Does this seem right for 4x74GB Raptors in RAID 0?

[Revivalist]

http://img464.imageshack.us/img464/1152/hdtachandattoon4x74gbraptors1f.jpg

I'm not sure why someone with 2x74GB Raptors on VIA was able to beat my 4x74GB Raptors on NForce4 when it comes to read speed in HD Tach. But the difference comes in clearly with the burst speed. Could it be that their drive was less full or they had defragged it better than me? :eh?:

I know my ATTO score is where it's supposed to be because I've compared it with others that also have 4x74GB Raptors in RAID 0. . . .

Overall, most people probably wouldn't say that the cost is worth the performance, but I definitely notice the real world speed when I transfer files. Whether 2x74GB would give the same real-world feel, I don't know, but it's worth a few extra hundred dollars to me to have the best in that area. (I just wouldn't be willing to spend thousands for 2x7800GTX or a FX-57/60. :p )

[Rattle]

this is what I get with 2 hitachi 80gb 7200 rpm 8mb cache drives I paid $110 for...

http://photobucket.com/albums/b6/Rattle77/?action=view&current=omgwtfsata3.jpg

[Revivalist]

Ok, I'm starting to think that there's something wrong with HDTach. Notice that your burst is way above the maximum bandwidth possible with SATA-150. The most you could possibly have with two drives in RAID 0 is 150MB/s x 2 = 300MB/s. But you got 329MB/s. . . .

I actually had those same exact drives in RAID 0 before I moved over to the Raptors. Here's what I got in HDTach when I used the VIA controller and when I used the Promise controller on a ASUS A8V

http://img107.imageshack.us/img107/4421/hdtachcomparison7td.jpg

But when I used Windows64, I got this with the same drives . . .

http://img374.imageshack.us/img374/765/hdtachcrazyhighscore5xd.jpg

How CRAZY is that??!! . . .

Maybe we should try comparing ATTO scores. . . Do you have any benches with ATTO?

[Rattle]

Ok, I'm starting to think that there's something wrong with HDTach. Notice that your burst is way above the maximum bandwidth possible with SATA-150. The most you could possibly have with two drives in RAID 0 is 150MB/s x 2 = 300MB/s. But you got 329MB/s. . . .

I actually had those same exact drives in RAID 0 before I moved over to the Raptors. Here's what I got in HDTach when I used the VIA controller and when I used the Promise controller on a ASUS A8V

http://img107.imageshack.us/img107/4421/hdtachcomparison7td.jpg

But when I used Windows64, I got this with the same drives . . .

http://img374.imageshack.us/img374/765/hdtachcrazyhighscore5xd.jpg

How CRAZY is that??!! . . .

Maybe we should try comparing ATTO scores. . . Do you have any benches with ATTO?


thats because its not sata 1.5 LOL

these are sata 3.0 drives as is my board.. thats why you see that.

before i enabled sata 3.0 on the drives I was still getting more than 150 anyway.........

also you gotta take into account the controller your using, the controller on the DFI happens to be badass. Another reason why its the best board on the market....

I have never uses ATTO

got a linky ?

[Revivalist]

Ooooh, SATA II . . . that explains it. ;)

Yes, the NF4 controller on the DFI is supposed to be one of the best. But I'm not seeing the performance from 4 Raptors that I was expecting, at least not from the benchmarks. . ..

You can find ATTO here. . . .

www.s93916419.onlinehome.us/Intratech/bench32.exe

[shadowdr]

With reads and writes so far apart,I would run chkdsk and be sure there are no bad sectors.I have usually had some file corruption when they are that different.You should increase Que depth to 10 and run a couple of times for consistency.

[ajrettke]

IIRC raptors usualy write faster than read....kinda odd but I almost always remember that being the case.

Also, RAID 0 has diminishing returns (and sometimes negative effects) after 2 disks. With 4 disks RAID 0 you have to wait for the slowest disk each time before sending the data (which grealy increases seek times). Also the max bandwidth withers off as well as it's just not effecient to split small files up so many times and controllers become saturated.

[Revivalist]

Hmm, interesting info guys. Thanks. . .

I'll run Check Disk right now . . .

As for the efficiency of 4 disks in RAID 0, do you think it might hellp if I increased the stripe size from the usual 16Kb to something like 32kb or 64kb?

Also, about waiting for the slowest drive, I thought that since they're all 10K RPM that none of them should really cause that much delay compared to others. :eh?:

[Rattle]

Hmm, interesting info guys. Thanks. . .

I'll run Check Disk right now . . .

As for the efficiency of 4 disks in RAID 0, do you think it might hellp if I increased the stripe size from the usual 16Kb to something like 32kb or 64kb?

Also, about waiting for the slowest drive, I thought that since they're all 10K RPM that none of them should really cause that much delay compared to others. :eh?:

although they are the same speed they are not "synchronized" so whichever one loses a rev when they spin they all have to wait for it... this is why you lose performance with 4 from what I understand..

This is my first venture into raid with my hitachis, I figure bang for the buck it was worth the venture and they are as fast as any raptor. I reformat so much anyway I can easily just go back to non raid 0 setup. I like it though, gaming rules with them, especially in new games the load times have plumeted and I was using a single 74k raptor before...

[shadowdr]

Did chkdsk find anything?
You should see this thread (http://www.ocforums.com/showthread.php?p=4254338#post4254338) to see what they can come close to.