|
Untitled Document
Testing
Testing was performed using an Athlon 64 3500+ which clocks in at 2.2GHz. For
some reason, this board only ran at 199MHz FSB which meant we were running at
2189MHz. The graphics card used was an X700 Pro 256MB with clock speeds of 425MHz
on the core and 432MHz on the memory. We used 2 x 512MB Corsair PC4400 memory
modules.
Considering the nature of this board and lack of comparison products, testing
is minimal and will be left until a later date. By the running the system memory
in a single channel configuration, we can get an idea of what performance would
be like on Socket 754.
Doom 3 and Far Cry were run at 1280x1024 with 2x AA and 2x AF. 3D Mark 03 was
run at default settings. All results are in frames per second.
| |
3500+ (Dual, X700 Pro) |
3500+ (Single, X700 Pro) |
| Doom 3 |
27.0 |
27.0 |
| Far Cry |
34.33 |
34.31 |
| 3D Mark |
7125 |
6932 |
Gaming performance is obviously being limited by the graphics chip itself,
as a 9800XT on a similar Socket 754 platform was capping out at 27.1fps with
the same Doom 3 configuration. 3D Mark shows a slight performance difference
in single channel configuration but this is more than likely being skewed by
the software rendering tests.
By reducing the settings of Doom 3 to 1024x768 with no AA or AF, we can see
if any real world difference is shown.
| |
3500+ (Dual, X700 Pro) |
3500+ (Single, X700 Pro) |
| Doom 3 |
53.9 |
52.9 |
These results indicate that the extra memory bandwidth is hardly needed when
it comes to gaming. With a faster graphics card, results may be more interesting. For those of who aren't aware, a 3500+ on Socket 939 runs at the
same clock speed as a 3200+ on Socket 754. The extra 300+ is considered worthy
because of dual channel memory support. By running our system in single channel
mode, we have direct comparison to the K8T800 chipset system used in our Athlon
64 Chip Guide. We used the same video encoding test as performed in this article.
As a recap, this was done using XMPEG, encoding to DivX 5.1.1 (780Kbps) and MP3
audio (128kbps). We used a 20,000 frame section of "The Mask" DVD.
| |
3500+ (Dual) |
3500+ (Single) |
3200+ (Single) |
| XMPEG |
33.87 |
32.42 |
32.68 |
In this case, an extra 1fps is gained with dual channel memory. Although this
doesn't make a huge difference in reality, this is a similar increase to an extra
100MHz clock speed. But we still don't see why this justifies the extra 300+.
<< Back | 1 | | 2 | | 3 | | 4 | | 5 | Next >>
|
