Apple's Transition to PowerPC put in perspective
Contributed by
Thomas Hormby | osViews
January 3, 2005 3:57 PM
The short story of RISC on the desktop.
In the early nineties and late eighties, Apple was more successful than it had ever been. It's line of computers were widely respected for their ease of use and speed. The recently introduced Macintosh IIfx outperformed PC's and Macs alike. The Motorola 68k, the processor used in the Macintosh, was not performing as Apple had hoped. Designed for machines with limited resources, the 68k was very complex, but required very little code. The complexity helped companies deliver machines when RAM and storage were both very expensive, but when prices dropped, the complexity made for slower programs. Apple decided that a new architecture would be required to run the next generation operating systems and applications.
The 68k was introduced to be used in household applications, but it was quickly adopted by companies peddling UNIX workstations because of its similarities to the PDP-11 and VAX and for its optimizations for the C programming language. Early Sun, SGI, Apollo, HP and NeXT machines all used the new 68k series.
Jef Raskin, the original manager of the Macintosh project, had envisioned the Macintosh with a very inexpensive processor, the Motorola 6809E. Bud Tribble, a software designer at Apple had the project switch over to the more powerful 68k so that they could use LISA code. The change stuck, and the Macintosh was designed around a workstation-class processor.
Several years after the introduction of the Macintosh, Apple began the Jaguar project, to create a RISC based workstation around the 88110 processor from Motorola. The project manager, Jack McHenry, feared that the machine would flop if it was not compatible with the existing software for the Macintosh. He had the team create a emulator for the 68k processor, codenamed Cognac.
Unlike emulators before, like VirtualPC and RealPC, the 68k emulator performed very well. The emulator operated on the 10/90 principle, only 10% of the code is being used 90% of the time. Cognac never emulated an entire Macintosh for the legacy program all the time, only the elements that were used most often. Everything else was emulated only when it was needed. The Jaguar project was canceled in June of 1990 when Apple executives feared that it would kill Macintosh sales.
Sales of Motorola's new RISC processor, the 88110, were not doing well. Motorola's biggest buyers for the chip, Ford and Apple, both canceled their contracts and were looking for more viable alternatives. In response, the Apple-IBM-Motorola, or AIM, alliance was formed to create a brand new processor, architecture and operating system.
The new processor was to be the PowerPC, an IBM POWER processor, with modifications form Motorola. The new architecture was the be PReP, a slightly modified RS/6000 and the new operating system was to be based on Pink, and produced by a company funded by IBM and Apple, Taligent. Only the PowerPC was successful from the alliance, the other products were late to the market and never adopted.
The PowerPC 601, the first chip from the AIM alliance, was delivered to Apple on September 2, 1992 in Christmas wrapping paper. Work began immediately on porting the Mac OS to the new chip. Using the Cognac emulator, the first PowerPC prototype, the Pilt Down Mac, booted the Mac OS one week after it was delivered. The PDM became the low end PowerPC machine at Apple, the PowerMacintosh 6100/66. Apple still needed a midrange and high end design, and it turned to Jaguar technology, and used a slightly modified Jaguar motherboard for Cold Fusion and Carl Sagan, the 7100/66 and 8100/80.
The new machines shipped on time, during the Mac's tenth anniversary. The public and business alike loved the machines, and quickly adopted them, with the last 68k Macintosh, the PowerBook 190, being discontinued in 1996, two years after the PowerPC was introduced.
