Disk performance test

Discussion in 'Mac Pro' started by BSartist, Feb 23, 2008.

  1. BSartist macrumors newbie

    Joined:
    Feb 16, 2008
    #1
    I have a new mac pro with the stock 320 GB drive in it. I added two WD3200AAKS drives are 7K RPM with 16M cache.

    I will be doing video editing as well as running parallels.

    I see that some people move the Mac boot drive off the stock drive and I want to figure out if my setup would benefit from that.

    My prior (windoze based) NLE software came with a utility that tested drive performance. Is there something on the Mac that I could use to determine if my new drives are faster enough to warrant the boot move?

    Or does anyone have similar setup/knowledge to give me some hints to the answer. I couldn't find anything concrete by searching around.

    Thanx much
    Bill
     
  2. GotPro macrumors 6502

    Joined:
    Jan 29, 2007
    #2
    One test is XBENCH

    Xbench

    Honestly...

    The best way to test both the read/write capabilities of your various drive is to take, say, your iTunes folder (or other such folder with many files)... try to have 10-20 GB of data in it... and copy it to each drive.

    Then... get your stopwatch out... and "DUPLICATE" the folder on each drive... that will test the read and write speed of a single drive, the throughput, as well as test it with many small and medium files (more the type of things you'll be dealing with in a boot drive scenario)...

    My Hitachi 1TB was ALMOST 3X as fast at duplicating my 22GB iTunes folder than the stock 320 from WD!!!
     
  3. BSartist thread starter macrumors newbie

    Joined:
    Feb 16, 2008
    #3
    I ran some of those tests...

    Interesting if not completely explainable results.

    I used a directory with one nested directory. Total of about 7,300 items of 24.4GB (music files as you suggested).

    Boot is the stock drive and Disk1,2 are the new ones - partitioned for the first time. Times are in mm:ss:hh (hh=hundredths)

    Nothing else was happening on the system. I repartition Disk2 between each trial initially - later I started erasing between each trial because I was puzzled by the results.

    07:14:98 - Boot to Disk1
    06:01:20 - Disk1 to Disk2
    06:07:57 - Disk1 to Boot

    So it seems that stock disk is somewhat worse on read but comparable on write? Then I wanted to see if Journalling made a difference since I'm not sure whether to journal my video drive.

    06:15:45 - Disk1 to Disk2 (not journaled) (slower than journaled.)
    06:22:60 - Disk1 to Disk2 (journaled) (above it was only 6:01!)
    06:11:41 - Disk1 to Disk2 (journaled - erased first)
    06:22:40 - Disk1 to Disk2 (not journaled - erased first)
    06:59:83 - Boot to Disk2 (journaled - repeating first trial).

    I guess these results are probably inconclusive and I guess too close to be significant?


    Bill
     
  4. GotPro macrumors 6502

    Joined:
    Jan 29, 2007
    #4
    Ok... if I see what you wrote correctly... I think you misunderstood what I was trying to get you to do.

    The times will be very close when copying from one drive to another... one drive, obviously, will be always be in the equation.

    Ok... copy the data from the boot drive to the other two drives.

    Now that the data is there (don't worry about timing how long it takes to copy it from the boot drive to drives 1 & 2)...

    Do a "duplicate item" from the finder... (Apple + D) first in the Boot Drive, time it... then on Drive One... then time it... then on Drive Two... and time it :)

    Or did you do that? LOL
     
  5. BSartist thread starter macrumors newbie

    Joined:
    Feb 16, 2008
    #5
    Test rationale

    No I didn't do that, I could try that. But I actually thought that my first interpretation of what you meant made sense. Basically this is what I figured:

    By trying enough combinations I would be able to isolate out the drives read performance from its write performance. For example, copying from Boot to Disk1 tests Boot's read and Disk1's write performance. Going back the other way would do the opposite. By throwing in a Disk1 to Disk2 (same types) that should give me a baseline to know where the limiting factor is.

    But I also see the point of your duplicate on the same drive test.
     
  6. GotPro macrumors 6502

    Joined:
    Jan 29, 2007
    #6
    The duplicate on the same drive test will showcase a much more strenuous read/write exercise while isolating the performance data to a particular drive!

    You'll know which drive is fastest, unconditionally, overall... this way!:apple:
     
  7. BSartist thread starter macrumors newbie

    Joined:
    Feb 16, 2008
    #7
    Ran the other test

    I hear ya. Initially I thought that separating out read and write was valuable (in some cases it is) but I see how compounding them together could give clearer results.

    I ran the test on the boot drive and then on one of the others.

    15:06 - Boot Drive
    14:27 - New Drive

    Still probably not statistically significant. Probably leave as-is unless I end up with something blindingly faster.

    Thanx for the advice.
     
  8. GotPro macrumors 6502

    Joined:
    Jan 29, 2007
    #8
    Yes... doesn't seem to be much of an improvement... certainly not something you're likely to notice.

    If you upgrade in the future... look for at least a 2x improvement over the stock drive (unless you go with another WD drive)
     
  9. BSartist thread starter macrumors newbie

    Joined:
    Feb 16, 2008
    #9
    thx again for your prompt and v.useful replies

    ...much appreciated.

    Now on to setting up the rest of my system :)
     

Share This Page