Benchmark Results

3D Games

Doom 3 sees huge gains at nearly all resolutions as a result of overclocking the system to its maximum CPU speed and minimum RAM latencies. Is this a sign of things to come?

F.E.A.R. seems much more biased towards the graphics card, which wasn’t overclocked in our system. Moderate gains drop towards zero as resolutions increase.

Oblivion shows reasonably good gains at most resolutions, and proves something else that’s interesting: our build was more stable at each overclock setting than it had been at stock speed, where it had crashed at the highest setting in the outdoor scene.

Audio

Lame and OGG are the first benchmarks to show moderate performance gains for the initial overclocking attempt, but these also exhibited exceptional results from the final overclock.

Video

DivX and XviD suggest that when you’ve seen one CPU-bound benchmark, you’ve seen them all. The video encoding results appear to shadow those of audio encoding benchmarks.

Applications

Synthetics

Leads from overclocking hold up to increased resolutions far better in 3D Mark 2005 than they did in actual games.

Most of the abbreviated-applications (called "viewsets") from SPECviewperf showed reasonable gains from CPU and RAM adjustments, but Solidworks strangely showed a slight decrease.

PC Mark CPU and memory scores are spot-on for what we expected from our settings. The hard drive performance probably shouldn’t have changed, but the results seen here might have been caused by differences between runs.