cbjaust
August 27th, 2007, 12:50 AM
In your guide TTR's Guide to Determining Optimum Overclocked System Performance Points (http://www.thetechrepository.com/showthread.php?t=195)you state that the calculation for CAS latency is CAS x 2000 / Freq MEM (Data Rate MHz) .
My question is why do you use the DATA rate as the frequency in the calculation and then also use a constant of 2000 to make the calculation come out right?
Would it not be simpler to make it like this?: CAS latency (ns) = CAS / Freq MEM (Base Frequency in MHz) and then convert to your favourite units at the end?
eg DDR2-1000, CAS 4: CAS latency = 4 / 5x10^8 = 8x10^-9 s (x 1x10^9) = 8.0ns
Also, does this mean that CAS2 PC3200 has a CAS latency of 10ns ?
Thanks
PS your articles are very informative and any articles pertaining to how all of the timings and (A64) subtimings relate would be very welcome.
My question is why do you use the DATA rate as the frequency in the calculation and then also use a constant of 2000 to make the calculation come out right?
Would it not be simpler to make it like this?: CAS latency (ns) = CAS / Freq MEM (Base Frequency in MHz) and then convert to your favourite units at the end?
eg DDR2-1000, CAS 4: CAS latency = 4 / 5x10^8 = 8x10^-9 s (x 1x10^9) = 8.0ns
Also, does this mean that CAS2 PC3200 has a CAS latency of 10ns ?
Thanks
PS your articles are very informative and any articles pertaining to how all of the timings and (A64) subtimings relate would be very welcome.