Quote Originally Posted by gosh View Post
I need to match what I am seeing in the pattern from other tests. And this explanation matches very well to this test.
http://www.guru3d.com/article/cpu-sc...e-processors/9

If you want me to listen on your arguments then you need to say something that is different from the pattern that I have seen on other reviews. What I have seen about nVidia 8800GTX is the same as this review jack did. It does have the same pattern.
Also one wish, could you please stick to the technology and skip the personal attacks. Yes I have been trying to get some information and posting on some forums in order to get some information that contradicts what I am saying. I have also done a test program and have done some test on different types of operations, it is possible to get more than twice the speed in memory use with a low clocked AMD compared to one high clocked Intel if you use instructions that works well on AMD just using memory (no GPU).
memory bandwidth means jack on desktop, you could quadruple the memory bandwidth and you wont see different results on 90% of standard desktop apps.

Memory bandwidth is a important factor in HPC/server (4p) environment, but not on desktop. And thats why we see nehalem goes with IMC to cover server (4p) HPC space where amd is still top.