raid question

[eh?]

I did raid through windows 2000, and well i did a benchmark using sisoft and the drives scored lower than a 4 gig hard drive... it is 2 10gigs. is there any thing i could do to improve this?

[redduc900]

Just a few questions about when you set up the array...

1. What stripe size did you set up the array with?
2. Did you format the array using FAT32 or NTFS? (the partition that the OS resides on).
3. If you formatted the boot partition using FAT32, what cluster size did you set it up with?
4. I'm not familiar with your boards' RAID BIOS settings, but if there's an option for enabling " Drive Write Cache"...is it enabled?
5. In the RAID BIOS, if there is an option to enable/disable "Boot Sector Virus Protection"...what is it set to?
6. If there is an option in the RAID BIOS for enabling/disabling "DMA Read/Write"...what is it set to?

[eh?]

Originally posted by redduc900
Just a few questions about when you set up the array...

1. What stripe size did you set up the array with?
2. Did you format the array using FAT32 or NTFS? (the partition that the OS resides on).
3. If you formatted the boot partition using FAT32, what cluster size did you set it up with?
4. I'm not familiar with your boards' RAID BIOS settings, but if there's an option for enabling " Drive Write Cache"...is it enabled?
5. In the RAID BIOS, if there is an option to enable/disable "Boot Sector Virus Protection"...what is it set to?
6. If there is an option in the RAID BIOS for enabling/disabling "DMA Read/Write"...what is it set to?
1. Im not sure i did it through windows 2000 i think its 0....
2. They are 2 10 gigs, on 1 i formated 1.5 gig to install win2k on and then left the rest of both of them un partioned, and when i got into win2k then upgraded to dynamic drives and then it made a 1 big 17 gig harddrive.
3. I used NTFS and left the cluster size at default.
4. Its not done with hardware its done with software through win2k. so there is no hardware fir the raid.

[Jon]

Using Sandra to benchmark it with?

If so, use something like ATTO or HDTach. Sandra is a very unreliable disk benchmark.

[Sonny]

Software Raid? Isn't there a big performance hit when you do that?

[eh?]

Originally posted by sonny
Software Raid? Isn't there a big performance hit when you do that?
i hope not i did it to make it faster....

[Sonny]

Let me rephrase that. You take a BIG performance hit when you do software raid.:eek:

[Jon]

Software RAID really doesn't use many more CPU cycles than a "hardware" RAID card will. Both are primarily software dependant.

I used the software RAID function before the HPT370 controllers came out and my CPU usage only dropped from about 8% usage while copying large files to about 3% when getting an add-on card. Didn't notice a whole lot of speed difference either. RAID controllers might give you 10-15% increase in speed over using the software built into Win2K...biggest plus to me is that you can install the OS on the stripe when using a controller.

[eh?]

so was it worth doing, and is ther eany way to make it faster?

[Jon]

It was worth it, in my opinion.

As far as making it faster...not really. All depends on what benchmark you're using to see the throughput you believe is slow. Like I said before, I've always had problems with Sandra when benchmarking RAID arrays.