Tom's Hardware > All Reviews > Storage > External Storage > In Search of True DDR2 Bleeding Edge Memory

More Speed By Tightening Latency Timings

03:00 - Monday 27 February 2006 by Tom's Guide Team

We conducted the following latency timing test with an FSB speed of 1066 (266 MHz) using the premium Corsair CM2X512-8000UL memory modules.

An overview of speed improvements as they relate to latency timings:

Module Latency Speed (CL X.0-3-3-8) FSB 1066
Latency	Read	Write	Latency
CL 3.0	7458 MB/sec	2499 MB/sec	76.9 ns
CL 4.0	7435 MB/sec	2447 MB/sec	80.1 ns
CL 5.0	7405 MB/sec	2375 MB/sec	81.9 ns
Latency	Read	Write	Latency
CL 3.0	100.7 %	105.2 %	106.5 %
CL 4.0	100.4 %	103.0 %	102.2 %
CL 5.0	100 %	100%	100%

Even at a high FSB speed of 1066, boosting CAS latency to 3.0 barely improves memory read speeds, but memory write speeds improve on the order of 5%.

The speed improvement that results from tightening latency timings at an FSB speed of 800 (200 MHz) is likewise only minimal. Thus, it makes no sense to leave the FSB speed unchanged, and only boost memory timings. This strategy does not offer enough gains to justify purchasing high performance, but more expensive, low-latency RAM.