OVERCLOCKING Overclocking provided stable frequencies of 410MHz/1.1GHz with no signs of
artifacts or overheating. To test initially for excessive temperatures and
other graphic problems I used the synthetic benchmarks 3DMark2003 and
AquaMark3. I closely monitored temperatures and found running 3DMark2003
in one instance, generated the maximum temperature of almost 90C.
Temperature at idle fluctuated between 58 and 61C. Playing games rarely
exceeded 84C with the norm being around 77C for the average. While I
experienced no artifacts or other graphical deterioration while using overclocked timings in game play during this review, approaching 90C in a
synthetic benchmark is of concern. I intend to try to find a more
efficient cooling solution for this card after this review. It should be
noted that temperatures never exceeded 84C while gaming with default clock
frequencies.
While overclocking enhanced performance in Call of Duty, at high resolution
and settings, cpu limited games such as UT2004 proved the A64 3000+ to be a
bottleneck. This resulted in overclocking the 6800 GT of little or no benefit over default
clock frequencies.
CALL OF DUTY We will begin this review's tests with CoD which maintains high popularity
among gamers. The benchmarks used are nV News' Dawnville and POW Camp recorded
demo maps. Resolutions used were 1024x768, 1280x1024, and 1600x1200.
Antialiasing and anisotropic filtering settings were in combinations of No AA /
AF, 4x AA / 8x AF, and 8XS AA / 16AF. Each resolution and AA/AF setting is
identified in the result tables along with whether the eVGA 6800 GT was
overclocked (410/1.1) or set at default timings of 350/1.0 for the particular
benchmark. For comparison the X800Pro was run through the same resolution and
settings. No overclock results were provided for the X800Pro and default
settings of 475/900 were maintained.
nV News' custom demos from the Dawnville and POW Camp maps were used to test performance of the
eVGA 6800 GT. The ATI X800 Pro was used for comparison of
performance. Note that all in-game
graphics and performance settings were configured to their highest quality. Sound was set to Miles Fast 2D
Positional Audio at 44khz.
Call of Duty Results
AMD Athlon 64 3000+
Dawnville Demo
6800GT
No
AA/AF
4X
AA/8X AF
350/1.0
410/1.1
350/1.0
410/1.1
1024x768
177.2
176.1
174.8
174.7
1280x1024
176.5
176.0
166.2
168.6
1600x1200
170.5
171.7
149.3
156.8
X800
PRO
No
AA/AF
4X
AA/8X AF
1024x768
154.5
155.0
1280x1024
153.0
152.4
1600x1200
146.2
147.6
6800GT
No
AA/AF
4X
AA/8X AF
350/1.0
410/1.1
350/1.0
410/1.1
1024x768
120.1
120.3
118.9
121.2
1280x1024
119.7
119.9
118.1
119.3
1600x1200
108.6
112.9
85.6
94.3
X800
PRO
No
AA/AF
4X
AA/8X AF
1024x768
104.2
104.1
1280x1024
103.6
103.7
1600x1200
101.5
101.4
CoD provided very little stress on the eVGA 6800 GT as evident by the
results. No more sub 40 frame rates here, even at the highest resolutions
used. Actual game play was excellent even at the highest settings of 8XS
AA / 16X AF with resolutions up to 1600x1200. X800Pro makes a good
comeback at 1600x1200 with antialiasing and anisotropic filtering on. In
fact, application of AA/AF had little affect on the X800Pro performance.
Going to AA/AF gave just an average of 0.1 frame rate per second drop on the
average of three runs of the POWcamp demo. The 6800 GT lost 13 FPS or
almost 8% going from no AA/AF to 4x AA / 8x AF, at 1600x1200, on the average
through three runs.
Overclocking gave an increase of 4.3 FPS average at 1600x1200, no AA/AF,
while applying AA/AF gave 8.7 average FPS increase over default settings on the
6800 GT. With these results it is evident that overclocking is of no real
benefit on my system with such high frame rates as produced by the 6800 GT.
