Titanium PowerBook Report Part 5

Part V: Using the Titanium PowerBook

Once I'd gotten my TiBook personalized and moved my important applications and files over from my Wallstreet, it was time to examine the Ti's speed.

While some hardcore types like scientifically precise benchmarks, the truth is that all speed tests are subjective. Just because one computer runs a certain Photoshop filter 0.68 times faster than another doesn't mean you'll notice the extra speed (unless you happen to run that particular filter on 500 MB files all day long every day). For most people, real world tests are far more significant, and that's why I prefer informal benchmarks (though I'm including the results of a few standardized tests below).

In this regard, the Titanium shines. In fact, it feels as fast or faster than my 500-mhz desktop G4, at least in terms of booting and launching applications. For regular software tasks, I can actually notice a slight improvement over the 400-mhz G4 I use daily at work. The fact that I can notice a difference is impressive -- often times there's a technical speed difference but it's negligible in real life, or only noticeable during times of extreme calculation (compressing a QuickTime movie, running Photoshop filters, ripping MP3s, etc.).

Since this machine's so fast, I experimented by turning on power-saving features I'd never used on my G3: I set processor cycling and reduced processor speed (which slows my 500-mhz processor to 400-mhz) to be on when using the battery.

Surprisingly, I didn't notice much of a difference. I even played several minutes of a DVD and while the software warned me on launch that I was running at reduced speed, I told it to go ahead and I didn't notice any skipping or navigation problems (but I'll confess I didn't watch an entire movie). I did have virtual memory turned off, which is supposed to help DVD playback. At one point, doing some word processing while having SoundJam play MP3s in the background, I found typing a little sluggish. I went to SoundJam's preferences and checked the "Reduce processor use when in the background" and the problem went away (I didn't notice any different either way with SoundJam's playback).

Another interesting test: Virtual PC 4 (which is accelerated for the G4 processor). I found Windows 95 to be splendidly fast. Windows popped open instantly, instead of the stuttering motion I used to see. Solitaire actually kept up with me!

The slowest aspect of a portable system is the hard drive. To see how fast Apple's 20 gigabyte drive is I ran a disk utility on an empty partition and compared the results with an IBM 45-gig 7200 RPM installed in a desktop G4.

Hard Drive Speed PowerBook 20GB as Percentage of IBM 45GB (in desktop G4) (Higher percentage = faster)
Test Size	Read Sequential	Write Sequential	Read Random	Write Random
16 KBytes	54.90%	98.94%	43.12%	23.01%
64 KBytes	39.06%	108.10%	45.62%	20.45%
256 KBytes	42.18%	86.00%	46.23%	47.96%
1 MByte	33.75%	60.27%	39.00%	70.55%

(Test was performed with QuickBench™ 1.5.)

As you can see from the above table, the drive is on average 50% slower than the drive in the desktop Mac -- but a great deal depends on what you are doing with the drive. Writing many small chunks of data on a fragmented disk (Write Random) is much slower than on the desktop system (as slow as 20% of the desktop model's speed). But in a few cases the laptop drive performs as well or better than the desktop drive: writing sequential data is quite respectable.

For my CPU tests I turned to Photoshop, but my tests are not scientific. I tested Photoshop the way I tend to use it: the screen set to millions of colors and to whatever resolution I usually use. So, for instance, the screen resolution of my desktop Mac and Titanium don't match, which probably skews the results, but I don't care. I wanted to see how the Ti would fare in a real world, unfair test.

