Performance in Intel IOMeter
Sequential Read & Write Patterns
IOMeter is sending a stream of read and write requests with a request queue depth of 4. The size of the requested data block is changed each minute, so that we could see the dependence of the drive’s sequential read/write speed on the size of the data block. This test is indicative of the maximum speed the drive can achieve.
The numeric data can be viewed in tables. We’ll discuss graphs and diagrams.
The newcomers are competitive to their opponents at sequential reading. For example, the Hitachi is average if compared with the others whereas the Samsung is slow on small data chunks and fast on large data chunks. It achieves its top speed on larger data blocks than the other drives, though. The Toshiba remains the leader in reading small data blocks – its ability to glue multiple requests into a single large request puts it beyond any competition.
Interestingly, the drives from Hitachi and Western Digital both suffer a performance slump on 4KB data chunks. This is their response to the specific synthetic load when the controller is installed into a PCI-X slot. We will not observe this kind of a slump in the other tests, yet the other HDDs do not have it at all.
The standings are totally different at sequential writing: the Hitachi is now competing with the Western Digital for first place while the Samsung shows a very low top speed for such dense platters as it has. This time around it is the HDD from Western Digital that shows its ability to effectively glue multiple small requests into a large one.
Disk Response Time
In this test IOMeter is sending a stream of requests to read and write 512-byte data blocks with a request queue of 1 for 10 minutes. The total number of requests processed by the HDD is over 60 thousand, so we get a sustained response time that doesn’t depend on the HDD’s buffer size.
The HDD from Western Digital is unrivalled at both reading and writing. It must have a very fast mechanism of the read/write heads backed up by efficient firmware algorithms. The Hitachi is a disappointment: its response time at reading is too high. Even platters with lower recording density were better in this respect.