INTRODUCTION
I spent long hours up last night finding max frequency at each standard CAS Latency.
I was freezing my behind off in front of an Air Conditioner to keep
case cool. I drank loads of scotch and coffee to stay awake.
I also corrupted every second file in windows, RAM has a tendancy to do this
when it is undervolted for the speed you are trying to boot at.
The reason for this testing Frequency VS Cas Latency, Results found
in bandwidth (read, write and copy) and also in Latency (ns)

I was also testing how far I could take 1t Command Rate.
Interestingly 1t had little effect on total RAM OverClock for CAS3,
absolutely no effect on CAS4 (maybe a few Mhz but I did not want to change FSB).
However it did effect how far I was able to push CAS5.

TEST METHOD
All testing done with AC blowing into side of case (freezing cold) and vMEM
at 2.425v BIOS for about 2.47 REAL volts.
I used Everest and SPI (XS mod) to test results. One thing should be noted,
regardless of whether I had Max speed at CAS3, CAS4 or CAS5, all of my
times on SuperPi were nearly identical, No more than 0.05 second's of
difference, so I did not keep results. SuperPi1m times were all around 14.141s.
I Did not have any fancy hardware to use, so I just used my rig.
CPU: C2D E6600 @ 3.6ghz (1604 FSB)
Mobo: Asus P5N32-E (680i)
Ram: OCZ Sli ready PC8500 CL5 (DDR 1066)
GFX: 8800GTX
PSU: Thermaltake TP 850w

Each test was run twice, I took the screens of second run.


Now onto the results.


Tight Timing's
Max Frequency @ CAS3 and CR1t


Max Frequency @ CAS3 and CR2t



Medium Timing's
Max Frequency @ CAS4 and CR1t


Max Frequency @ CAS4 and CR2t



loose Timing's
Max Frequency @ CAS5 and CR1t (the 1T CR limited me badly here.)


Max Frequency @ CAS5 and CR2t



CONCLUSION
It is easy to sit down and do the mathamatics to work out
which combinations of timings vs frequency will benifit you the most,
but there is very little I was able to find where testing was performed to
back the theory up.

Here is a field example for anyone who can not decide whether they want
to use stock timing's and high frequency, or whether they want to set up
tighter timing's.

It was a BIG surprise to find that CAS5 max frequency had such a low latency but CAS3
max frequency had higher latency but more bandwidth. Quite the opposite of what I expected.

So in the end which is the best? Well you can decide, here are my results.