||Intel Core i7-3930K
||Asus Radeon HD 5550
||GeIL Evo Corsa GOC316GB2400C11QC
||OCZ Vertex 2 60Go
||Seasonic X-Series 750 Watts
We were thinking a lot about how to test memory in a way that one can come up with a qualitative conclusion rather than a quantitative one. Our goal is to be able to comment on the memory kits quality. Furthermore, standard performance tests with memory do not have any significance at all. This is because almost the only thing you're measuring is the error in measurement.
What we do instead of performance tests are "scaling test". First we start with
the standard voltage and increase the frequency until the system becomes
unstable. To test the stability we use Memtest86+ v4.2.
Using this method one can do statements on the modules potential and on the quality of the chips used. Generally you can say the lower the voltage needed for high overclocking results the better the chips.
Discuss this article in the forums