I'm a newbie when it comes to memory dividers, what exactly does it do for you, and I hear all the time people talking about 1 to 1 just like you mentioned there, almost implying that 1/1 was better then the other options?
Concerning computer architecture, the cpu communicates with the main memory (RAM) through a bus called the Front Side Bus (FSB) or HyperTransport (Terminal?) (HTT). These three components (CPU, FSB, Memory) usually all run at the same frequency. The memory is typically not able to run at the same high speeds as the cpu and FSB, therefore the memory will run at a slower speed. This slower speed of the memory needs to be aligned in some manner to still be able to communicate with the higher speed of the cpu so certain speeds are selected, which correspond to either a (A:B), 1:1, 5:6, 4:6, or even a 3:6 relationship. This means that the memory is running at 'A' clock cycles for every 'B' clock cycles of the cpu. If the memory is running slower than the cpu, then some cpu cycles are not able to access the memory.
Therefore, a 1:1 relationship means the memory runs one cycle for every one cycle of the cpu and no cpu cycles are lost. In a 5:6 relationship, the memory runs 5 cycles for every 6 cycles of the cpu and one cpu cycle is lost for every six. The AMD64 possesses an onboard memory controller, which allows the cpu to more efficiently prepare for these lost memory accessible cycles and perform other work at this time. This increases the overall cpu efficiency and allows the use of cheaper RAM memory to overall achieve only slightly lower performance than the better, more expensive memory.
In my example with my system, my DDR400 memory is not able to reliably run faster than 225MHz (it fails at 230) but my cpu is able to run at 290MHz and my FSB can handle 320MHz. This means my memory is limiting my overall system performance. I could purchase DDR500 memory that is able to run at 250MHz but I have already spent money on the DDR400, which is rated at 200MHz. To allow the faster speeds, I set my memory to its maximum of 225MHz and then find a ratio with the CPU to allow the fastest and most efficient system speed. A higher CPU MHz speed requires a lower memory divider to allow the memory to keep up with it.