With memory so cheap, lots of people (and I’m just as guilty) have been buying lots of memory for little more reason than “because it’s there.”
However, what isn’t as generally well known is that more memory usually means less FSB overclocking.
How much so?
I began to test this, and ran into a major problem.
As I observed the other day, one 256Mb stick of Crucial PC2100 managed to pass our memory tester at fast settings at 156Mhz.
I was rather dismayed to find that three sticks only managed to do 137MHz. (At least it was nice enough to blow up right away; with three sticks, the complete test takes four hours.) I then began the battery of tests with two sticks, the exact same two sticks I ran regularly at 150Mhz on an AMD760 board.
So far, I can vouch for 140MHz, and know that 145Mhz doesn’t work. This is cause for more than dismay. This makes me wonder about the motherboard and/or the chipset.
I’m going to play with this some more, but at this point, the best way for me to find out whether this is typical or not is to ask.
If you’ve been FSB overclocking, and have noted how far you could go FSB-wise with one, two, or more sticks, could you drop me a note?
Please indicate what mobo you’re using, the kind of RAM, and how far you got FSB-wise with the same settings, but different number of sticks?