A very significant review for SSD enthusiasts, it's great to see TT is looking at I/O consistency and latency. The one number access time results most benchmarks supply does not begin to show how a SSD performs with different workloads.
While the very high queue depth workloads used for enterprise SSD testing are not typical for consumer usage, the inconsistent latency that these tests reveal may explain some of the seemingly strange benchmark results some SSD users get.
This type of testing also shows that testing SSD performance is not simple, and one simple benchmark test does not begin to indicate the overall performance of any SSD. It also shows that declaring one SSD as "the best" based on a few general benchmarks is incorrect, and very few if any SSDs are superior to others in every aspect of performance.