|
 |
|
SiSoft Sandra 2003 Professional |
|
The Synthetic View |
|
To give us a clearer picture of
how a particular system compares to other popular
configurations, we enlisted the help of SiSoft's Sandra 2003
Professional. Sandra has a comprehensive internal
performance database that can be used to gauge a system's
overall potential, helping us to get a clearer idea of what
a system should be capable of.
CPU @
1.84GHz.
|
Multimedia @ 1.84MHz.
|
Memory @ 333MHz.
|
|
Hard
Drive
|
CPU @
2.07GHz.
|
Multimedia @ 2.07MHz.
|
Memory @ 376MHz.
|
With the Barton Processors, the
CPU tests are a bit deceptive. If you compare the
results you will see that the CPU based scores are on par
with that of an Athlon XP 2200+. This is because the
test uses the processors actual speed as a gauge. The
Barton 2500+ does run at 1.8GHz with a bus speed of 333MHz
as opposed to the 2200+ running at 266MHz. The
synthetic test is also unable to show the performance
benefits of the 512KB of on-die cache compared to the
2200+'s 256KB of on-die cache. So when looking
at the CPU tests, you'll need to read between the lines.
Since our processor runs with a 333MHz bus, that is the
speed at which the memory was tested.
The memory results show the improvements of the memory
performance with the KT400A when compared to the KT333.
Once we overclocked the system, the memory scores really
improved, pushing the KT400A board to the front on the pack.
While SiSoft Sandra is a good
tool for us to get an initial impression of a system there
are a few other results we like to look at to give us a more
balanced picture.
|
|
|
Futuremark's 3DMark03 and PCMark 2002 |
|
Futuremark's Finest |
|
Next up we have some numbers
from PCMark2002Pro and 3DMark03. PCMark2002Pro has proven
it is a good synthetic tool for gauging a system's overall
performance. 3DMark03 on the other hand, is turning
many people off. There is
a big debate going on as to the validity of the results with
certain hardware. Nonetheless, we've included it in
this review since we used the same video card with all
tests. We'll start
things off with a run at PCMark2002Pro, focusing on the CPU
and Memory results, then we'll follow that up with a run at
3DMark03. With each test we compared the scores to
those of an ASUS A7N8X Deluxe motherboard.
With the CPU results, the KT400A
board was outpaced by the nForce2 board by roughly 100
points, a 1.9% difference. The real surprise came when
we compared the two board's memory scores. The A7N8X
Deluxe took a firm lead over the "LAN Party" board,
exceeding it close to 10%. While VIA has made some
improvements to their memory controller performance, it
appears that the nForce is still on top.
The results from the 3DMark03
test was far closer, although the ASUS board still held the
lead by a few points. Nonetheless, the DFI "LAN Party"
KT400A put up respectable scores and we can't draw any
conclusions based on synthetic benchmarks alone. Next
we'll run Comanche 4 and Quake 3 to give us a look at some
real world gaming potential.
|
|
|
Quake 3 and Comanche 4 |
|
Gaming Tests |
|
When we start talking about
gauging a motherboard's gaming performance, Novalogic's
Comanche 4 has proven to be an excellent test because it
relies so heavily on the CPU. Let's see how the DFI
"LAN Party" compared to the nForce2 based system when we
ran the default benchmark with audio disabled.
With this DirectX benchmark, the
DFI board kept things very close, tipping the lead to the
ASUS board by less than half of a point.
The
Winstones and Final Words
| 
