- Jan 16, 2001
- 31,528
- 3
- 76
In AT's recent RAID comparo article, it was shown that HD data transfer rates went up, in step with an increase in stripe size.
In this thread from yesterday, I listed some benchmarks (SANDRA/HDTACH) showing the results of my RAID O array. I was using a 32kb stripe size. It was recommended that I go to a 64kb stripe size. I told you that I would change the stripe size and report back my results.
Well, unfortunately, the results weren't significantly higher. As a matter of fact, you could say the diff would be negligible. Same tests SANDRA and HDTach. I even ran each one twice, just in case.
Now, there's one definite fact that stands: the RAID 0 array's data transfer AND Random Access time is much faster than my single 45GB DM60+ drive. So, overall, I'm pleased that I did it.
I'm just curious why there was little to no difference in performance when changing stripe sizes. I realize this is probably a really deep technical issue that's gonna be over my head, but I want to learn. Thanks.
In this thread from yesterday, I listed some benchmarks (SANDRA/HDTACH) showing the results of my RAID O array. I was using a 32kb stripe size. It was recommended that I go to a 64kb stripe size. I told you that I would change the stripe size and report back my results.
Well, unfortunately, the results weren't significantly higher. As a matter of fact, you could say the diff would be negligible. Same tests SANDRA and HDTach. I even ran each one twice, just in case.
Now, there's one definite fact that stands: the RAID 0 array's data transfer AND Random Access time is much faster than my single 45GB DM60+ drive. So, overall, I'm pleased that I did it.
I'm just curious why there was little to no difference in performance when changing stripe sizes. I realize this is probably a really deep technical issue that's gonna be over my head, but I want to learn. Thanks.