I have been wondering what the relationship between RAM timing and frequency is for some time, and I cannot find any answers.
I've heard both referred to as the "speed" of your memory, but to my knowledge, if you increase frequency, the timings increase, and by reducing the frequency, timings decrease. I thought that a high frequency was preferable, but I've noticed enthusiasts using super low frequencies to achieve lower timings.
Also, if you buy DDR3 1600 memory, will it perform as well as say DDR3 1066 memory at these lower latencies?
Can anybody elaborate on this topic? It would be helpful if you could direct me to an article about it, if one exists. If not, one should be made, as RAM seems to have common misconceptions.
Well, to be honest, there isn't much of a noticeable difference between 1333MHz 9-9-9-24 and 1866MHz 7-8-7-24. Lower timings do make more of an impact than memory frequency IIRC.
Buying DDR3 1600MHz memory at higher latencies will probably average around the same as DDR3 1066MHz with lower latencies.
I'm running my Ripjaws which are meant to be at 2000MHz, 9-9-9-27-2T at 1340MHz, 7-7-7-20-1T, and I can't really say I can be bothered to push the speed up to 2000MHz, mainly due to the fact that I think CL7-1T will probably be more beneficial than that extra 660Mhz at CL9-2T.
Yes, increasing frequency at a certain point would require some increase to timings as the ram modules simply couldn't cope up. The increase in timings has a lower negative effect than increasing the frequency so you'd most likely increase the frequency anyway.
Some may be lowering their ram frequencies just to sync it with the FSB.
From the image you see that it is faster to have CAS5 800MHz modules than CAS4 533MHz modules. It's harder for manufacturers to lower latency than increasing clock speeds, so the CAS4 533MHz module might end up costlier than the CAS5 800MHz module.