Hands on the FireWire iBook
October 9, 2000 - Dan Knight
A year ago I wrote about the first (and
still only) iBook we got at work. Today I'm writing about my
wife's new indigo iBook.
There's a sad story behind 
it,
which involves her PowerBook G3
Series II (also known as WallStreet) and some Diet Coke. The
PowerBook lost.
Thank goodness we run backup religiously on our home network.
Although she lost some work, we were able to restore everything as of
the last backup. In fact, I put those files on her partition on my
computer so she could work until the iBook arrived.
The first step was partitioning the hard drive. I wanted a
separate partition for Virtual PC, one for a spare copy of the OS,
and one for browser cache files - all in addition to the main
working partition. We ended up with 7 GB for work, 512 MB for Virtual
PC, 2 GB for emergencies (and OS X Beta!), and 32 MB for cache
files.
The next step was using the Software Restore CD to put back
everything wiped away when we partitioned the hard drive. Next boot
from the emergency partition, connect ethernet, turn on sharing, and
copy all the files from her partition on my hard drive to the main
partition on the new iBook.
Benchmarks
MacBench 5
That done, the next step was to run MacBench 5.0 and see how the
iBook's new PowerPC 750CX processor performs. (Yeah, I'm something of
a geek at heart.) The first test was with Virtual Memory on and set
to 65 MB, the iBook's default. The second test was with Virtual
Memory turned off. To test a friend's theory that VM works best when
set to multiples of 32, I also tested at 96 MB. (Since I can't locate
my MacBench CD at the moment, I was unable to test graphics or CD-ROM
performance.)
test CPU FPU disk
VM 65 693 1221 1121
VM off 693 1222 1146
VM 96 696 1220 1113
There's no significant performance difference between these tests.
Even the disk tests show no more than 3% difference from best to
worst.
It's quite surprising to see such a low CPU test score, since
MacBench 5.0 sets the beige Power Mac
G3/300 at 1000. In fact, the original 233 MHz iMac scored 696!
Why a 366 MHz PowerPC 750CX, the processor used in the new iBooks,
should score so poorly under MacBench 5 was beyond me. For
comparison, the original iBook
scored 893 on the CPU test and 973 on the FPU test. Hoping it was
some extension or control panel causing the slowdown, I booted with
extensions off and obtained virtually identical CPU (697) and FPU
(1234) scores.
In the final analysis, I must say I was disappointed to find the
366 MHz 750CX chip scoring worse than the 300 MHz 750 processor in
the original iBook. All the PR from IBM pointed to significant
efficiencies in the 750CX that should let it outperform a G3 at
the same clock speed (with a 512 KB 02.1 cache). Instead, a chip that
clocks 22% faster actually benchmarks 22% slower.
On the other hand, the new iBook scores 26% better on the FPU
test, or as well as a G3/400 would if Apple had put one in an
iBook.
Speedometer 4.0.2
I couldn't leave it at that, so I pulled out good old Speedometer
4.02, which is so old it uses the Quadra
605 as it's base score of 1.0. Time to see how the indigo iBook
scores.
I ran Speedo 4 on my SuperMac
S900 with 333 MHz Newer Tech G3 card for comparison. After seeing
the MacBench 5 scores, I thought my S900 would blow past the new
iBook.
test CPU disk math
iBook 28.2 3.24 1016
S900 25.2 2.95 891
Here the 366 MHz PowerPC 750CX in the new iBook outperforms the G3
by 12.2% on the CPU test and 14.1% on the math test. Considering the
iBook runs at a 10% higher clock speed, that's impressive.
According to Newer Tech's Clockometer utility, the G3 upgrade
actually runs at 330 MHz with a 220 MHz 1 MB backside cache on a 44
MHz system bus. That accounts for another 1% of the iBook's
performance.
