clock cycles

[Dragon]

question: . . . ram timings. address strobe delays etc, etc, etc. there all measured in clock cycles(i think), so how much info does a clock cycle move in terms of the processecor, and what effect do the ram timings have on how fast you move information. i.e., what is a clock cycle specifically? and is info moved through the ram each cycle no matter what its timings are? if so what affects how much?

OR

am i completely off, if so.. how does it work?

worded that as best i could....

[[LIK]*Spiker]

I think, but not 100 % sure:

Take an intel P4 3.2 gig for instance. Every clock cycle, it moves 3.2 gigs of "stuff." Ever seen a wavelength chart for audio? Well a clock cycle looks sort of like that, so every "peak" and "ditch" (for lack of better scientific words) in the wave represents a clock cycle.

Ok I think i'm done now, 'cause I really don't think I know what I'm talkng about, but thats the extent to what I know. I hope somebody more qualified posts in this forum 'cause I'm curious too.

[Jason425]

data moved is not directly proportional to clock speed. Other factors, such as cache, bus speed, chipset or other things determine how much your chip moves.

Simply put, the lower your ram timings are, the more overall performance you'll have. go to http://www.corsairmicro.com/memory_basics/153707/ to find out more about ram

[Dragon]

that was helpful but i mean like, if something is put in ram one cycle and needed the next, will it get it thne, or have to wait for the timings, 2, then 2, then 2, then 5 - in my case?

[Jason425]

I think everything revolves around the timings.. the first number (cas latency) is the most important

[james]

this all depends on things like instruction sets, the architecture (32bit-64bit), latencies and whatnot. there's not a quick answer to the question. try looking for a white paper if you are really interested. (arstechnica.com has some good explanations of how cpu architecture's work and whatnot, i suggest maybe starting there)