Why MacIntel Is Good News For PowerBook And iBook Fans
by Charles W. Moore
Steve Jobs' WWDC bombshell announcement that Apple will be switching from Power PC to Intel processors over the next couple of years pretty much puts paid to the prospect of a G5 PowerBook -- now almost certainly to remain forever vapor. It's hard to conceive of Apple expanding any more development resources or Energy on the G5, PowerBook project, even if IBM were willing to play ball, which they likely aren't. Apple will be focused on getting Intel-based books ready for 2006 release.
Which begs the question of what will happen in the meantime. Apple laptops are already lagging way behind X86 Intel PC notebooks in the megahertz race, albeit not necessarily in real-world, practical performance. However, the puny speed-bump from 1.5 GHz to 1.67 GHz at the top end of the last PowerBook revisions indicates that G4 speed advances have slowed to a near-standstill, and the year that will elapse before the debut of the first Intel-based Macs could be a very long one indeed if some sort of portable performance boost can't be engineered in the meantime.
Whatever transpires, I am pretty confident that the PowerBook will remain G4-based until new Intel versions are ready. It's hard to imagine that Apple will try to soldier it out for another 12 months with the 7447 series G4 chips, which are partially crippled by slow bus speed which prevents the full speed of fast DDR memory from being fully utilized with the processor, and which may or may not be have the potential to be squeezed to 1.8 GHz or so. One possibility, although it may be a long shot, is that the PowerBook will get Freescale's G4 7448 processor, which has reportedly been sampling at clock speeds of 1.4-2.0 GHz and a bus speed of 200 MHz, along with 1 MB of L2 cache. Since the 7448 is also pin-for-pin and software compatible with the current MPC74xx processors, it could easily be plugged into Apple's existing 'Book designs as a stopgap.
Otherwise, the current MHz logjam will continue and probably become even more rest. Currently, only 170 MHz separates the entry-level PowerBook from the top of the line, and the lowliest iBook (which is overdue for a speed bump) is just 470 MHz lower in clock speed from the fastest 15" and 17" PowerBooks. The PowerBook high-end has advanced a meager 670 MHz in the past two and one half years plus.
Which highlights why Apple needed to do something drastic about CPU supply. Portables represent more than 50 percent of Apple's system sales, and are too important to let languish. Intel has been putting a lot of engineering effort into low-speed chip development for the portable computer market, and whatever was going on at IBM in that context, it obviously wasn't enough.
I've long been skeptical about the G5 as a laptop chip candidate, and am on record saying as much in this column. The problems that the Power PC 970 has presented in terms of heat management even in desktop machines has been formidable. Witness the elaborate, multi-fan cooling systems in the G5 Power Macs and iMacs. Apple's VP of Worldwide Sales and Operations, Tim Cook commented last winter, getting a G5 to work in a PowerBook was the "the mother of all thermal challenges." It just didn't seem likely that IBM would be able to produce in the foreseeable future a G5 chip with modest enough power demand and heat generation qualities to operate in a portable without unacceptable compromises.
As for an Intel-based PowerBook (how odd that still sounds!), there are a bunch of existing mobile Pentium/Centrino chips on the Intel shelf:
• Itanium 2
• Xeon
• Pentium Extreme Ed.
• Pentium 4 HT
• Pentium 4
• Mobile Pentium 4 HT
• Pentium M
• Celeron D
• Celeron M
The Pentium M (Centrino) is faster than Apple's top-end Freescale PowerBook chips, both in clock speed and system bus throughput, and include advanced technologies like PCI Express and Serial ATA, DDR 2 SDRAM support, and HD audio.
But even more interesting, and fitting nicely into the time frame for Intel-based Mac laptops, is a new line of mobile computing CPUs Intel has under development called "Yonah," which reportedly will be fabricated using a 65-nanometer process, and incorporate a number of enhancements over Intel's current Pentium M line of notebook chips. Yonah's dual cores will share a 2MB cache, while a single-core version of the chip will be available for low-priced laptops. The technology used in Yonah will focus on reduced power consumption and heat dissipation, with the objective to reduce power consumption to the point by 2008 for Yonah-based laptops to be able to run for eight hours on a battery charge. Sounds very promising.
On another, related note, I was disappointed that there were no iBook upgrade announcements either in the month of May or in Steve Jobs' WWDC keynote. The iBook is now seriously overdue for a speed bump and video upgrade so that it can support Core Image in OS X 10.4 Tiger without qualification.
However, a single clock speed increment upgrade would put the top end iBook at the same level as the 12" and base 15" PowerBooks, which Apple probably finds unappealing from a marketing standpoint.
What we may see is more emphasis placed on other sorts of enhancements for both the iBook and PowerBook, such as the widescreen iBooks rumored for release in the fall. We got a taste of this with the January PowerBook revisions, which played up the new motion sensor and scrollable trackpad features, and other, non-clock speed specification enhancements, rather than the modest speed bumps.
Getting back to Intel and Apple, the sketchy road map that has been laid out thus far has Apple's consumer machines -- the Mac mini, the iMac, and the iBook projected as the first models slated to receive the Intel makeover. That presumably means that there will be Intel-based iBooks before we will see PowerBooks with Intel inside, although I'll be very surprised if the PowerBook doesn't get an Intel injection fairly early on as well.
Make no mistake, the need for more powerful laptop processors was instrumental to Apple's decision to make this paradigm-shifting switch.
In a long run, this move should prove to be beneficial for we Apple portable aficionados. It was ever more evident that neither IBM nor Freescale had a great deal of interest in pursuing the laptop chip market aggressively. In some slightly sour grapes post-keynote reaction, Freescale CEO Michel Mayer was quoted saying in an email to employees that he was disappointed, but noted that the company's Mac business accounts for less than three per cent of total revenue, while IBM said in a statement that the company is focused on the highest-value opportunities in each marketplace. Intel, on the other hand, is strongly focused on laptop chip engineering.
It is sad, of course, to contemplate the end of the Power PC era. Originally, it seemed as if the Power PC would prove a more elegant and efficient solution with its Reduced Instruction Set Computing (RISC) technology as opposed to the Complex Instruction Set Computing (CISC) architecture used in Intel's X86 chips as well as the Motorola 680X0 processors used in all Macs up to 1994, and finally phased out in 1996. However, Intel gradually incorporated some of the RISC advantages into its X86 chips, making the distinction much less compelling than it had been.
The G3 and G4 Power PC processors did prove to be superb portable computer CPUs, running at relatively low wattage is that allowed industry-leading battery life and relatively modest heat generation. If the engineering effort had been applied to keeping clock speed competitive with PC laptops, and we now had 2 GHz plus PowerBooks, the Power PC party might have continued. But that didn't happen, so thanks for the memories, and it's time to move on.
It's going to be an interesting ride, and doubtless a bit bumpy in spots. Hold on to your hats!
