The state of Intel’s processor architectures was of only passing interest at Macworld – until June when Apple CEO Steve Jobs dropped the bombshell that, starting next year, Macs will feature Intel chips and not those of the stalled PowerPC range.

Steve probably feared that IBM would retreat to the pre-transistor days of digital circuits composed of vacuum tubes, light bulbs and bike pedals.

Intel has just held its autumn Developer Forum. As usual it talked up the processor roadmap, and then vainly lectured the PC makers about how they should build their computers to take advantage of its latest technologies. Everyone listens politely, claps at the right moments, and then goes home to design even cheaper PCs that are sometimes less sophisticated than the ones they used to make before the Forum.

Except that this year, sitting in the front row were some new faces. Probably dressed in the Apple uniform of jeans and black polo necks, these guys design the best-looking, most sophisticated and innovative personal computers in the business. Wary of cost but determined to build the best, these smarty-pants may have been the only developers present who actually listened and plan to utilize these latest techs.
At last Intel has a partner that will work alongside it to create computers worthy of the processors within – computers that move the industry and all of us in a forward direction.

After years of toe-tapping static progress in chip-speed advancements via IBM and Motorola, Apple’s engineers heard Intel boldly announce that as of next year its processors would run a lot slower than they do right now. Slower… Did they sit there open mouthed and distraught? No, the Apple guys just smiled.

The basic thrust of Intel’s processor announcements is that naked speed is out and cool is in. In explaining the switch to Intel, Jobs compared the PowerPC with Intel’s chips for “performance per watt”, not gigahertz and teraflops. No matter how many transistors it’s got, even the speediest chip isn’t much good if half that power escapes as heat.

The Pentium 4 is based on an architecture designed to deliver maximum performance. Clock speed is the most easily understood aspect of processor performance, even though it’s just one of many metrics used to increase chip performance.

Increases in clock speed – the rate at which a processor executes instructions – require increases in the electrical power used to run the chip. Keeping this power under control wasn’t a problem for Intel until the advent of new manufacturing technologies in 2003. As always, Intel used these new technologies to shrink the size of its transistors, but as lines within chips decrease in size to average widths of 90 nanometres, more electrons are able to penetrate the thin walls and leak out of the processor as heat. Excessive amounts of heat can damage a system, and as chips get hotter more expensive cooling equipment is required to maintain performance. And Intel’s next set of chips uses an even smaller 65-nanometre process.

The power density of processors used in PCs today is on a par with that inside a nuclear reactor. And the failure rate of a chip doubles with every increase in temperature of 10°C.

The Pentium M architecture is designed to do more work per clock cycle, so it doesn’t have to run as fast or use as much power to produce results similar to the Pentium 4. This also makes it easier for Intel to create chips with four or more processing cores that will fit into PCs and servers without the need for expensive and bulky cooling equipment.

In order to achieve the speeds that eclipsed those of the G4 and G5, Intel lengthened the pipeline of the Pentium 4 processor to 31 stages, almost three times the length in its Pentium III. Think of a processor’s pipeline as a series of steps needed to process data. Data is moved into the first stage of a pipeline, processed, and then sent to the next stage. At the same time, new data is brought into the first stage – a process that is repeated on down the pipeline.

Processors, like the Pentium 4, with longer pipelines that only work on a small piece of data in each stage have to run very fast in order to get the same amount of work done as a chip with a shorter pipeline – such as the PowerPC.

A few years ago Apple tried desperately to tell everyone exactly this when exposing what it called the “Megahertz Myth”.

Intel’s new architecture dials down the power and the clock speed by using only 14 stages to process data. This means that the new chips will do more work per stage, and can therefore run at slower clock speeds.

Another improvement in the new design is the ability of its memory caches to share data. Cache memory stores frequently used instructions close to the processor, where they can be accessed far more quickly than if they are stored in memory. With Intel’s first dual-core chips, if one processor core needs information stored in the cache located adjacent to the other core, it has to laboriously request the data by going off the chip through its front-side bus connection to memory. Shared caches exchange data without having to leave the chip – significantly improving performance.

Intel also disclosed a number of power-saving technologies built into its forthcoming Napa mobile-computing platform. Laptops based on Napa will offer longer battery life with a significant increase in performance over older Pentium M notebooks. The chipset will be able to automatically dim a laptop’s display based on the amount of power left in the battery in order to extend the lifetime of that charge. And it can avoid processing graphical data that does not affect the quality of the final image.
And Intel isn’t just saving juice in its chips.

It’s been working with Panasonic to improve battery technology. The work is focused on improving current lithium-ion battery technology to the point where a PowerBook or iBook will be able to run for eight hours on a single charge – at least a third longer than currently possible.

The new long-life battery should be available in April 2006 – just in time for Apple’s first Intel-based laptops. Combined with the power-efficient processor, such a power-portable should quickly dispel any lingering doubts we had over Apple’s chip switch. If you thought Apple’s move to Intel processors was hot news you were wrong. Very soon, the Mac is going to be super-cool. MW