The POV-Ray 3.7 beta, our standard rendering benchmark has been updated to the latest build, version 22. We use the standard benchmarking scene, with performance reported in pixels rendered per second.
Figure 2 – POV Ray Performance
For POV-Ray, Harpertown is only slightly faster than Clovertown on a per-clock basis – about 2.3%. This is not surprising, since POV-Ray is almost entirely cached to start with; the extra bandwidth from the chipset, and additional cache does not help much and none of microarchitectural changes provide much of a benefit either.
Note that in general, comparing the IPC between a 2.33GHz and a 3GHz chip is not entirely accurate. However, in the case of POV-Ray performance scales pretty much linearly with respect to frequency, core count, etc. so the comparison is valid.