Photoshop Results 500-mhz PowerBook G4 versus 400-mhz PowerMac G4 (All times in seconds -- smaller is faster)
Action	Desktop G4 500	PBG4 500 (64MB)	PBG4 500 (220MB)
Rotate 90	1.1	29.4	1.2
Rotate 9	12.9	49.3	10.6
Rotate .9	11.8	46.9	21.5
Gaussian Blur 1	2.5	33.0	3.2
Gaussian Blur 3.8	7.2	36.9	7.2
Gaussian Blur 85	7.6	32.5	7.6
Unsharp Mask 1/50/0	4.5	34.1	4.2
Unsharp Mask 3.8/50/0	8.2	39.4	8.3
Unsharp Mask 10/50/5	8.1	39.6	8.1
Despeckle	3.3	33.6	3.9
CMYK Convert	21.4	34.0	23.1
Image Size	1.9	26.4	2.0
Lens Flare	15.1	67.2	15.2
Color Halftone	15.6	73.2	15.8
NTSC Colors	16.1	78.1	17.4
Accented Edges	42.9	62.9	44.1
Pointillize	62.7	115.2	62.1
Watercolor	90.9	127.0	93.1
Polar Coordinates	14.4	62.2	14.5
Radial Blur	195.2	243.1	191.0
Lighting Effects	8.2	67.6	8.4
Total:	551.6	1331.6	562.5

(The test used was PS5bench Advanced which manipulates a 50MB test image. Red cells indicate fastest test.)

As you can see, the Titanium fares well against the same speed desktop G4. It's less than two percent slower! The desktop's faster speed on some tests could be due to occasional hard drive access.

To see how much of a difference the slower hard drive made, I ran the tests on the Titanium with only 64MB of RAM allocated to Photoshop. The results proved that you need to be sure to put plenty of RAM in your Ti, because the slow disk performance will kill you. A 400-mhz desktop G4 did the same 64MB test in 1060.1 seconds -- 25% faster than the 500-mhz PowerBook! (Of course the Ti beat the 400 when both had plenty of RAM.)

While the Ti has a faster video card than my old Wallstreet, it is certainly no match for even an ordinary card like the ATI Rage 128 in the desktop PowerMac 500. I don't do a lot of gaming, so 3D performance isn't of much interest to me (though my Titanium ran Oni at what seemed to me blinding speed), but I do do a great deal of scrolling. Unfortunately, the Ti can't compare to the desktop Mac, even when the Ti is set to just 256 colors.

Scroll Test Time to scroll through 3500-line text document (All times in seconds -- smaller is faster)
Test	Desktop G4 500 (millions)	PBG4 (256)	PBG4 (thousands)	PBG4 (millions)	Wallstreet 250-mhz G3 (millions)	Wallstreet 250-mhz G3 (256)
Line	15.47	19.7	21.92	31.07	122.10	29.36
Page	6.49	7.49	7.50	7.51	9.49	9.23

As a final speed test between the desktop and the portable, I decided to RIP a CD using SoundJam and see how the machines compared. (SoundJam is accelerated for the G4 chip.) With SoundJam configured to immediately convert any CD inserted, I timed from inserting the CD until SoundJam reported being finished. (The CD I used was Sara Groves' latest, Conversations.)

CD to MP3 Test (All times in minutes:seconds -- smaller is faster)
Test	PowerMac G4 500-mhz	PowerBook G4 500-mhz
Convert CD to MP3	7:04	12:56

I suspect a big part of the difference in rip speed has to do with the speed of the DVD drive. The laptop drive is no doubt slower than the desktop model. I got 3.5x conversions on the Titanium, but nearly 8x on the desktop.

Next Segment: In Part VI, Marc makes some final observations and concludes his Titanium PowerBook G4 report.

Titanium Report - Part 1
Marc receives his Titanium PowerBook and reports his initial impressions.

Titanium Report - Part 2
Analysis of the exterior of the Titanium PowerBook continues with an exploration of the screen and keyboard.

Titanium Report - Part 3
Marc powers up the TiBook and discovers the first problems.

Titanium Report - Part 4
Marc installs RAM, partitions the hard drive, and begins personalizing his Titanium wonder.

Titanium Report - Part 5
Performance results: Marc benchmarks the Titanium PowerBook against a 500-mhz desktop G4 in a variety of tests.

Titanium Report - Part 6
Should you buy a Titanium PowerBook now or wait for version 1.1? In this final segment, Marc comments on a variety of issues and concludes with his analysis of the Ti's worthiness.

.