When IBM introduced the PowerPC 750CX processor, they claimed that
the 256 KB on-chip cache would provide similar performance to the old
G3 with a 512 KB backside cache running at half processor speed. (For
a lot more on the 750CX, see Should
Apple Use the New G3?) Although I don't have Speedometer scores
for the recently discontinued 366 MHz iBook
SE, I suspect they would be very similar to the indigo iBook's
scores. If anyone out there has run Speedometer 4 on the 366 MHz
iBook SE, I would be very interested in receiving your results.
On Benchmarks
What's curious is the discrepancy between MacBench 5 and
Speedometer 4. As we note on our
benchmark index, "Remember that benchmarks are arbitrary. They
measure certain types of performance that may or may not reflect the
way you work."
Well, MacBench 5 and Speedometer 4 apparently measure CPU and
math/FPU performance very differently.
Perhaps this is part of the reason Macworld has moved away from
MacBench to a new test suite that uses application software. It's
impossible to predict how a processor that has a 22% CPU score on
MacBench and a 26% higher FPU score will compare in the real
world.
I guess we'll just have to wait for Macworld to Speedmark the new
iBooks and compare them with the older ones. I'm guessing overall
performance will be about the same for the old 366 MHz iBook SE and
the new 366 MHz FireWire iBook.
It's hard for my wife to judge the performance difference, since
she has also migrated from Mac OS 8.6 to 9.0.4 in changing computers.
As I commented when making the move
a couple weeks back, switching between programs seems a bit more
sluggish, but stability is better.
In the end, what counts is how well the computer works for you.
Benchmarks can be a helpful way of comparing computers and
components, but they only tell part of the story.
Back to the iBook
Regardless of benchmarks, the indigo iBook is faster than the 233
MHz WallStreet it replaced. And if performance is really no better
than the old iBook, it also sells for $100 less, so the value remains
excellent.
Anyhow, after porting all the files over and running the requisite
benchmarks, the next step is integrating files from the old PowerBook
with those on the iBook. Since the PowerBook had Mac OS 8.6 and the
iBook comes with 9.0.4, you have to keep an eye out for software
incompatibilities. It's a good thing I went through that process
last week on my SuperMac
S900 and made notes of the problem software.
The most important tool for this project is Clean-Install
Assistant by Marc Moini. It
goes through a System Folder and pulls out all the third-party
control panels, extensions, fonts, etc. It places them in a new
folder, which you then put on the same level as the System Folder on
your hard drive. Then you dig out problem software and use C-IA to
reintegrate the rest with your new System Folder.
Next, download the latest version of TechTool (1.2.1) from
Micromat to wipe the desktop
and force the iBook to rebuild it and fix all the generic icons.
After all that, I suggest you run your favorite utility program
(Norton Utilities, TechTool Pro, or Disk Warrior, my current
favorite) to make sure all the files are in good shape and optimize
the hard drive.
The Computer Itself
As I said last year, the iBook was never
intended to be a business computer, but it certainly has the right
features. And this year it has a far more acceptable color.
The iBook still has the best keyboard of any portable computer
I've used, and that included the most costly PowerBook. The feel, the
feedback, the typing experience is great.
So is the texture and feel of the wrist rest, a cool plastic with
an almost soft texture that feels just right. As I noted in my
first iBook review, the trackpad isn't recessed below the work
surface. Instead, it's flush with the case - and, like the keyboard,
its feels great and works very nicely.
In fact, I'd have to say the iBook continues to provides the best
tactile experience I've ever had with a portable computer. Everything
about it feels right, from the soft curves and metal reinforced
handle to the grippy plastic casing and the trackpad button.
The screen is positively luminescent; my wife says it's better
than the one her WallStreet PowerBook had. I'd say it's at least a
match for the one in the current PowerBook, except for the smaller
size. As Steve Jobs said when introducing the iBook, it is an iMac to
go.
Conclusion
Benchmark anomalies aside, the iBook remains a lot of laptop for
the money. It has adequate speed and a big enough screen for most
users. It has great battery life, excellent ergonomics, and even a
handle to facilitate transportation. All this and a lower price tag
than last year's model.
I'd have to call it the best portable computer under $2,000.
Further Reading
- <discuss
this article on MacSlash>
