Each memory kit was tested at three speeds; stock speed, rated speed, and highest overclock. For "stock speed", we used set the motherboard to AUTO, which causes the memory modules to operate at standard JEDEC DDR2-800 specifications. This was done to provide a good "ground-floor" reference point for the modules. For "rated speed", the modules were manually configured to operate at the settings specified by OCZ. Finally, for "highest overclock" we set the modules to the highest stable overclock settings we were able to achieve.
- Intel Core 2 Duo E6400 (2.13 GHz) Dual-Core
- 1 x Nvidia nForce 680i Series Motherboard
- 1 x Western Digital Raptor 10,000 RPM Serial ATA Hard Disk
Microsoft Windows XP Professional (32-bit)
- Reaper HPC PC2-8500 @ Stock (JEDEC DDR2-800 specs)
- Reaper HPC PC2-8500 @ Rated (1067MHz, 5-5-5-15)
- Reaper HPC PC2-8500 @ Highest OC (1143MHz, 5-5-5-15)
- Reaper HPC PC2-6400 EB @ Stock (JEDEC DDR2-800 specs)
- Reaper HPC PC2-6400 EB @ Rated (800MHz, 4-3-3-12)
- Reaper HPC PC2-6400 EB @ Highest OC (1067MHz, 4-5-5-15)
We began our testing with SiSoftware's SANDRA, the System ANalyzer, Diagnostic and Reporting Assistant. SANDRA consists of a set of information and diagnostic utilities that can provide a host of useful information about your hardware and operating system. We ran SANDRA's Memory Bandwidth and Latency tests.
As expected, at stock speed, both the Reaper HPC PC2-8500 and the PC2-6400 EB modules scores nearly identically in all three tests. At stock speed, the PC2-8500 performed slightly better than the PC2-6400 EB in the bandwidth tests. A similar trend is observed when both modules are overclocked. The difference in speed between the two memory kits is much more noticable in the latency benchmark where the PC2-8500's small lead increases dramatically.
While both memory kits performed significantly better at their rated speed when compared to their performance at stock speed, overclocking didn't produce a significant performance gain. In fact, the PC2-6400 EB's performance actually took a hit in the SANDRA integer bandwidth test when it was overclocked. However, the performance difference between rated and overclock is within benchmark error margins. These counter intuitive results may be due to a number of factors but overall they illistrate that the system doesn't necessarily benefit from very high memory speeds when the front side bus speed isn't increased.