Test Methodology: How Do You Make It A Fairer Fight?
Once again, all of the tests we're running today are being driven by AMD's Phenom II X4 955 BE at its default clock speed of 3.2 GHz. After we ran all of the tests, we disabled one CPU core using the operating system, rebooted, and then ran the tests again until we had results from one to four CPU cores. The method we use to disable the CPU cores in the Windows Vista operating system is documented in the previous article, here.
To test the validity of the results, we will run all of the tests again on a separate CPU, the Phenom II X2 550 BE. With its multiplier raised to 17, its clock speed will be identical to that of the Phenom II X4 955 BE. This way, we can compare the results of the Phenom II X4 955 BE with two CPU cores disabled to a true retail dual-core Phenom II X2 550, validating our test methodology.
For our new concurrent application test, we will run an AVG anti-virus scan in the background at the same time as the World in Conflict game benchmark. AVG is an ideal application for this test because real-world users will often find automatic scans running in the background while they're performing other tasks, and World in Conflict is an ideal game benchmark to use because it captures both average and minimum frame rates.
With a good grasp on our metrics, let's move on to the hardware and examine the details of our test rig.