Interactive VGA Charts 2006 Completely Inaccurate?

[Ian]

I was just browsing the VGA comparison charts, http://www23.tomshardware.com/graphics.html, and I can't figure out the results. Are the testing methods not comparable to the ways we test?

The results for 3DMark 06 are not anywhere close to those found by the consumer. Anyone know where the discrepancy is?

Also, results for FEAR, AOE3, COD2, etc, all seem WAY WAY WAY too low. I realize that the settings are on max details, however one would think that the results wouldn't be that low. Also, SLI configurations often perform WORSE then single card solutions in these tests. Anyone have any ideas? because as far as I can tell Toms wasted a lot of time doing these tests, yeah they show relative performance, I guess, but still...

 

[PX7800GT]

Also, results for FEAR, AOE3, COD2, etc, all seem WAY WAY WAY too low.

You Are 100% Right :twisted:

Also, SLI configurations often perform WORSE then single card solutions in these tests. Anyone have any ideas?

It`s True , Some Games Like Half-Life 2 Shows Lower Frame Rate In SLI Mode ...

 

[mad_fitzy]

if u read a few of the sli/crossfire reviews u will see that in quite a few of the tests single card solutions outperform the dual card configurations.

 

[ltcommander_data]

Yeah that is weird. Part of it might be the drivers since they are still using Catalyst 6.1 and 6.2 and ForceWare 81.98. I also don't agree with their CPU choice. I would think that if you are comparing graphics cards you would use the fastest CPU available to avoid bottlenecks, but they only used a 3.2GHz Pentium D 840. Even if they wanted to stick with a dual core Intel, they could have at least used a 3.46GHz Pentium Extreme Edition 955 or even a 3.73GHz 965EE.

Also, while the describtion of the RAM tells you what model they used it doesn't tell you what it was actually set at. The ATI and nVidia setups are also using different models of RAM. Updating to Windows XP SP2 would probably make it more representative as well.