Detecting Drive Speeds
SATA Hard Drives are supposed to be much better than IDE (PATA) drives but how fast are they really? I decided to find out with hdparm. I have three hard drives on my desktop and interestingly enough they are three different types (one is a SATA II dive, while the second is SATA - I and the third happens to be an old PATA drive that I have had with me for quite a long time.)
As you can see from the table below there is a difference in speeds between the three different drives. But that difference is small and not worth writing home about. The 3.0 gbps bandwidth on the SATA II bus is completely wasted while the PATA drive is only marginally faster than a high speed USB drive.
I have another set of three drives on my media/file server. All three happen to be 250GB Hitach SATA II drives. Two of them have been combined together to create a stripe set using linux software RAID. Their speeds as reported by hdparm -tT are as follows:
|Soft RAID||HDP725025GLA380 X 2||Stripe||500GB||988.17||168.42|
These drives apear to be slightly faster than the drives on the desktop. While two off them had nearly identical speeds the third was about 20mbps slower than the other two! I suspect it might be caused by a noisy cable. A RAID device can also be tested with hdparm and I was pleasently suprised to find that it’s speed really is the sum of the speeds of the devices that are used to build it. I expected it to be much lower than the sum.
From these results, I can only conclude that a bigger drive is faster than a slower drive - wheter it’s SATA or PATA is completely irrelevent.
It’s important to note that the figure reported by hdparm is the highest possible speed for your disk. For day to day use with Random reads it’s going to be a lot slower. If those random reads are mixed with write operations, it’s going to be slower still.