Luminous Landscape Forum

Equipment & Techniques => Computers & Peripherals => Topic started by: barry685 on August 01, 2011, 05:54:30 am

Title: Hard drive performance
Post by: barry685 on August 01, 2011, 05:54:30 am

Which will likely perform better? 2 Samsung spinpoint F3 drives in a Raid 0 configuration, or a Wd Velociraptor 450gb 1000k rpm drive?

Title: Re: Hard drive performance
Post by: nemophoto on August 02, 2011, 03:55:27 pm

From my limited experience, I believe your best bet is still a 10K single drive. I use to use a RAID 0, however, problems I encountered, which I believe had to do with a bad sector (which entailed me having to go all the way back to DOS to work around crashes to get back to a single drive) has made me gun-shy of RAIDS. I found when I benchmarked my RAID, while some of the the throughput was indeed fast, the jumping back and forth between drives actually caused fairly significant "valleys" in the performance.

My 2-cents.

Nemo

Title: Re: Hard drive performance
Post by: John.Murray on August 03, 2011, 12:10:38 pm

Assuming you have a SATA 6GBs controller, definately the WD

Title: Re: Hard drive performance
Post by: DeanChriss on August 04, 2011, 12:51:12 pm

You might want to check thePuget Systems article at
http://www.pugetsystems.com/labs/articles/SATA-Controller-Performance-Explored-87

They found that "no traditional platter SATA hard drives can actually run fast enough to benefit from a 6Gbps interface", including the 10K RPM Velociraptor. The conclusion is that unless you're using SSD, it doesn't matter whether you plug your SATA III drive into a SATA III or SATA II interface.

Title: Re: Hard drive performance
Post by: John.Murray on August 04, 2011, 03:27:12 pm

Interesting Article!  Thanks.  I noticed they used Crystal Disk Mark  (http://crystalmark.info/software/CrystalDiskMark/index-e.html) for the tests, I'm not familiar with it, preferring instead the tried and true Iometer (http://www.iometer.org/).  That said, I'll definately check it out, as Iometer is not fun to use....  One thing that jumps out at me is no indication of the size of the data-sets used for testing.  My experience has been a noticeable increase in throughput, particularly when multiple drives are being used.