August 15, 2000
Overview: A while back we were going over the upgrade cards
we have reviewed in the past and realized that Newer Technology's
cards were poorly represented. To begin the process of filling
in the gaps in our coverage, we take a look at two PCI upgrades
from Newer Technology, the G3/500 and G4/350. The G3/500 represents
the high end of the G3 processor line whereas the G4/350 is
a modest entry to the world of G4's and AltiVec (aka "Velocity")
processing. We put both cards through their paces to help
you decide which is best for your needs.
| Supported Models
Apple: 7300, 7500, 7600, 8500, 8600, 9500, 9600
Power Computing: Power Tower/Pro, PowerWave, Power Curve*,
UMAX SuperMac S900 and J700.
* The G4 card will not work in the Power Computing
PowerCurve or PowerCenter/Pro machines
Installation: Installing both cards is a breeze thanks to
the well written and clearly illustrated instruction manuals.
The G4/350 has no dip switches or jumpers to adjust, further
streamlining the installation process. The G3/500 on the other
hand has a row of 4 switches, all set to the "off"
position. This setting will work for most users with the exception
of folks who own 8500 and 9600 Power Macs. For installations
in these machines, switches 1 and 4 will need to be flipped
on for the card to work.
In the process of conducting a previous review, we damaged
one of our ZIF processors by not taking appropriate static
precautions. After this last experience we gained a renewed
appreciation for the included grounding strap and extension
bar on the card itself. The latter gives you a safe place
to grip while nudging the card into place and the former will
help you avoid our unpleasant experience. Use the grounding
The installer places several items on your hard drive including
an extension, control panel, and the Gauge Pro software. Unfortunately,
this means that information on the operation of your card
is divided between these two latter locations. The control
panel, for example, doesn't include information on the card's
operating temperature, processor version or system bus speed,
relegating these details to the Gauge Pro software. I would
like to see this information centralized, but this is a relatively
Stability/Performance: Both cards performed as expected.
That is to say the G3/500 easily bested the G4/350 in applications
whose functions have not been optimized for the AltiVec instruction
set. In cases where AltiVec could be utilized, the G3 began
to show its age, coming in last in spite of its 150MHz advantage
over the G4 card. Either card represents a significant performance
boost over the stock 9500 thanks to their 1MB backside caches.
The cache on each card runs at the standard 2:1 ratio, 250MHz
for the G3/500 and 175MHz for the G4/350. Always wanting to
push the envelope and not having the option of overclocking
the processor, I tried setting the cache to a faster operating
speed via the MaxPowr control panel. The G3 card simply didn't
allow the faster setting to be selected. The G4/350 was more
cooperative, allowing me to make the selection, but the 9500
repeatedly crashed with the faster setting. Both cards were
rock solid at their default settings though, and ran through
all of our tests without any glitches. If you don't use a
lot of applications that can take advantage of the AltiVec
instruction set, the G3/500 represents the better deal of
these two cards. If, however, you use Photoshop or other AltiVec
enabled applications on a daily basis a G4 card will quickly
pay for itself in time saved. Read on for benchmark scores...
Test Machine Configuration: Our test machine was a 9500 with
96MB RAM and OS 9.0.4 installed. We tested with an extension
set comprised of all OS extensions plus those installed by
Newer. For the MacBench tests virtual memory was turned off
and disk cache was set to 512k. These settings are consistent
with those used on the MacBench base reference machine, a
beige G3/300. For the real world tests we turned virtual memory
on and set it to 97MB.
|Estimated Street Price:
||Better performance for your $$$, stable
at default settings, well written & illustrated manual,
extension bar on cards helps during installation.
||Beats "faster" G3/500 on AltiVec
enabled tasks, stable at default settings, well written
& illustrated manual, extension bar on cards helps
||Processor and backside cache can't be overclocked,
control panel lacks comprehensive information.
MacBench 5.0 Results
The scores below are from MacBench 5.0. MacBench
5.0 is a subsystem-level benchmark that measures the performance
of a Mac's processor, disk, and graphics subsystems to name
a few. MacBench normalizes all scores relative to the base
machine, a Power Macintosh G3/300. The base machine receives
a score of 1000. For all MacBench tests, higher numbers mean
better performance. Be aware that MacBench 5.0 was released
before the G4 and AltiVec
hit the scene and, consequently, does not take advantage of
the extra instruction set. For more detailed information on
MacBench click here.
"Real World" Tests
(Shorter bars are better)
Time to Scroll a 574 page AppleWorks document
from top to bottom.
Using the same document as above we did a
search/replace command to replace the word "the"
with the word "macbench"
Photoshop 5.5 "Real World" Test
Render Boy 2.2.0
Time to render "Pool Table" Example
SoundJam MP3 Encode
Time to encode a CD track 4 minutes 29 seconds
Log File Analysis
For this test we used FunnelWeb Enterprise 4.0 to analyze
our web server's log file. The file was 17MB and contained
1.8 million lines to be processed!
Variable/Fixed Clock Rate
Toll Free Phone
Toll Free Phone