Perspective: Processor Design vs. Development

Posted by:
Date: Sunday, August 18th, 2002, 19:44
Category: Archive

PowerPage’s Bob Snow thinks jumping the PowerPC ship for Intel’s inferior designs isn’t the answer: expanding Apple market share and fixing the PowerPC’s R&D and production woes is possible and worthwhile. Read on, and weigh in.

There has been a lot of talk lately, suggesting that Apple might move to Intel as a supplier for it’s processors. I for one would hate to see such a move, but this is based on design aesthetics, not horsepower. Intel derives quite a few advantages from its volume of production and dominance of the market. It also carries some serious baggage. The Pentium has an architecture that is a throwback to the early days of the PC, but through impressive process engineering it has been brought forward to performance levels that are impressive. It is however a complex and power hungry beast that uses deep pipelines to reach high clockspeeds, often at the expense of real world performance. The new Itanium chip also seems to be a problematic product and Intel is currently fending off the bad publicity resulting from a nuisance lawsuit which alleges that they over hyped the Pentium 4 — duh!

I think the PowerPC has a much more elegant architecture, but is hamstrung by the fact that it lacks development resources due to the small production volumes. This hurts the R&D budget and the unfortunate split between Motorola and IBM over Altivec has put production and development of the G4 in the hands of the weakest partner. I think the PowerPC has far more development potential than the Pentium, it’s simply a question of putting the resources in place to do the job. Apple can help by boosting market share. Motorola relies on the embedded market to spread the cost of development, often at the expense of the the resources needed so desperately for the versions Apple wants. If market share for the Mac eventually moved from 5% to 10%, this would make plenty of resources available and the problem would get solved. One option is to move to IBM for production and development including the possibility of using the Power4 for desktop Macs. Apple might even consider buying out Motorola?s chip making division or entice a company like AMD to put their resources into moving the PowerPC chip forward.

Early dual processor G4?s were stop-gaps. Running OS9, they were mostly a marketing gambit, since they initially provided a modest boost, but only for programs that were optimized for dual processors. OS-X has changed all that, especially Jaguar. Now, the biggest bottleneck for new Apple hardware has to do with Motorola?s inability to get the latest chips into production to take full advantage of DDR memory.

Recent Posts

Comments are closed.