2.5" HDD Galore: Samsung, Seagate, Toshiba


Feb 13, 2006
i have to ask.. why weren't these drives tested in notebooks? kinda silly to test them inside of a server. although it is pretty sad to see my hard drive (seagate 7200.1 100gb) look very low in the benchmarks. you would think 7200 rpms would make a huge difference. it only overheats inside the laptop. i'll be switching to a 5400 rpm hard drive eventually. you guys should have done a temperature check also!


Apr 5, 2006
I was thinking the same thing - 7200rpm drives should be quicker. I like the 7200rpm 200GB drives, but a 500GB 5400rpm drive would make more sense for some. For those few rich guys out there with a 3-drive quad-core notebook, I can imagine the impact a RAID0 1.5TB hard drive setup would provide.

Another issue I (and others also may) have is I'm still using an older notebook with PATA interface and a 128GB limitation in the BIOS. I've been looking to upgrade to the faster 7200rpm and max out my capacity, but the only 7200rpm PATA drives out there come as either 100GB and 160GB. For the sake of upgrading, I would desperately like to see a 120GB 7200rpm PATA drive in the market.
the 100 gig drive(seagate 7200.1) is most likely not a PMR (perpendicular drives get the bits closer together by standing them up. this increases size and speed) drive. As such its still slower.


Why test in a server? hmmm because its tons faster....plug test...plug test....plug test....not open laptop....install drive....close......test....

7200rpms are not as effective on such a small drive...but there is a improvement on seek times....its like asking why some new drives can match and beat a raptor for speed over the first 150gigs.....the raptor will still win with its lower access times....The closer you can get the data, the faster the drive...

Temps in some kind of box would have been nice...but there is always next time.

All the other drives are just for reference anyway....


Feb 2, 2004
Old article I know but still a great resource for people looking at power consumption and speed differences. Wanted to help people understand hard drive performance a bit.

Hard drive performance isn't a flat number. It falls into overlapping categories of read, write, random (bits of information in different places) and sequential (large blocks of information in one spot, then move on)

Drive performance factors are seek time and data rate.

More data density on the same spin speed (i.e. bigger drive, 100GB vs 250GB) means the data rate goes up. That is why a 5400rpm drive with a high data density can "beat" a 7200rpm drive with a lower one. HOWEVER, that only effects SEQUENTIAL read and write.

To increase random read/write performance, you increase the RPM of the drive. This is because in random read/write, the drive spends more of its time seeking new locations than doing things at those locations. (11ms to seek, 1ms or less to write... big ratio!)

90% of the time, when you care about drive performance, you will be experiencing RANDOM read and write. The reason for this is that most things fall into these random read/write performance buckets:
* reading from one file, processing, and writing to another. (like copying files)
* reading from multiple small files
* searching through one large file, reading bits of it at a time.
* etc.

You experience sequential read/write performance under the following situations:
* Copying large files from one DEFRAGGED disk onto another defragged disk
* Creating a large amount of information using a small amount of information or information that is not stored on the disk (converting compressed video to uncompressed, rendering video from a small scene file, recording your screen to disk, recording video or audio.)
* Streaming something from the disk for display.

Note that some of those situations will not necessarily run faster with a faster disk. Can't record video or audio faster than it is coming into your camera or mic. ; ) Large file copies are a bigger deal but I don't frequently copy more than 2-3 gigs at a time. So a 5400 disk with high data density can win benchmarks, but it will lose in RL use most of the time because sequential performance is actually quite difficult to get in RL.