PiFast
|
|
Time
in Seconds
|
|
Albatron
PX925X Pro
|
52.61
|
PiFast
can hit the CPU and memory hard, so without a decent motherboard
to back them, you won't get the best performance. The Albatron
PX925X Pro provides a respectable time of 52 seconds.
CDeX
1.51
|
|
Time
in Minutes:Seconds
|
|
Albatron
PX925X Pro
|
01:15
|
Like
PiFast, our CDeX test shows a good time.
TMPGEnc
2.521
|
|
Time
in Minutes:Seconds
|
|
Albatron
PX925X Pro
|
06:19
|
Video
encoding really shows of the benefit of the extra instructions
in the Prescott CPU's, and like the two previous tests you will
want a good motherboard to make the most of it. Again the Albatron
PX925X hasn't slowed our 3.2GHz CPU down.
Gaming
Tests
Our
gaming tests again give us results we expect with nothing to really
write home about; the numbers are where they should be.
Subsystem
Testing – Audio
We
ran a few tests using Benchemall! and UT2K4 to see what impact
the onboard HD Audio had on system performance. The demo was run
3 times with sound on and sound off using a high quality 1024x768
configuration such as any end user would use, and the average
results recorded.
br-colossus
The
impact as you can see is minimal which is as it should be. I'd
of been worried at any huge differences but this is fine.
In
terms of sound quality, the HD Audio is a lot better than previous
incarnations of Realtek's onboard sound, and any other onboard
sound found on previous generation Intel based boards. The lows
and highs have a greater clarity and power to them and while overall
the solution isn't up there with Soundstorm, the
sound quality certainly is. One nice touch with the Realtek
software supplied is that the ports are auto sensing of devices,
so if you plug in a microphone on the rear pink port, the software
kicks in and asks you what has been plugged in and how you wish
to use it, all displayed on an icon driven list. This functionality
extends to the front panel audio too, allowing you to use 5.1
from the rear and instantly change to 2 speaker headphones from
the front (assuming you can work out the pin layouts for the front
audio panel).
Hard
Drive Performance
We
used HD Tach to gauge read and write performance with a Maxtor
80GB SATA drive. As usual, the disk was freshly formatted, and
configured with only one partition. We also tested with an 80GB
Western Digital on the ITE RAID controller for PATA performance.
Read
Speeds
|
|
Min
kps
|
Avg
kps
|
Max
kps
|
|
SATA
|
27056.1
|
52354.7
|
62984.9
|
|
PATA
|
25998.3
|
50623.9
|
60051.1
|
Write
Speeds
|
|
Min
kps
|
Avg
kps
|
Max
kps
|
|
SATA
|
17432.2
|
27810.3
|
62127.3
|
|
PATA
|
16987.9
|
27054.8
|
59354.8
|
Both
the ITE and the Onboard SATA controllers performed similarly
although of course the SATA interface had the upper hand. CPU
usage was about what we expected averaging around the 8% mark
for the SATA and a slightly higher 9% for the ITE connection.
Network
Performance
Unfortunately
I don't at this time possess any other gigabit network devices
so I was unable to test the Broadcom Gigabit NIC to its full potential.
However I did get numbers from DU Meter for the VIA Rhine III
NIC, along with Windows Task Manager for CPU usage. Our standard
testing method was used; the transfer of a group of small files
totaling 749 MB (varying from 200kb up to 108MB per file) as well
as one large file of 760MB (an ISO). Data was transferred to an
AN7, XP2500+ machine via a 3com 24 port Switch.
Small
Files Test - 749MB Total
|
|
Time
to Copy
|
Avg
Transfer Mb/sec
|
CPU
%
|
|
Upload
|
1:17.1
|
9.10
|
18
|
|
Download
|
1:15.8
|
9.30
|
21
|
CPU
Usage was a little on the high side during the small files test
although this is something we have seen before on other motherboards
with VIA solutions. That said when you are running at 3.2GHz
with Hyperthreading enabled, it's not something that will
affect you overly much. Speeds were OK although we have seen
better.
Large
File Test - 760MB Total
|
|
Time
to Copy
|
Avg
Transfer Mb/sec
|
CPU
%
|
|
Upload
|
1:19.1
|
9.22
|
16
|
|
Download
|
1:18.3
|
9.55
|
19
|
The
large file test gives us better numbers, but transferring one
file rather than multiple small ones inevitably has this effect
anyway. One thing of interest is the fact that running the same
tests on the Broadcom Gigabit NIC yields similar transfer times
but a 10% drop in CPU usage.
NEXT
