Apple to dump PPC for future Macs?

[JetRacer]

What I noticed was that bullcrap PPC vs x86 "benchmark".

They prolly used dual 333MHz FSB x86 mobos (the pricetags sure suggests so) pitted against single 166MHz FSB / 333MHz DDR PPC mobo. And the software was probably also non-altivec (G4). Or in other words: foul x86 propaganda at it's best...

It's the same old trick that Intel used (still use?) against Motorola; release first and do the benchmarks against the competitors old CPU line. Here it's Apples last years h/w against the brand new x86.

So don't worry about the 600MHz A1SE... It won't run at 75% performance of an x86 at equal MHz. Just keep those clusters coming :-)

[JetRacer]

@Hammer: Yeah, you're right. The x86 figure is wrong in the case of the Dell. I was speaking in more general terms and compared with AMD.

The Articia chipset which Apple uses have›three versions (S, Sa, P), All 133MHz, All 166MHz and a combo 166MHz FSB 333MHz DDR (some extra b/w for the gfx board).

The dual PPC are able to work directly agaist a single 133/166MHz FSB (since beginning of time). And is generally compared with newer x86 machines with far greater FSB bandwidth or even dual FSB. Point is that the only thing that matters in large, simple data processing jobs is FSB bandwidth (cache and altivec is useless). Do the same tests with f.ex. lightsource rendering using three clustered 600MHz PPC (preferably, but not necessarely G4) and the result will be devastating for the single 2GHz x86.

To someone else: x86 is crap. The quality of the PPC allows the use of Asm optimizing. And we all know what the difference between Asm and C in critical code. Try writing a 3D demo for the classic in C using maximum priority and you'll get the point. It's dead slow no matter priority. It's the assmebler that makes the difference, not the lack of multitasking. It's no sensation if someone manage to squeeze out twice as much performance due to asm optimizing. Nor x4 for that matter.