Performance Benchmarks :
Games related tests were carried out using the Creative
Labs 3D Blaster GeForce 3 Ti500.
Quake 3 Arena :
We couldn't make it through the benchmarks without
digging out the evergreen Quake 3 Arena. Even with a Ti500
you can't complain about 109fps in 1600x1200. Q3A does benefit
from the Morgan's added SSE functions and again convinces
us here that the Duron 1.3 is at last a fully fledged choice
for the gamer on a budget as well as the home and business
user looking to save a few quid.

GL Excess :
A great OpenGL benchmark that regularly finds
favor here at 3DV for its consistent results. The similarity
between the Duron and XP results show how graphics dependent
many of the tests are, though the CPU/FPU specific test point
to a 35% advantage to the XP.
|
Legend
|
|
FILL
RATE Tests
|
|
POLYGON
COUNT Tests
|
|
VRAM
Tests
|
|
CPU/FPU
Tests
|
|
Creative
Ti500 - Duron 1.3
|
|
6024
- XSMARKS
|
Scene 1
|
Scene 2
|
Scene 3
|
Scene 4
|
Scene 5
|
Scene 6
|
Scene 7
|
Scene 8
|
Scene 9
|
Scene 10
|
Scene 11
|
Scene 12
|
|
1024x768
32bpp
12/15/2001
13:36
|
MIN
|
117
|
76
|
113
|
67
|
105
|
123
|
82
|
90
|
93
|
54
|
122
|
124
|
|
AVG
|
163
|
101
|
212
|
77
|
175
|
157
|
108
|
114
|
117
|
84
|
158
|
119
|
|
MAX
|
201
|
127
|
258
|
100
|
312
|
192
|
130
|
129
|
144
|
166
|
220
|
167
|
| SUMMARIES |
5241
|
6280
|
3440
|
4404
|
|
|
Creative
Ti500 - AthlonXP 2000+
|
|
6024
- XSMARKS
|
Scene
1
|
Scene
2
|
Scene
3
|
Scene
4
|
Scene
5
|
Scene
6
|
Scene
7
|
Scene
8
|
Scene
9
|
Scene
10
|
Scene
11
|
Scene
12
|
|
1024x768
32bpp
12/15/2001
13:36
|
MIN
|
121
|
108
|
146
|
66
|
110
|
124
|
82
|
99
|
99
|
53
|
124
|
125
|
|
AVG
|
164
|
140
|
216
|
77
|
179
|
157
|
108
|
114
|
120
|
85
|
158
|
120
|
|
MAX
|
201
|
167
|
252
|
100
|
401
|
191
|
142
|
126
|
144
|
143
|
209
|
167
|
| SUMMARIES |
5294
|
7045
|
3419
|
5968
|
|
SYSmark 2001 :

SYSmark 2001 uses a
suite of 14 applications to recreate a day in the life of
your average user. The benchmark runs a variety of routines
in a variety of applications including Photoshop 6. Flash
5, Dreamweaver 4, Microsoft Excel and many more. Natural
pauses are inserted into the routines to mimic real life
usage and the results are based on average response times
to called routines. I've heard many people complain that
SYSmark is strongly Intel biased, but in all honesty, provided
it does exactly what it claims it does when compiling the
final ratings, it simply can't be.
This isn't the greatest
of showings from the Duron, with even the Athlon 1.2 showing
it the way with an average overall SYSmark score of 133.
Quite why this should be the case is not certain, perhaps
it's entirely due to the lack of L2 cache, something that
this suite of tests would make ample use of.
Page10 - Overclocking
and conclusion 