|
FIC
AD11
|
640
x 480
|
800
x 600
|
1280
x 1024
|
|
Fastest
|
162.9
|
155.1
|
128.0
|
|
Normal
|
149.2
|
147.8
|
110.1
|
|
High
Quality
|
148.4
|
139.2
|
64.1
|
|
Tyan
KT-A
|
640
x 480
|
800
x 600
|
1280
x 1024
|
|
Fastest
|
151.7
|
142.3
|
121.0
|
|
Normal
|
137.6
|
134.4
|
106.0
|
|
High
Quality
|
135.3
|
126.0
|
63.2
|
|
AK73
Pro
|
640
x 480
|
800
x 600
|
1280
x 1024
|
|
Fastest
|
121.3
|
118.1
|
96.3
|
|
Normal
|
114.2
|
112.4
|
95.1
|
|
High
Quality
|
113.3
|
103.4
|
62.2
|
I think this shows that the Absolute
MultiMedia GTS is fill rate limited. Look at the high resolution "High
Quality" scores to see what I mean. Those scores are within 2 points
at 1280 x 1024. That is why I picked that resolution in the first place.
But there is no comparison in the resolutions that most gamers would use.
The Trinity KTa scores are pushing the results that I got with the FIC AD11
at 1.2 GHz and with DDR memory and simply destroying the results with the
AK73. This benchmark shows clearly not only how memory dependent a game
like Quake III Arena is, but also how much BIOS optimizations can affect
gaming performance. I am really starting to wonder why every board manufacturer
is not including 4-way interleaving in their BIOS setup area.
