by
rob-ART morgan, mad scientist
This second page was added to the
original
test page
on 8/20/01 in response to reader suggestions for
more tests that stress disk and graphics
subsystems.
Updated 8/27/02 with expanded
Quake 3
results
using the same three graphics cards in both the DDR
and SDR systems.
Mahalo to my remote mad scientist, Greg, for
results from his new DDR Dual
1GHz.
(Note
to skeptics that doubted Greg's DDR results: On
8/26/02 we re-ran every test in the Bare Feats
local lab using a different DDR 1GHz/MP Power Mac
with even more memory (1.25GB) and got the same
results as he got. This report is not a mistake,
not a fake, and not a fluke as some suggested.)
This first graph
shows what happens when you run two "crunchers"
simultaneously.
Next we tried
running Cinema 4D while duplicating a large file
folder to see if the disk subsystem would steal any
cycles. And if it did, whether it would steal more
from the SDR system than from the DDR
system.
So far we haven't
done anything to stress the Graphics subsystem. We
decided to use Quake3 Arena as our "stressor." We
put the same graphics cards in both machines and
ran them with textures and geometrics set at
maxiumum. Graph below shows average of 3
runs.
LEGEND:
GeF4 Ti = GeForce4 Titanium
Rad 9000 = Radeon 9000
GeF4 MX = GeForce4 MX
I'm not sure what
this next graph proves but it certainly doesn't
help the DDR Power Mac's case.
CONCLUSION
As we saw on
PAGE
ONE,
the NEW DDR Power Mac has no performance
advantage over the old SDR Power Mac running at
the same clock speed.
I'm not saying
that the DDR Power Mac is a dud. I would choose it
over the older SDR Power Mac in a heart beat, what
with 2GB memory max memory, four drive support, and
extra optical drive bay. And the price is
tempting... $2495... a big improvement over the
price of the original dual 1GHz machine.
But the faster
system bus and the DDR RAM don't add the
performance increase we were all hoping for. See
the FIRST
PAGE for
a list of theories on the performance
bottleneck.
RELATED
LINKS
Read about what
Apple has to say about the nVIDIA GeForce4 Titanium
on the GRAPHICS
page of
the Power Mac section. It includes an interesting
graph showing the relative speed of the three DDR
Power Macs running Quake3 Arena. (It looks like
they were running at "High Quality" instead of "Max
Quality" like I use. Regardless, the 1.25GHz DDR
Power Mac is only 13% faster than the 1GHz DDR
Power Mac!)
See nVIDIA's
specs on the GeForce4
Titanium.
(Apple's card is comparable to the 4600 chip
set.)
ATI has a
description page for the Radeon
8500 Mac Edition
and Radeon
9000 Mac Edition
on their site.
Anandtech
compares the Radeon
9000 to
the 8500, GeForce4 MX, and GeForce4 Titanium AGP
cards for Windows PC's. They also compare the
Geforce4
Go to the Radeon 9000
Mobility
(both are NEW chips for laptops.)
Read about the
new Radeon
9700 (Mac
Edition coming in a few months), which, according
to Anandtech.com's
tests, is 30 to 50% faster than the GeForce4
Titanium... at least when run on a Windows
PC.
TEST
NOTES
The "SDR" Power
Mac 1GHz MP had 1GB of PC133 CL2 SDRAM.
The "DDR" Power Mac 1GHz MP had 1.25GB of PC2700
CL2.5 DDR RAM
Both booted and ran apps from an IBM 120GXP drive
with OS X "Jaguar" installed.
Graphics cards
tested on both the DDR and SDR system were:
The 64MB GeForce4 MX (standard in the SDR Power
Mac)
The 64MB Radeon 9000 (standard in the DDR Power
Mac)
The 128MB GeForce4 Titanium (optional on both SDR
and DDR Power Mac)
For details on
each real world test, read "HOW
I TEST."
|