Testing Continued - IO Meter READ Performance

     Once IO Meter prepared the drives for testing, the access specification was setup with a 5 megabyte 512 kilobyte test pattern. That pattern was used for all of the following tests on each of the drives. Data was collected after the test had been allowed to run for 15 minutes.

     When the read pattern changes from sequential to random, hard drives seek heads must be reposition several times during the read in order to collect all of the data. The above charts show the decrease in data read rate from sequential to random, except with the OCZ SSD. With no moving parts, the SSD can logically retrieve all of the requested data without mechanical latency.

     Look at what happens to the response times when require the data to be access randomly. The seek times of the Toshiba drive were clearly the longest at a little over 4 full seconds (4000 milliseconds = 1 second).  Without the handicap of moving parts, the OCZ SSD was able to respond in a mere 96 milliseconds. Just to put that into perspective for you, it takes about 300-400 milliseconds to blink your eyes.

IO Meter WRITE Performance

     The proverbial chink in the armor for the SSD would be its general writing performance. In a sequential write, the solid state drive lagged far behind even the Toshiba notebook drive. Once the pattern was shifted to a random order the SSD closed to within four megabytes per second of the Toshiba drive.

Page 1:  Intro Page 2:  Up Close Page 3:  Testing (Part 1) Page 4:  Testing (Part 2) Page 5:  Testing (Part 3) Page 6:  Conclusion