Hard Drive Benchmarking from Linux

Wednesday, June 06, 2007

If you trying to find out a similar application likes HDBench in Linux, perhaps you would like consider to use hdparm. This small application basically figuring out the harddisk parameters. Combined with several options, it could be a powerful toy to stressing your hard drive. For instant, see below default structure command:

# hdparm [ flags ] [device]

As written on it’s manual, to start benchmarking process you can simply set the flags parameter with –Tt. While the device parameter replaced with your hard drive assigned to device files (/dev/hdx). Watch for the example below:

# hdparm –Tt /dev/hda

The command above will resulted timing cache reads & timing buffered disk reads. Both on MB/sec unit. So, what is –Tt actually means? The first “T” parameter perform timings of cache reads for benchmark and comparison purposes. For meaningful results, this operation should be repeated 2-3 times on an otherwise inactive system (no other active processes) with at least a couple of megabytes of free memory. This displays the speed of reading directly from the Linux buffer cache without disk access. This measurement is essentially an indication of the throughput of the processor, cache & memory of the system under test.

The “t” parameter will perform timings of device reads for benchmark & comparison purposes. Again, for meaningful results, this operation should be repeated 2-3 times on an otherwise inactive system (no other active processes) with at least a couple of megabytes of free memory. This displays the speed of reading through the buffer cache to the disk without any prior caching of data. This measurement is an indication of how fast the drive can sustain sequential data reads under Linux, without any file system overhead. To ensure accurate measurements, the buffer cache is flushed during the processing of -t using the BLKFLSBUF ioctl.

If both flag is already specified, then a correction factor based on the outcome of -T will be incorporated into the result reported for the later -t operation. See the result test over the same Linux Fedora Core 4 with my very old stinky laptop Pentium III 600 Mhz, 128MB SDRAM, 20GB Seagate 5400 RPM hard drive: