1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Need your opinion

Discussion in 'Storage' started by repman244, Aug 25, 2011.

?

What is faster

  1. Hitachi Ultrastar 15k 450GB

    0 vote(s)
    0.0%
  2. 3xSeagate Savvio 10k.2 in RAID 0

    0 vote(s)
    0.0%
  1. repman244

    repman244

    Joined:
    Apr 7, 2011
    Messages:
    1,099 (0.92/day)
    Thanks Received:
    450
    For quite some time I was thinking of using the 3 SAS drives from my DL380 G5 in my PC instead of my Hitachi SAS.
    Today I finally decided to benchmark both my drive and the 3 drives in the DL which are in RAID 0. I always come up with weird results which probably come from using different RAID controllers.
    My PC uses an Adaptec 2405 with 128MB cache (write-back cache), on the DL there is a HP P400 controller with 256MB of cache (no idea if it's write-back or write through I tried different settings regarding caching and come up with even worse results).

    Hitachi drive is a HP branded drive (450GB) with 4 platters spinning at 15k RPM and it's a standard 3.5'' drive.
    The drives in RAID 0 are 3 Seagate HP branded drives (146GB each) with 2 platters spinning at 10k RPM and are 2.5'' (If I remember correctly they are Seagate Savvio 10.2).

    I remember doing benchmarks on the Seagate drives before putting them in RAID 0 and I was getting around 75-80MB read, that is why the following results seem weird to me.

    This is my Hitachi drive (ignore the burst speeds and CPU utilization):

    [​IMG]

    This is the RAID 0 on the DL (ignore the dips since they are probably cause by system access):

    [​IMG]

    I proceeded with ATTO:

    My Hitachi drive:

    [​IMG]

    RAID 0 on the DL:

    [​IMG]

    Now this is my problem, the numbers with RAID 0 are all over the place this is only one of the runs I made.
    Some came much better than my Hitachi, others about the same, the one posted is probably the worst of all, am I missing something?

    Either way I continued to do more benchmarks, next is HD Tune (I heard it doesn't do well with RAID's but I ran it anyway).

    Seeks (Hitachi):

    [​IMG]

    RAID 0 on DL:

    [​IMG]

    Obviously the Hitachi is better because of 15k RPM

    Next, sequential read.

    Hitachi:

    [​IMG]

    RAID 0 on DL:

    [​IMG]

    Again, ignore the bursts and the CPU usage.

    Before you ask why I didn't run the 3 drives from my server on my Adaptec controller, well I need to get 3 adapters for the connectors, since the 3 drives are HP hot-swap and don't have separate connector for the interface cable and power cable.

    So my question is: is it worth using the 3 SAS drives instead of my Hitachi, or should I get the adapters and try them on my Adaptec which is probably a better card than the HP P400 controller.
  2. repman244

    repman244

    Joined:
    Apr 7, 2011
    Messages:
    1,099 (0.92/day)
    Thanks Received:
    450
    Okay I did a re-test today so the results aren't that much affected by the controller cache.
    (ignore bursts and CPU utilization)

    Hitachi:

    [​IMG]

    RAID 0:

    [​IMG]

    I think the results for the RAID 0 are much better now and I think the cache has any influence on it anymore.

    Hitachi:

    [​IMG]

    RAID 0:

    [​IMG]

    Hitachi:

    [​IMG]

    RAID 0:

    [​IMG]


    So is it safe to say that if I use the RAID 0 in my PC I would get better performance than with my Hitachi I'm currently using?

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page