious cooling upgrade.
IBM do make a PPC series of chip, but it is really expensive for the performance, and IMHO it is not really suited to the desktop. The IBM engineers have been pushing great speeds out of these chips but they have not been able to produce them in any significant quantity.
I think it is important to remember that Amiga Corp. and apple chose the 68k because it was the fastest thing around for the price. I think Amiga Inc. should remember that.
It would have been great if Motorola had continued the 68K, like Intel did with the x86. By now the two chips would have been almost indentical anyway!!!* But Motorola didn't have a big enough market to do that and teamed up with IBM etc to make the PowerPC chips instead, which in theory would have allowed for a bigger market and allowed them to jump over the x86 emulation stage. But the PPC did not take off, maybe a more powerful 68k would have done? Isn't hindsight great
Apple are very lucky, they have their own CPU which they can control, but they are never going to be able to beat the x86 chips with the PPC, IMHO!
*I have spent a long time thinking about the evolution of the 68k, Motorola would have added a simd unit (maybe even Altivec?). Then they would have continued with the 060 RISC core and, then culled more of the less usefull instuctions and then added a few newer instructions (conditional moves etc...) to allow better branch prediction. Then the bus interface would have been changed to allow better cache and multiprocessor support. The RISC core would have grown lots of registers, to allow register renaming and out of order execution. The chip would be fully super scalar and super pipelined. As you can see, now the intenals of the chip would look just like an Athlon, but with the external instruction set based on the 68K and not the x86, and it would be big endien.
We all hate the x86 legacy, but it's time to face facts, the modern x86 is actually a great lump of silicon and very cheap to buy. Once real mode is switched off, the chip is great. The only thing bad is that motherboard manufactures still insist on using the IBM-PC BIOS!?!?!? I've been working on putting an AROS kernel in place of the BIOS so that a standard PC mobo would boot directly into intuition in 3 seconds (no more black and white text based BIOS screen), it's fun but I lack the technical knowledge about PC Chipsets to progress at any speed. But it really looks good to see!
Ok, my rant is over... you can get back to eating your lunch now
Remember that internally, a 600Mhz PPC and a 600Mhz Athlon don't look very different!!! Simply becuase to run a chip at such speeds, there is only one way to do it. I will grant you that the die size of the Athlon is bit bigger, but the Athlon is also a lot cheaper so that doesn't matter in the slightest!!!!
Oh, and to say that the modern x86 (eg Athlon) is unoptimised and inefficient is blatent FUD!!! I suggest you get yourself a good book on cpu design, and read it.
I already stated that the worst thing about the x86 is the archaic "IBM-PC BIOS" (which no OS other than MSDOS and Windoze9x uses anyway...), but that is flashable!!!! I've been looking at the OpenBIOS and LinuxBIOS projects as to how AROS can replace the BIOS. It's hard but not impossible, and great to see.
