Cool Mac Gear

Most MacWorld Expos leave us with a raft of new product announcements to digest, but Expo New York 2002 left us with a couple of modest hardware upgrades -- the 17 inch iMac and the tweaked iPod, but mostly controversy over the for-fee .mac replacement for free iTools, and continued anticipation of the August 24th debut of OS X 10.2 Jaguar -- itself not without controversy thanks to Apple’s decision to charge a full $129 fee for a .1 upgrade.

It Apple might have dodged one of these bullets by naming Jaguar OS X 10.5, which its promised features inventory would arguably have justified. However, it would also have been smart policy to make the Jaguar free upgrade being offered to those who purchase new Macs between July 17th and August 24th retroactive to, say, the PowerBooks/iBook upgrade announcements in late April/early May. That would have been a lot more classy, and not cost Apple a great deal, while paying dividends in consumer goodwill -- something that has been battered by the .mac bombshell.

I had my say on that topic last week, so I won’t belabor it again here. What I want to address in this column is the growing megahertz gap, in regard to which Apple has been barely more than treading water.

It’s just short of three years since the 500 MHz G4 was first announced, and two and one-half years since it actually shipped, and Apple is still stuck at 1 GHz with perhaps 1.2 GHz on the near term horizon. Meanwhile, Intel’s Pentium is within striking distance of 3 gigahertz, a threshold that will likely be crossed before the year’s out. And while megahertz, like horsepower in automobiles, is an overrated issue for many users once an adequate level is reached, a 3 to 1 or even 2.5 to 1 deficit is not something that can be rationalized away -- gigaflops and velocity engines notwithstanding.

Motorola, a troubled corporate entity these days, appears to have only perfunctory interest in PowerPC Development, and the recent layoff of another 4000 personnel in Moto’s semiconductor division is not an auspicious lookout with respect to a speedy debut of G5 chips. Motorola’s troubles put Apple in a precarious position, and it’s unlikely that Moto executives are losing any sleep over Apple’s increasing MHz gap woes. I expect some of them are still disgruntled over Steve Jobs pulling the plug on Motorola’s StarMax Mac clone license five years ago -- a move that cost Moto millions of dollars, and left them with an installed base of orphaned computers with five-year warranties to honor.

Why Apple has not availed itself more extensively of IBM’s somewhat faster progress in the PC development is a mystery. While IBM’s 600 MHz and 700 MHz 750FX “Sahara” chips are being used in the iBook and the holdover G3 CRT iMac, 800 MHz, 900 MHz, and 1 MHz versions are available, and IBM has 1.3 GHz Power4 CPUs, which are also PowerPC chips.

Presumably the it’s the hype that Apple has invested in the G4 and its Altivec (“Velocity Engine” that is the main impediment to more extensive use of available G3 horsepower, along with Apple’s policy of not selling “consumer” level computers with higher clock speeds than the G4 “professional” models. However, the problem is that PC vendors are selling cheapo bargain-basement entry-level machines with higher clock speeds than Apple’s professional machines.

Altivec’s virtues notwithstanding, has just never caught on in a big way. Only a handful of applications, notably Photoshop and a bunch of Apple’s own programs, really take advantage of it, and there is always the untidy detail of keeping performance acceptable on non-Altivec G3 machines, which Apple still sells.

Once IBM parted company with Motorola on the Altivec issue, the technology became a lame duck and an impediment to growth of the PowerPC platform, especially given Motorola’s growing indifference to the personal computer market for its chips.

My guess would be that if the PowerPC has a long-term future as a desktop and laptop CPU, IBM will be in the game longest, with Motorola a question mark.

These are of course more business and corporate political considerations than technological ones. There remains the issue of whether the PowerPC can keep the x86 in sight in the megahertz race. I’m a complete layman in these matters, and I have no idea what the Power PC’s real potential is -- whether the clock speed stall is due to inherent technical limitations, or whether it’s just that IBM and Motorola haven’t put the required development effort into keeping up with Intel and AMD in boosting clock speed.

Whatever, the bottom-line is that Apple is getting left behind, not least in terms of consumer perception, and it could be that the only ultimate solution will be to make a move to x86.

Personally, I would prefer a scenario in which IBM and Moto get it in gear and light fuse under PowerPC development. I don’t think that maintaining clock speed parity with the Pentium is necessary, but the ratio needs to be a lot less than 2.5 to 1 or 3 to 1. However, that may not be possible. And perhaps Steve Jobs was hinting at a possible shift to x86 in his cryptic, widely quoted comment in a July 17th interview on what might happen once Apple finishes the transition to OS X:”Then we’ll have options, and we like to have options,”

On the other hand, Jobs is also quoted saying: “But right now, you know, between Motorola and IBM, the roadmap looks pretty decent,” and perhaps the “options” remark was aimed more at sending a message to IBM and Motorola than signaling an imminent CPU platform shift for Apple.

Not that Apple couldn’t pull off a migration to x86, at least if past history is any guide. They already did it once back in 1994 with the switch from the Motorola 680x0 chip to PowerPC, and I think everyone, probably including Apple engineers, was amazed at how smoothly that went. 68 K emulation made the transition largely transparent from a software perspective, and the PC revolution was virtually painless for most users.

Of course, things are a bit more complicated right now, being as the Mac community is in the middle of another paradigm-shift, with the move from the Classic Mac OS to OS X, which already involves one layer of emulation in the form of Classic Mode support.

Rumor also has it that ATI and Nvidia have set up teams to investigate porting OS X to an x86 CPU, which implies that Apple has given these key suppliers a heads-up (and an OS X x86 prototype OS to work with) in anticipation of a potential move.

And as John Siracusa of Ars Technica recently observed:

What I don’t expect to see happen is a direct port of OS X 2 to x86 PC architecture. If Apple does go x86, look for some set of BIOS or ROM engineering that will continue to limit OS X to Apple hardware. The mid-’90s adventure in clone-licensing nearly killed Apple. Letting OS X run on any old cheapo PC box would be suicide for Apple’s hardware sales, which are the company’s most profitable cash cow. And with Microsoft contract stipulations stating that PC OEMs must not bundle alternative operating systems on penalty of losing their Windows license, going head-to-head with Windows in the OS market would be too risky.

Like I said, my preference would be for PowerPC development to advance to a degree that x86 Macs would not be necessary. Especially as a portable computer enthusiast, I regard the PowerPC, with its relatively low power consumption and more moderate heat generation issues as a much more attractive laptop CPU than the x86, provided that competitive performance levels can be maintained. However, I also think that Mr. Jobs is wise to keep exploring the “options.”

Charles W. Moore