Originally Posted by zmatt
In all fairness whether its a "true" quad or not doesn't make that big of a performance difference. That is probably why Intel doesn't talk about it much. AMD sorta shot themselves in the foot by hyping Phenom like that and it ultimately didn't deliver.
being monolithic, the Nehalems will have a distinct advantage over MCMs (ie Core 2s)... you see, because Core 2s are 2 distinct dies in a single module, the bandwidth between the cores on one die is huge and the latency is tiny, but between the separate dies, the bandwidth is limited because they must communicate through the FSB... because Nehalems are built on a single die, each core can communicate with the others directly, decreasing the overhead at load.
The cache can be accessed by every core on a Nehalem processor as well, instead of the older core 2 architecture, in which the cores could only access the cache associated with their die. this increase in cache bandwidth and reducing in intercore latency allows nehalem to excel in complex calculations and high end benchmarks (eg SuperPi)