Quote Originally Posted by Viper666 View Post
For the last time its been posted here already 100 times,
The x87 math part of Phenom is exactly the same as K8, that part of the processor was not changed, only the SSE has been changed, Super PI doesnt use SSE, Real world apps like the tinker core from Folding @ Home use this part of the processor(X87), and K8, Beat C2D , And K10 does 2. Why would AMD change a part of the processor that they already Murder Intel @, Im pretty sure Daily Tech put up numbers backing this up.Its simple, Super PI is a quick hitter, C2D fills its pipe alot faster than AMD, but its a much smaller pipe, AMD has a fat pipe it takes a lil longer to fill, Super PI likes the smaller pipe better, Phenom will kill Intel at Memory Intensive stuff, If the data set doesnt fit in Intels cache, AMD wins! If it does fit in the cache Intel wins, Ive seen this posted on many Super Computer forums, where people are doing real work on clusters, not Running bench marks.
My thoughts exactly. SuperPi should be burried deep under ground , or re-write a new one from ground up that fully supports the new multi-cpu's. Heck , even P4 still gets better numbers in superpi then K8. But what about 99.9% of all other applications. All those who think superPI is the chit then you're the weak link. Goodbye.