While in the previous post I focused on generic network access to my NAS over a gigabit network, in this post I will detail the results I got using the hard-disk utility HDTune to measure the performance of my four local disks. The results are pretty stunning, and they highlight the difference in transfer rate and access time between conventional HDD and new solid state drives (SSD).
|Model||Min transfer rate MB/s||Max transfer rate MB/s||Avg transfer rate MB/s||Access time ms||Burst rate||CPU usage %|
|Crucial M4 64GB||289.7||349.4||336.3||0.1||112.6||12.8|
|WD Caviar Green 2TB 5400 RPM||54.2||117.8||89.2||13.1||150.7||6.2|
|WD Caviar 500GB 7200 RPM||38.9||83.6||67.9||20.3||118.8||5.4|
|Samsung HM320JI 320GB 5400 RPM||32.5||70.5||54.2||18.4||80.7||4.3|
The Crucial M4 SSD (plugged on a SATA III port) performs very well and confirms my expectations from this kind of device. Of course I did not need such a performance in a NAS, but I wanted in my build for three reasons which are its low power consumption, it's coolness and it quietness (due to the absence of moving parts).
And having an SSD gives me a brand new sensation: I'm no longer waiting for my computer to boot, but it is waiting for me, and that's the way it should be. Even the Velociraptor I have on another station looks so sluggish to me today!
|Crucial M4 64 GB tested with HD TUNE 2.55|
As you can see in the following picture, my WD Caviar Green is a little bit bottlenecked by the SATA II controller, and you can see it from the beginning of the perf test (highlighted in red) being a straight line and not going beyond 120 MB/s. Also this 32 bit version of HD Tune cannot test more than 1 TB of disk, that's why the yellow dots stop half-way in the graph (where I put a yellow line): this means that the access time is probably higher than 13.1 ms, but this version of HD Tune cannot test it.
|WD Caviar Green 2 TB performance|
|Western Digital Caviar 500 GB performance|
|Samsung 320 GB performance|
In a forthcoming post I will put the results I get with AS SSD and maybe other bench tools. Comments are welcome!