For game testing, we decided to focus our limited testing time on low screen resolutions and low visual quality settings, so that graphics subsystem bottlenecks wouldn't mask differences in CPU performance. Some of the image quality settings were awfully low given the potency of the Radeon X1900 CrossFire cards in these systems, but we wanted to keep the spotlight on the CPUs.
We ran F.E.A.R. at a lowly 640x480 resolution (confirmed to be the same on both boxes) with the computer performance set to "Maximum" and the graphics card set to "Low." The test was the game's built-in "Test settings" benchmark based on an in-game cut scene animation.
For Half-Life 2: Lost Coast, we used the IQ options you can see below at a resolution of 1024x768. We then ran a timedemo of a custom demo recorded by Intel.
We kicked off the UT2004 timedemo using a script that ran the game and then launched into a custom demo at a resolution of 1024x768.
Finally, we have a Quake 4 timedemo, which we confirmed ran at the same settings on both machines. Like a genius, I completely failed to type those settings into my notes because my laptop battery was dying, and I didn't manage to snap a screenshot, either. This one was also a low-res, low-quality setup to emphasize CPU performance above all else.
These gaming numbers for Conroe look tremendous. We also have some media encoding benchmarks.
|The SSD Endurance Experiment: Only two remain after 1.5PB||65|
|Friday night topic: Conspiracy theories||221|
|GeForce 344.11 WHQL drivers support new cards, new games, G-Sync||7|
|Deal of the week: A 23'' IPS monitor for $150, a 200-mm fan for free, and more||24|
|GeForce GTX 970, 980 cards already widely available||31|
|Curved VA panel powers 27'' Samsung monitor||25|
|Android L to encrypt devices by default||7|