Quote Originally Posted by gosh View Post
No it isn't a fact but is one possible explanation. Most people just say it is badly optimized but no explanation on what could be bad.

If it is using three threads and each thread works against same memory.
Nehalem will the load the memory once, because each thread can access same L3 cache. C2Q needs to load the memory twice. C2Q = 2xC2D = no communication between caches on each C2D. That could also be one explanation why C2Q need to work that much more compared to Nehalem
You and your FSB issue again. Is that why GTA 4 runs much better on a C2Q with its FSB than Phenom?