Benchmarks are more prevalent in SSD testing than many other components, because comparisons between SSDs in real-world usage are difficult, if not impossible in many applications. For instance, when timing most program load times by hand, you might have an error of several tenths of a second, and the actual difference in the load times might be less than a tenth of a second. What this means is that comparisons become somewhat artificial and results are semi-random based upon how you each case happens to be timed. These comparisons are inaccurate without writing special software to execute and time the tests, and, in many cases, can be simply wrong. However, certain program suites still push SSDs and may still be used for comparison between drives. When it's all said and done, benchmarks are still the best way to determine a quantifiable difference between SSDs, and PCMark7 provides a great solution as an objective measure of real-life performance.
We ran three different synthetic benchmarks: CrystalDiskMark, ATTO, and AS SSD. There are four major attributes to SSD performance, which are sequential write and read speeds and random write and read speeds. Synthetic benchmarks often test an array of different file sizes in its tests, and many reviewers will make sure they report the 4K file size. Since most applications and normal OS usage will send 4K transfers to and from the SSD, the 4K file size provides a good measure of performance during normal usage. Different benchmarks also show differences in performance with compressible or incompressible data transfers.
There is a certain amount of variability between each run of the benchmarks, so we ran them with larger sizes to allow the speeds to average over a longer period of time. We ran four runs of each benchmark and averaged the numbers to further offset variability.
CPU: Intel i5-3770K
Motherboard: Gigabyte Z77X-UD3H
RAM: 2x4GB Patriot Viper Xtreme II 1600MHz DDR3
OS SSD: Patriot Pyro SE 60GB
Secondary SSD: Kingston SSDNow V200 128GB w/120506a Firmware