|
|
||||||||||||
|
Test Bed
Perhaps the most interesting part to measuring the performance of DDR 2 is just how dependant it is on the FSB. No matter how fast the memory can go, or how low it's latencies can get, it will be forever limited by the constrictions of the Front Side bus. Thus, we do four tests versus the conventional two. We test the ram at stock, then overclocked speeds on a non-overclocked CPU, then we ratchet up the FSB, and re-do the tests. Read on to find out the results - you may just be surprised. Stock CPU Interestingly enough, we can safely say that speed is not king with DDR2 at the moment. This is largely due to the system bandwidth restrictions put into place via a narrow front side bus, which undoubtedly is holding the playing field even. What does this mean for the average consumer? Essentially, the only reason to buy the fastest memory possible at this moment in time is if you choose to overclock, after all, when you're pushing your FSB in excess of a gigahertz, it's best to have memory that can scale up with it. With the similarity in performance between DDR2-667 and DDR2-800 at this point, let's move on to see if an overclocked CPU helps out.  Overclocked CPU Once again, it seems that having a wide open FSB actually yields the best performance, as we can clearly see the memory working at 1:1 with the FSB, going along at 875 MHz is by a small stretch, the slowest of the bunch. This further re-enforces the statements made earlier, that the lone purpose for high speed memory at this point in time is to overclock. As a side note, the CPU is only clocked to 3.28GHz during the 875MHz portion of this test.    
|
|
||||||||||
