Potholes on Apple's chip roadmap?

Three wheels on their wagon. But they keep rolling along.

Macintosh site Architosh has some reasonably intelligent speculation on unreleased Motorola processors, and offers a roadmap.

The site has a good track record, and based its article from information in a Motorola document that has subsequently been withdrawn. The unsigned piece considers revisions to the 745x processor that was originally launched in January this year. Its author(s) suggest a tentative timescale which sees faster revisions of the current 745x leading to the introduction of 970-based systems in 2004.

But what happened to Motorola's 7470? This is one of the most frequently asked questions we get from professional Mac users.

We corroborated that the processor was ready back in May. It should be in volume production now, and its non appearance gives credence to the suggestion that Motorola and Apple no longer have the close working relationship they previously enjoyed. If they have any kind of relationship at all: the changes to the older 745x processor are all modifications that can be made by Apple's engineering team. Motorola appears to be contractually obliged to provide the older processor, and it's up to Apple to bend the space-time continuum to make them faster.

This theory caused a stir recently as it included the accusation that Apple is using "overclocked" processors in its most recent "Windtunnel" G4 Macs. Overclocking is when the user raises the clock frequency beyond the recommended frequency marked on the processor. Chips are capable of operating at several speeds, and are graded as they leave the fab: good batches are judged capable of running higher frequencies. Since the parts Apple is using in the latest Macs run at their marked frequency, Apple can't strictly be accused of "overclocking".

But it's a distinction without a difference. These are last years chips, running hotter and at higher frequencies than they were last year.

The real story, hinted at in Architosh's article is the non appearance of new Motorola processors, and what that tells us about the Apple-Motorola relationship.

Apple badly needs a speed infusion for its professional range: as an analyst at Microprocessor Forum pointed out, it now has a workstation class OS, so it needs a workstation class processor. Not for you and me perhaps, who get along just fine, but for the big content creators that represent a potentially lucrative new market for Apple.

Bootnote: Architosh makes one error in suggesting that 970-based Apple notebooks would be the first 64bit notebooks. Actually, you've been able to buy a 64bit notebook for some years, from these nice people. IBM has produced hybrid 64/32 designs before, and used the POWER as the basis of moving its AS/400 from 48bit CISC to 64b it RISC. Hitachi has produced PPC-based derivatives, but it's actually one of the last to announce.

Our 64bit timescale looks like this - and please mail us corrections. The dates are date of the announcement, not ship dates:- October 1991: MIPS (R4000); March 1992: DEC (Alpha); October 1994: Sun (UltraSPARC); November 1994: Hewlett Packard (PA-RISC 8000); October 1997: Intel and Hewlett Packard (IA-64); October 1999: AMD (x86-64). (It's a bit hard to tell when IA-64 was "announced"; the original declaration of intent between Intel and HP was made in June 1994. We're plumping for 97. Paging Mr Huck! … ) ®