Sequential Read Performance
Our sequential tests are conducted in the same manner as our random IO tests. Each queue depth is tested for three minutes without any idle time in between the tests and the IOs are 4K aligned similar to what you would experience in a typical desktop OS.
In sequential read performance the difference between drives is rather marginal, but in power consumption we start to see some differences. At 250GB and 120GB the 850 EVO is very efficient, but the 500GB and 1TB are again more power hungry.
Looking at performance across all queue depths doesn't reveal any surprises: at QD2 and higher all drives are practically saturating the SATA 6Gbps interface. What's notable, though, is that the 1TB degrades in performance as the queue depth increases. I wonder if this is a thermal issue (mSATA/M.2 don't have a chassis to use as a heatsink) or just poor firmware optimization.
 | |||||||||
Sequential Write Performance
Sequential write testing differs from random testing in the sense that the LBA span is not limited. That's because sequential IOs don't fragment the drive, so the performance will be at its peak regardless.
In sequential write speed the capacity plays a more significant role because the 120GB and 250GB are noticeably behind the 500GB and other higher capacity drives. The poor performance of the 1TB model is once again a surprise, though.
The full graph of all queue depths shows the reason for the 1TB's low performance: it starts high at ~430MB/s, but after that the performance decreases.
 | |||||||||
ncG1vNJzZmivp6x7orrAp5utnZOde6S7zGiqoaenZIZxfpJoq6GdXaiurr%2FUp55mq6OZenmBj2acr6ddosCiwMBmpGtloprDqrHWaG4%3D