RAID 0 on Abit AT7

[smitbret]

I have 2 20gb Seagate Barracuda III's that I have set up in RAID 0 via the onboard High Point controller on an Abit AT7, IDE connectors 5&6. The set up works, but when I benchmark it on HD Tach I am only getting an average of about 36000 mb/s transfer, almost identical my single Maxtor ATA 133. When I benchmark the file system with SiSoft Sandra it's coming in at about 18000. I am running my system and program files off of the array. Shouldn't I be seeing something quite a bit higher? What am I missing here?

 

[HammerBot]

The baracude III isnt the worlds fastests. Try testing the drives in stand-alone. Whatever score you get, you can expect an performance increase of max 70%, with the best combination of stripe/clustersize and drive.
What stripe/clustersize do you use. What file system?

<i><b>Engineering is the fine art of making what you want from things you can get</b></i>
<A HREF="http://www.btvillarin.com/phpBB/viewtopic.php?t=655" target="_new">My systems</A>

 

[smitbret]

Stripe size is 32k. Is that too big to be running the system and program files on? I'm running Windows Xp Pro.

 

[HammerBot]

There isnt a stripe/clustersize combination that fits all usages and drives/controllers. But it may have a significant influence on the performance. If you want to get the most out of your raid0 array, I suggest you make some benchmarks with various settings.
If you have a third drive you can install temporary, an easy way to go about benchmarking RAID is described in the RAID FAQ

<i><b>Engineering is the fine art of making what you want from things you can get</b></i>
<A HREF="http://www.btvillarin.com/phpBB/viewtopic.php?t=655" target="_new">My systems</A>

 

[Ncogneto]

HDD tach and Sis sandra are noterious for giving erroneous results when dealing with Raid Arrays.. both are pretty much useless for anything other than single drive tests. Give winbench a go and try the disk inspection test and report back.

It's not what they tell you, its what they don't tell you!<P ID="edit"><FONT SIZE=-1><EM>Edited by ncogneto on 11/11/02 00:17 AM.</EM></FONT></P>

 

[smitbret]

Winbench didn't give an average read speed like SiSoft and HD Tach do. Instead it gives beginning and end figures. Are these averages? I hope so, cuz they're pretty good. Thanks.

 

[Ncogneto]

Winbench performs like HDD tach pretends too. From the begining of the drive to the end testing all areas...you should see a graph while the test is taking place. Meanwhile HDD tach only takes random samples across the whole disk and extrapolates its curve accordingly. What were your scores?

It's not what they tell you, its what they don't tell you!

 

[smitbret]

Scored better than I expected, although I'd probably set my sights pretty low after running SiSoft and HD tach. Beginning # was about 82000 and the finishing number was at 51000. Works for me, since I was just looking for something in the 50+ range. Pretty steady graph all of the way down, not all over the place like the HD Tach graph. Anyway, I'm getting what I need/want out of it now. Are those scores reasonable or should I expect more (or less)? Thanks.
-Brett