There are some very impressive statistics for newer solid state hard drives such as the Corsair Force, OCZ agility 2 and OCZ vertex 2, all of which share the same controller chip. I initially tested the drive with a hard drive test program to discover excellent performance. I then found that when I actually wrote and read my ordinary data to and from the hard drive, the performance was nowhere near as fast. I decided to write 0x55 (binary 0101010101010101010101010101...) across the whole drive with badblocks. Badblocks appeared to work very fast, writing 40Gb in 2m 45s - 242Mb/Sec! I decided to try using different patterns with badblocks, then discovered the random pattern took 13m 24s to write 40Gb (49.75Mb/sec). I checked to find this was not caused by a bottleneck in the pseudo-random number generator. On further investigation, as exposed here, the drives behave completely differently with highly repetitive data characteristic of hard drive performance testing programs, compared to real-life data, In the test I present in this video, I generate 2 files. One contains just binary zeros -- highly repetitive - a lot like hard drive test programs generate. The other contains incompressible data much like avi, mp3, jpeg, zip, or iso files. As you can see here, this drive behaves totally differently. The drive "writes" highly repetitive files at about 220Mb/Sec, but writes realistic everyday files at about 50-52Mb/sec. Somewhat different to the claimed 270Mb/sec. I have not <b>...</b>
SSD
solid
state
hard
drive
performance
Corsair
Force
OCZ
agility
vertex
