Originally Posted by paulrbeers
Yes, but as I stated then the difference between bandwidth and real world difference (i.e. how much faster does a video encode, frames per second in a game, etc) is negligible. I proved that by providing my own benchmarks. I ran benchmarks on my brand new quad core mini with 4GB of RAM at 1600mhz vs 16GB of 1333mhz and the 1333mhz RAM was almost as fast as the 1600mhz. Video benchmarks were a hair faster, as was the memory bandwidth (obviously), but for handbrake the difference was a second or two (well within a magin of error). So again, the overall system speed is barely affected by faster RAM.
Yes, this makes perfect sense if you know how software typically accesses memory. First of all, software tends to access the same memory over and over, so that data gets cached in the processor chip itself (processors have several megabytes of cache these days) and thus the speed of your RAM is irrelevant. Second, software tends to access small amounts of memory in a basically random fashion, which stresses the *latency* of RAM chips and not *bandwidth*. It's entirely possible, if not common, for "faster" RAM (1600MHz) to have the same or even higher latency than slower RAM.
So basically, yes, it should not surprise anybody that "faster" RAM makes little to no difference in performance in almost all typical use cases.
It amuses me that people burn so many calories trying to figure out what RAM is fastest and what SSDs are fastest. The speed of SSDs is similarly usually irrelevant since when people copy large files to/from SSDs it's usually from much slower hard drives. No point in being able to write 500 MB/s to an SSD when you can only read it at 90 MB/s from your hard drive.