How often do you replace your hard drives?

I guess you PC must magically work differently from everybody else's then. Could you tell me more about your "benchmarks"? Were they real-world application load times or irrelevent peak throughput measurements?

I had a RAID-0 at one point and the reason why I got rid of it was that it didn't make one lick of a difference. And I did the stopwatch thing I suggested.

It must, because I *noticed* a performance difference in I/O-bound situations. But hey, what do I know. It's not like I used the PC every day for a year and a half or anything...
 
Is there still anyone crazy enough to run a Raid-0 rig?
Yep. 2xWD250JBs on ICH5R in RAID 0 for storage/edit of TS streams from HDTV tuner. Does it make a difference? Not heaps. Software RAID incurs some overhead, so things are a wash excepting for the convenience of a single bigger volume.
 
Yep. 2xWD250JBs on ICH5R in RAID 0 for storage/edit of TS streams from HDTV tuner. Does it make a difference? Not heaps. Software RAID incurs some overhead, so things are a wash excepting for the convenience of a single bigger volume.

Interestingly enough, software RAID-0 isn't neccessarily slower. The Asus P5W DH has a hardware RAID controller (Silicon Image 4723) and ends up being slower than Intel's software raid solution in all synthetic benchmarks. In stripping mode, that is... mirroring is significantly faster with the SI 4723 because mirroring it seems to cause a significantly larger CPU load than stripping, which I found surprising.
 
I keep all critical data on 1 hard drive, and keep a regular backup of it on another hard drive. The probability of both failing at the same time is miniscule (barring an outside event like a house fire). Since I don't mind losing my "less critical" data, I don't bother with over-the-top RAID solutions.
 
Back
Top