|
 |
|
OpenGL Benchmarks with Quake 3 |
|
Anisotropic Filtering |
|
In this final round of Quake 3 OpenGL benchmarking, we ran
the same combination of tests again, this time enabling 4X
Anisotropic filtering, otherwise known as 32-Tap.
With 32-Tap anisotropic filtering enabled, the performance
trend continued with each card. There isn't
much new to say here as the Asus V9280s simply holds its
own against the V8420. The GF4 Ti4600 is relatively
unchallenged in this benchmark, leading the V9280S by
9-10% in each set of scores.
|
|
|
OpenGL: Return of the Jedi |
|
May
the GeForce be with you |
|
Lucasart's
Jedi Knight 2 is another excellent tool for testing a
system's OpenGL performance. This game is
based on a heavily modified version of the Quake 3 engine,
and fully supports advanced shaders as well as high
texture resolutions and effects. With some modifying of
the multiplayer shortcut, you can use the built-in DEMO to
produce some benchmarks. Let's take a look and see
how things shaped up.
The odd thing
in these tests, at least at the lower resolutions, is that
the slowest card, the V8420, seemed to produce some of the
highest scores, even topping the Ti4600 in one case.
While all of the scores were close at 1024x768, we did
notice that the V9280 technically produced the lowest two
of the bunch - not a ringing endorsement of AGP8x.
At 1280x1024, things begin to shape up a bit better, as
the scores fall more or less where we had expected.
The overclocked V8420 still managed to beat out all of the
other card, however. Only at 1600x1200 do we get
what we feel are the best representations of the related
performances of the cards. At normal speeds, the
V9280S beat out the more standard V8420 by 6.7 frames per
second, an 8% difference, but fell a full 9 frames short
of the Ti4600. With the raised speeds, the V8420 at
AGP4x was 3.5 frames better than the V9280S. That
is, until we overlclocked it as well. Overclocking
the card resulted in an additional 6 and a half frames per
second, yet still behind the Ti4600.
|
