@strangestranger:
Thanks, and while I understand the math and get the principal... I'm not sure that really answers the question of which RAM is most approrpiate for a stock speed Q6600. But it simplifies some other information I've come to find out...
FWIW: I am not meaning to hijack the thread from KERMO, though since the list of components that were suggested didn't include the required RAM - I soothe my conscience in the thought that it's relevant to the main topic ;-)... If I can take a minute to flesh out my own thoughts - perhaps someone can breifly comment "yeah, sounds basically right" ... or "start a new thread to continue discussion"
.
■So I think the data rate figures you broke down just mean that even older DDR2-533 could be paired with a Q6600 at its stock 1066 MHZ FSB eh?
■That said, I've come to get the impression that doing so (as in either your example or mine) - i.e. getting the 'effective' clock ratios between CPU and memory to be 1:1 - provides some assurance of "stability" and this would be particularly desirable "if" overclocking the CPU/FSB...
■And so, "if" overclocking a CPU with a stock 1066 Mhz FSB... then RAM 'faster' than DDR2-533 (a 1:1 with 1066) operating at it's regular speed can get you back to the 1:1 ratio as you increase the FSB as you overclock...?
QUESTION: Assuming all that is basically correct... the gain in stability with a 1:1 ratio is something I'm reading is only really a 'must' in overclocking situations; is that correct?
QUESTION: If
NOT overclocking the CPU, is there performance benefit withOUT loss in stability in running 'faster' RAM resulting in a CPU to memory ratio of something other than 1:1 (like 1:2 if you went with DDR2-1066 / PC2-8500 memory) ?