Samsung’s 960 Pro M.2 NVME SSD is lust-worthy:
In Ars Technica’s benchmarks, the 512 GB model clocked in at over 3500 megabytes per second sequential read and 2000 megabytes per second sequential write. Those are jaw-dropping performance numbers.
What I find really interesting however, is that the 960 Pro does not perform much better than the previous generation 850 Pro SATA SSD in PCMark 7 and 8 real-world benchmarks. (Random IOPS performance of the drive is also really good.)
The implication is that most normal users will not subjectively experience much difference in speed between the 960 Pro and the older, much cheaper, model.
Does this mean that we have reached some sort of theoretical limit where increases in drive speed simply do not impact normal computing activities anymore (some sort of performance plateau), or are there changes conceivable to the rest of our hardware and software to increase the effect of drive speedups?