PowerBook G4 Diary: The Lab Test

When I entered the Lab that fateful day last week to retrieve the PowerBook G4, I didn't realize we were culturing viruses as well as testing Macs. But once the proper gestation period had passed, the results were there for all to see: I lost my voice and my will to live. Worse, I lost the will to test, and fell a couple of days behind schedule for these diaries.

This is my lame way of apologizing to all of you for the long wait for more juicy details on Apple's new portable, or as Plastic.com recently described this diary, "almost pornographic in its step-by-step telling of the unpacking of the sleek beast." A technological pornographer; wouldn't my mom be proud?

But seriously folks, your response to this series has been both overwhelming and deeply gratifying. The interaction with you on the Macworld Forums has not only helped guide my testing, it's also helped remind me what a bright and enthusiastic group Macworld 's readers are. Thanks especially for the little digital "gifts" a couple of you sent, including this cool desktop picture from Aaron Cooper that he has graciously allowed me to post for anyone to download (click to download the full-size version):

In the last two diaries we unpacked, toured, and did a comparison of the Titanium PowerBook G4 to its predecessor. In this diary, we get down to brass tacks and answer most of your most burning questions on how it performs.

It's times like these I really come to appreciate how incredibly lucky I am to work at Macworld. Not just because I got to use a Titanium PowerBook before most everyone else, but because of Macworld Lab. Macworld has the Mac market's only fully-staffed, fully-equipped product-testing lab, with six full-time lab technicians who do nothing but test Mac hardware.

In short, these folks make me look good. They ran the PowerBook through a long series of laborious tests and I get to write up the results.

I've just received the final numbers from speed, heat and battery life tests we ran on the 500MHz PowerBook G4. After living with the Titanium PowerBook G4 for a week, the results are not surprising, although in a couple of areas they are a little disappointing.

To see where the G4 has its greatest impact, all you need to do is look at the Photoshop scores. Photoshop 5.5 has been optimized by Adobe to take advantage of the G4's AltiVec (Velocity Engine) processor extension, and in general, Photoshop has been designed to use faster processors, and memory and system buses to their fullest. In other words, Photoshop results are about as close as you get to a real-world test of the CPU performance.

Clearly, the G4 gives the new PowerBook a major edge over the PowerBook G3 when it comes to Altivec-optimized tasks such as Gaussian Blur, Unsharp Mask, or, most amazingly, Lighting Effects. Results range from twice as fast to more than three times faster on the Titanium PowerBook.

But once you exit the realm of AltiVec, the speed difference between old and new PowerBooks drops considerably: in a Photoshop RGB to CMYK conversion, the Titanium PowerBook G4 is only about 10 percent faster than the PowerBook G3.

And in Speedmark 2.1 tests, which judge performance on a variety of everyday computing tasks, the difference in speed is smaller still. It's important to keep in mind that when it comes to Speedmark, a little means a lot. The baseline score of 100 is for a 350MHz G3 iMac, so a score of 137 means the Titanium PowerBook G4 is roughly 37 percent faster than an iMac.

But notice how much higher the Power Mac G4 533 scored in Speedmark tests? That's because the Power Mac has faster system and memory buses, plus faster storage and graphics subsystems. The moral of the story: Everything contributes to real-world performance, not just the CPU. Because the rest of the PowerBook's logic board remains largely the same as the PowerBook G3, its overall performance gains are modest.

No place is this more obvious than in graphics performance. The Titanium PowerBook G4 uses the same ATI Rage Mobility graphics subsystem as the PowerBook G3. You can see that in our Quake III tests -- the scores are very close. This is because Quake can't leverage AltiVec, and without AltiVec, the G4 is only marginally faster than a G3.

However, graphics do not live by frame rate alone. Rendering is also an important factor for how well a graphics subsystem performs. And the best test of that is, again, a game. This time I used Deus Ex, a real-time 3-D first-person perspective game. I took screen shots of the same scenes on both the PowerBook G3 and G4, with the exact same game display settings -- 1,024-by-768-pixel resolution, 32-bit color, high graphics detail and decals all turned on.

