|
 |
|
The Hothardware Test System & Testing
Methodology |
|
Pure P4 Power |
|
|
COMMON HARDWARE:
Intel Pentium 4 2.4GHz (2400MHz - 533MHz FSB)
Soyo P4X400 DRAGON Ultra (AGP 8X)
256MB Corsair PC3200 (CAS 2)
On-Board NIC
On-Board Sound
IBM 7200RPM 60GB HD
Lite-On 16X DVD-ROM
Standard Floppy Drive
Windows XP Professional SP1 (DirectX 8.1)
VIA 4-In-1 v4.43
NVIDIA Detonators v40.41 Drivers
VIDEO CARDS TESTED:
Gainward GF4 Ti Ultra 650/XP
Golden Sample (128MB)
Gainward GF4 MX Pro 600/8X-XP Golden Sample (64MB)
MSI GF4 Ti 4200 (128MB)
NVIDIA GF4 MX 440 Reference Board (64MB) |
Methodology:
Due to the fact that we have seen significant
variations in benchmark scores from one site to the
next, we feel it is necessary to explain exactly how
we configure each test system before running any
benchmarks. We chose to test these video boards on the
Soyo P4X400 DRAGON Ultra, with a 2.4GHz Pentium 4
(533MHz FSB), because it functioned at AGP8X.
The first thing we did when configuring this system
was enter the BIOS and select "Load Optimized Defaults".
We then configured the Memory CAS Latency and other
memory timings to be set by the SPD, but ran the
memory at 166MHz (PC2700). The hard drive was then
formatted, and Windows XP Professional w/ SP1 was
installed. After the Windows installation was
complete, we installed the VIA 4-In-1 drivers and then
hit the Windows Update site and downloaded all of the
available updates, with the exception of the ones
related to Windows Messenger. Then we installed all of
the necessary drivers for the rest of our components,
then disabled and removed Windows Messenger.
Auto-Updating and System Restore were also disabled,
and then we set up a 768MB permanent page file.
Lastly, we set Windows XPs Visual Effects to "best
performance", installed all of the benchmarking
software, defragged the hard drive and ran all of the
tests at the CPU's default clock speed. |
| |
|
 |
|
DirectX 8 Benchmarks with 3DMark 2001SE (Build
330) |
|
MadOnion's Flagship |
|
First well start off our tests with one of our favorite
Direct X 8 benchmarking applications, MadOnions 3DMark
2001SE. We ran each card at both 1024x768 and 1600x1200
resolutions
to give a wide range of performance scores. We threw in
an MSI Ti4200 and a nVidia GeForce4 MX440 reference card
in for
comparison.
When we take into consideration that AGP 3.0 provides
twice the bandwidth of AGP 2.0, we would've hoped to see
greater increases in performance. The 650/XP averaged a
gain of 1.5-to-2 percent in each test. The MX Pro posted
larger gains over the reference card, but keep in mind it
is clocked at 275/500 versus the reference card's 270/400 clockspeeds.
|
|
|
More
DirectX with NovaLogic's Comanche4 |
|
Bullet Time is so
cool! |
|
Another popular DirectX benchmarking utility is
NovaLogics Comanche4. Not only is it an effective test, it
is one of the best looking too! Once again we ran the
test at 1024x768 and 1600x1200 to cover the spectrum of
commonly run resolutions. Lets take a look.
Novalogics Comanche4 showed greater gains than we saw
with 3DMark2001SE. The 650/XP boosted its output by
4.7-4.9%. Taking into account its more aggressive clockspeeds, the MX Pro 600 yielded lesser
gains than we saw with 3DMark2001SE because Comanche4 is
very CPU dependant.
On to
OpenGL with Quake 3 |