In the past, I always ran the full battery of benchmarks when reviewing memory. System, graphics, and other performance benchmarks show little if any increases in results when merely changing memory clocks. It is kind of pointless to show a increase of 5 Marks in a Futuremark benchmark, or .5FPS in a game.regular benchmarks just don’t tell the whole story. Higher memory performance is just part of the entire system, when your memory performance is better, everything runs better, but isn’t necessarily shown in benchmarking. So I have narrowed my memory testing to purely memory benchmarks in SiSoft Sandra and Lavalys Everest, which do show.
I ran the benchmarks at stock clock and timings, highest overclock at stock timings, tightest timings, and highest overclock. The latter two showed decreased performance in every test, so to keep the charts simple I opted not to show them. Just keep in mind that both are available to add much flexibility to the memory’s capabilities.
Aug 18, 2014 0