HardOCP heeft een paar interessante Quake3 benchmarks gedraaid op een dual Celeron box:
The question we really want answered, of course, is how running high-poly settings will affect performance. And, to throw a new twist at the question, how running high-poly settings in a dual-processor setup will affect things. Could it be possible, as I suggested, that a second CPU could be used as a geometry accelerator of sorts? Could an SMP rig run Q3 with the polygon detail settings turned up at about the same speed as a single-CPU rig would run the default settings? In theory, at least, it seems like a possibility, since transformation and lighting happens on the CPU, not on the video card, in almost all the current 3D accelerators, including the Voodoo 3, TNT2, etc. Why not put that second CPU to use processing more polys per scene?
Check de resultaten op de frontpage van HardOCP.com.