[hikaricloud]

So I was skimming around (I'm pretty sure this hasn't been posted yet), and came across this interesting article about the speeds of Seagate's 1.5tb HDD in comparison to the 300gb Velociraptor. For the first test, they set the size of the 1.5tb down to 300gb to pit it against the Velociraptor to see how it fared...and well, see for yourself.







More benchmarks and such at the source: Seagate Barracuda 1.5 Tb MOD | TechwareLabs

[Yamikotai]

That's a fair bit better. I'd still much prefer an SSD though if I were looking for a performance drive.

[hikaricloud]

Well obviously, but the prices are still a decent bit high. Just thought it was an interesting comparison.

[Puddle Jumper]

In my opinion that article seems somewhat meaningless since they only use 300gb of a 1.5tb drive, essentially ensuring a best case scenario were all data is located on the most favorable part of the drive, making the results very different from anything you would see in the real world. The author also tries to downplay the results of the 4kb random write when in fact you write 4kb files when you are using your system far more often that you write 500mb ones.

[hikaricloud]

They also did an 819gb test.

I can personally attest to the 1.5tb being ridiculously fast, and more than a match for a velociraptor, a drive specifically marketed for its speed. The idea of the article was for more than 100 bucks less, you could get a similarly performing drive with 5 times the space. In that essence, the article is a great success.

[zedman3d]

I was just about to post this.
Pretty awesome.

[zmatt]

Originally Posted by hikaricloud

They also did an 819gb test.

I can personally attest to the 1.5tb being ridiculously fast, and more than a match for a velociraptor, a drive specifically marketed for its speed. The idea of the article was for more than 100 bucks less, you could get a similarly performing drive with 5 times the space. In that essence, the article is a great success.

It's not faster. the 4kb test is the most important test. they rigged it and downplayed it to make it look like the Velociraptor lost. Why? I don't know, but they did. Your pc makes many many many more 4kb reads and writes than it makes massive 500meg operations. The large file tests are mostly useless because you rarely move a large group at once. Normally you accumulate the files overtime. The most important numbers in storage are the latency and the random read/write speed. They determine the frequency and length of those system hiccups you get when opening the file browser and the general slow feeling you get whenever you mess with your files.

[hikaricloud]

It is really interesting to see what your money can buy at the moment, and just about how pointless the velociraptors really are for the money.

Though I will say if you were going to spend 230 bucks on a velociraptor, you may as well get an SSD, which essentially makes this story worthless in that sense, but it's interesting nonetheless.

zmatt: You obviously have never had to do multiple large scale backups. It was a nearly constant thing with both shops I worked for.

[Puddle Jumper]

Originally Posted by hikaricloud

They also did an 819gb test.

I can personally attest to the 1.5tb being ridiculously fast, and more than a match for a velociraptor, a drive specifically marketed for its speed. The idea of the article was for more than 100 bucks less, you could get a similarly performing drive with 5 times the space. In that essence, the article is a great success.

They did part of the test at 819gb but it was pretty short in comparison and is still for from a real world situation.

The 1.5tb may have decent bandwidth but the problem with it and conventional hard drives in general is random writes of small files which is a very common operation on a system. Of all the tests they conducted the best 4kb random write speed the seagate achieved even when the drive was artificially limited to 300gb was 1.966mb/s, and that's a best case scenario. While that sounds pretty slow already it seems even worse when you look at the the 4kb random write speed of a Intel X25-M in a worst case scenario, 23mb/s, which is a order of magnitude faster. Those numbers are also for the first gen version of the Intel, a faster and cheaper second gen X25-M was released earlier today.

Also that article didn't prove a drive with 5x the capacity beat the velociraptor, it proved a drive with 1x or 2.73x the capacity beat the velocirraptor.

I completely agree the velociraptor is pointless for the money but for different reasons.

[hikaricloud]

Sorry, I guess I just didn't realize that buying a drive based on a .001s faster 4k write time was a smart move when just about everything else is faster/close to the same performance.

If it was an SSD, that would be a bit of a different story, considering the speed of 4k would effectively be .000173s instead of .002 or .001, and that is a decent difference in comparison.