C2Q vs Core i7 - 3d Grap - Medium

[cnick6]

Has anyone noticed a huge difference between their 3D Graphics - Medium tests when comparing these two processors?

I'm just curious as to the technical reason. The 3D Graphics - Simple and Complex are both about the same differences. Why would the Medium test show a huge difference? Could this test be faulty?

FYI, I am comparing C2Q and Core i7 with ATI Radeon 5870 adapter. And both baseline results were using the same driver version. (In fact, a baseline result with an older version showed this same huge difference.)


Thanks
-Nick

[David (PassMark)]

It would help if you posted the full specs of the two machines and the actual results you obtained.

[cnick6]

Okay I'll do this when I get home later tonight.

But the difference from memory was over 1000% improvement just between the C2Q and Core i7 with the same video card and driver version.

The other 3D tests were roughly 140% better.

I'll post the results soon.

[cnick6]

Here's some screenshots of the systems in question and the 3D test results:

[David (PassMark)]

Several of the full screen test results look like they are limited to the 60Hz monitor refresh rate. Anything with a frame rate of 59 /60 is suspicious. There are settings in some device drivers control panels and in the PerformanceTest Edit / preferences window to check this. But might also be just coincidence.

The average results for the HD5870, with the Q9550 are about,
Simple, 3000
Medium, 800
Complex, 65
DirectX 10, 48

Any machine that varies a lot below this probalby has configuration issues.

[cnick6]

Thanks for your help. I figured out the problem. When I originally installed the ATI drivers, I must have modified their default settings. All of the 3D/AA settings were maxed out and thus slowing down the test results.

I was pretty certain that the Core i7 CPU shouldn't show such a wide margin of result.

I don't know if there's any way you can control this in the Passmark software, but it would probably help your baseline numbers if there was some way to temporarily "override" the ATI Catalyst settings while running the benchmark. Otherwise, you will collect misleading results.

OR -- you could also probably read the ATI registry key to see if the settings are default or custom. The summary report could then show the fact that the settings are custom. It's just an idea.

Here's a screenshot to show you it's nearly identical to the other system: