CAS latency on memory?

[platerra07]

i am trying to figure out what the CAS latency on memory means, i am having trouble understanding what it means when it says
Cas Latency: 2-3-3-6 T1
i am looking at corsair memory and some read
Cas Latency: 2-3-3-6 T1
and others read
Cas Latency: 3-3-3-8
and also
Cas Latency: 3-4-4-8 1T

what does it mean and which one is better?

these are the two i am looking at so far:

http://www.newegg.com/app/ViewProductDesc.asp?description=20-145-492&depa=1

and

http://www.newegg.com/app/ViewProductDesc.asp?description=20-145-449&depa=1

 

[The Merlin]

Go for the second one, it has better good timings

 

[bla!!]

Try searching the forum for your answer first, but here's a link from a different post that eplians quit a bit about memory.

http://www.philrees.co.uk/articles/memory_specs.htm


It goes over what DDR, SDRAM and all that junk is, scroll down to the latency section if that's all you're concerned about.

 

[YourPrinter]

Cas Latency: 2-3-3-6 T1 <---- lower the faster

 

[Demalii]

This website will clear everything up for you, it describes exactly what ram timings are and how they work.

http://www.dugu9tweaks.net/articles/ramtimings/

 

[zerozero]

As everybody says the lower the better but here is a layman's explanation about those numbers.

Imagine your memory bank in a grid with rows and columns. Each intersection point representing a location or address in your memory chip.
The computer will find those address by enabling a row and a column and where they meet is the addressed seeked by the cpu.
Depending on the quality of the RAM, it will take a certain number of clock pulses to address the proper row or column. CAS (column address strobe) is a signal expressed in number of clock pulses to find the proper column. RAS (row address strobe) is the same thing but for row.
CAS to RAS or latency time is the number of pulses required by your RAM to pass from a column to a row signal.
It is easy to see that the lower number of pulses required means faster RAM access time therefore better performance.

 

[platerra07]

thanks, that explanation was nice and easy