Here are the two scenes shot on a PowerBook G3. (I chose scenes that show both distant and close-up texture maps -- click on the small versions to see a full-size screen.):

And here are the same two scenes on the PowerBook G4:

Can you see any difference?

I should add that while there was no appreciable difference in either game performance or visual quality between the G3 and G4 portables, I thought Deux Ex and Quake III were both quite playable on both machines, even with the graphics maxed out. Using the TrackPad as a game input device was more of an impediment to a good gaming experience than either the frame rate or rendering quality.

So, I guess the moral of the story is, if you want the best possible Mac gaming experience, you should opt for a desktop machine. Big surprise.

An area much more critical to most PowerBook users than game performance is startup speed. Again, I turned to Macworld Lab to run wake from sleep time trials on the new PowerBook. They tested both PowerBooks with OS 9.1; here's what they found:

Seems that wake time is one of the few areas where PowerBooks have a distinct advantage over desktops! I was surprised to see that the Lab found the G4 slower than the G3 on wake time with AppleTalk on, so I ran the tests again myself. I found that the G4 was indeed waking up in about 12 seconds. But my PowerBook G3 was taking over 40 seconds to wake from sleep with AppleTalk on! This can no doubt be attributed to the accumulation of excess garbage files and extensions that have collected like barnacles in my G3's System Folder. It will be interesting to watch and see if the G4's wake from sleep performance follows a similar downward spiral over time.

Of course, another key component of portable performance is battery life. Apple now claims battery run times in the five-hour range. I chose to test this claim in three different ways. First, in normal use, with the hard drive coming up and down as determined by a script, with processor cycling enabled and screen blanking after one minute of no input. Second, a maxed-out test featuring everything running, including the hard drive, DVD drive, with processor cycling disabled, and the screen at maximum brightness for the entire test. Third, absolute minimum power: screen dimmed, drive spun down, processor cycling enabled, and the chip clocked down to 300MHz (you can set this in the PowerBook control panel under Expert settings).

In the maxed-out test, the PowerBook managed to eke out an hour and 47 minutes of battery life -- not bad considering the abuse I was putting it through. The standard profile delivered even better results, 3 hours and 12 minutes, again very respectable. And, yes, the minimum power configuration did indeed deliver about 5 hours of battery life. Of course, with the screen dimmed, drive spun down, and running at 300MHz, I'm not sure what kind of life that would be for a real road warrior.

Finally, in response to dozens of e-mails from readers asking us to test the PowerBook to find out how hot it runs, we put both the PowerBook G3 and G4 on a heat analyzer and ran them for several hours under the most demanding circumstances possible. They both reached the same maximum heat level -- between 115 and 120 degrees Fahrenheit. My own experience with the PowerBook G4 confirms this: in what are now close to 100 hours of usage under a wide variety of circumstances, the G4 did get warm, but never to the point where it was uncomfortable for me to have it sitting in my lap.

As to where the reports of hot-running PowerBook G4s are coming from, I suspect that most of those reports came from pre-production models on the show floor at Macworld Expo -- where they had been running for hours, plugged in and sitting on light tables. Just about the worst-case scenario for heating up a metal portable.

My conclusions on our Lab testing won't surprise you. With the exception of AltiVec-savvy applications, the PowerBook G4 offers only modest speed gains (although this would still make the PowerBook the fastest portable you can buy on any platform). Graphics speed is a wash between generations, as is wake from sleep time and heat levels, but battery life seems to have improved somewhat. It's a good thing, too, as spare batteries are selling for $129 a pop at Apple's online store.

Hey! We're not done yet! Stay tuned for the next installment of the PowerBook G4 Diary, when Dr. Gore gets out his tool kit and cracks open the Titanium PowerBook G4 to see what's inside.

Want to discuss the Titanium PowerBook G4 lab results with Andrew Gore? You can do it now, in our "Titanium PowerBook Lab Tests" message thread, part of our Portable Macs Forum.

1 2 Page 1
Shop Tech Products at Amazon