Performance in Intel IOMeter: Sequential Read & Write

We will first measure sequential read and write speeds in IOMeter which can show the influence of the interface on the drive performance. So, the drive is processing a stream of read/write requests with a request queue depth of 4; the data block size was changed every minute. As a result, we can see the dependence of the drive’s sequential read/write speed on the size of the processed data block. Then we can compare these data for each interface, too.

The sequential read diagram shows the considerably higher efficiency of the FireWire interface. The external drives are both much faster when using it. Without doubt, this interface allows the drives to show more of their speed potential than the other interface does. You can also see that the linear read speed of the Maxtor drive is higher than its opponent’s at any data block size. The results the drives have working via USB 2.0 do not differ much as their performance is limited by the lower bandwidth of this interface. Such factors as the cache buffer size may account for the variations in performance of the two devices in this case. We’ll check this out below in other tests.

Displaying the linear write speeds of the drives, the second diagram has much in common with the previous one. Of course, the speeds are lower here, yet we again see that the FireWire interface allows achieving the highest performance with either drive, but especially with the Maxtor. The difference between FireWire and USB 2.0 is less conspicuous with the Seagate which may be an indication that its write speed is in fact lower than the opponent’s. You can also see again that the USB 2.0 graphs of both the drives don’t differ much because the insufficient interface bandwidth levels their performance.