While I was reading the article link here showing the nforce benchmarks i noticed something.
<< Basically both the CPU and the GPU have 4.2 gigabytes of bandwidth availability at any given time. nVIDIA claims this will allow for up to a 30% performance gain over comparable RDRAM based systems (in terms of memory bandwidth of course). >>
But do you notice that in the stream benchmark they fail to include benchmarks from the Intel's 850 chipset, strange. If my memory serves me right, the Intel 850 shows better numbers in STREAM then the n-force, someone correct me if I'm wrong.
<< Basically both the CPU and the GPU have 4.2 gigabytes of bandwidth availability at any given time. nVIDIA claims this will allow for up to a 30% performance gain over comparable RDRAM based systems (in terms of memory bandwidth of course). >>
But do you notice that in the stream benchmark they fail to include benchmarks from the Intel's 850 chipset, strange. If my memory serves me right, the Intel 850 shows better numbers in STREAM then the n-force, someone correct me if I'm wrong.
