Intel VP: Apple should have used our chips
Interesting article from MacCentral. Sour grapes? Hmm.
http://maccentral.macworld.com/news/2003/08/05/intel/
http://maccentral.macworld.com/news/2003/08/05/intel/
Comments
Muahahahahahahaha.
It makes perfect sense: Intel had meticulously paved the way for the Itanium to take over the workstation market by killing HP's chip and the Alpha; SPARC and MIPS (SGI's CPU of choice) were weakened, and IBM had left a huge gap between the aged 604ev and the monstrous POWER4, and AMD was struggling, and had been all but entirely locked out of Intel's turf anyway. They even had the luxury of time to delay its introduction, and reintroduce the chip after the first one bombed.
Now, all of a sudden, they've got fierce competition. The 970 outruns and hopelessly outscales the 32-bit Xeon, while running neck and neck with the big, expensive Itanium at a lower price and a much lower price/watt - crucial for the emerging clustering and blade markets.
So out comes the FUD machine. Expect plenty of it, because not only is the 970 a real threat, the Opteron also attacks the Itanic from a completely different angle, and in a way that's made the likes of Dell consider AMD seriously for the first time - AMD is now offering something that the OEMs can't get from Intel, and it's targeted at exactly the markets that Dell is now moving into (there's much more profit to be had there than in desktops).
He thinks his chips are the best and that everyone should be using them. He's totally entitled to that opinion, just as we're entitled to the opinion that the G5 has the potential to whip the ever-living crap out of their chips.
One thing is true, Intel has a plethera of chips including mobile chips that Apple could have immediately implemented in the Powerbooks.
Is there any real advantage with going with IBM or was it just to be different?
Originally posted by coolmac
I wonder, how and why did Steve Jobs decide to go with the IBM 970 processor anyway, and why did he choose it over Intel?
It's high-performing, more attractive in many ways than Intel's offerings, and 100% backward compatible with Apple's legacy going back to the mid-90s (and farther, thanks to the stalwart old 68k emulator).
Moving from the G4 to the 970 is essentially transparent and painless. Jumping to a different ISA would not be, and that ignores other considerations (like being at the mercy of Intel, and thus Microsoft, trading AltiVec for the woeful SSE, adopting a dead-end x86 technology now that Intel is moving to Itanium, having to support two fundamentally different and utterly incompatible 32 and 64 bit platforms rather than one seamless architecture).
One thing is true, Intel has a plethera of chips including mobile chips that Apple could have immediately implemented in the Powerbooks.
Is there any real advantage with going with IBM or was it just to be different?
Nothing Intel offers could be immediately implemented in the PowerBooks. The ISA is different and the bus is different, so Apple would have to overhaul the motherboard; port over any parts of the OS that can't be recompiled and recompile the rest (losing all the benefits of AltiVec optimization that SSE can't come close to replacing) - and then ship a substantially larger distribution with both x86 and PowerPC binaries; drop Classic support; and then ship a machine that requires the buyers to purchase or download ported and/or recompiled versions of all their applications. Meanwhile, it would be much easier to move WINE over to OS X, which would mean that instead of developing native applications, companies could just tell Mac users to run the Windows version (in which case, why not just get the real thing?). Since x86 has fewer implementation registers than PowerPC, x86 would choke hard on any attempt at emulating a PowerPC, so even if Apple did ship an emulator it would suck mightily. Even the 68k emulator would suck wind, since that architecture had a generous number of registers as well.
Or, Apple can just use the latest and greatest iteration of the architecture they've been using since 1994, reap the rewards of being a significant enough customer to have a hand in the design of the CPUs they use (a luxury they would not have as an Intel customer), remain both independent of WinTel and allied with their most dangerous rival (IBM), spare themselves a tremendous amount of hassle, and release machines that can go head to head with workstations 3-5 times their price and win.
It's a dead simple decision, really.
In conclusion he's a big liar.