Intel Ssd Raid 0 : How To Interpret Results?

Discussion in 'Mac Pro' started by ildondeigiocchi, Oct 23, 2010.

  1. ildondeigiocchi macrumors 6502a

    ildondeigiocchi

    Joined:
    Dec 30, 2007
    Location:
    Montreal
    #1
    I recently decided to tackle my hard drive mess. So i deleted the info on Windows 7 Intel X-25 M SSD and created a RAID 0 array with that drive and its identical which i ran OSX on. My home folder is now located on a 2TB WD Caviar Black. The thing is I don't notice much difference between 1 SSD running alone compared to a Raid 0 SSD array.

    I know that Xbench isn't the most reliable tool.... but nonetheless, are these results satisfactory for a Intel X-25 M raid-0 array?

    Results 229.03
    System Info
    Xbench Version 1.3
    System Version 10.6.4 (10F569)
    Physical RAM 6144 MB
    Model MacPro4,1
    Drive Type Macintosh HD


    Disk Test 229.03
    Sequential 209.78
    Uncached Write 136.15 83.59 MB/sec [4K blocks]
    Uncached Write 240.64 136.15 MB/sec [256K blocks]
    Uncached Read 162.80 47.64 MB/sec [4K blocks]
    Uncached Read 701.86 352.75 MB/sec [256K blocks]
    Random 252.16
    Uncached Write 120.90 12.80 MB/sec [4K blocks]
    Uncached Write 153.05 49.00 MB/sec [256K blocks]
    Uncached Read 2425.86 17.19 MB/sec [4K blocks]
    Uncached Read 1548.74 287.38 MB/sec [256K blocks]
     
  2. bluesteel macrumors 6502

    Joined:
    Apr 5, 2007
    Location:
    Earth
    #2
    sorry, i but i wanted to ask how you liked having the SSD as the boot drive and 2TB Black as the home folder? this is kind of the setup i'm looking to create myself, although i'm looking to get an OWC 120 SSD. do you recommend?
     
  3. ildondeigiocchi thread starter macrumors 6502a

    ildondeigiocchi

    Joined:
    Dec 30, 2007
    Location:
    Montreal
    #3
    Having an SSD as a boot drive and a large conventional Hard drive like a Western Digital 2TB Caviar Black is usually an ideal setup. I'm happy with the overall system performance but I was wondering if my benchmark results demonstrate adequate results.
     
  4. bluesteel macrumors 6502

    Joined:
    Apr 5, 2007
    Location:
    Earth
    #4
    i'm sure somebody who knows about this will post. i wish i could answer your question :)
     
  5. ildondeigiocchi thread starter macrumors 6502a

    ildondeigiocchi

    Joined:
    Dec 30, 2007
    Location:
    Montreal
    #5
    Thanks... The only reason why I didn't purchase OWC SSDs is because I already had the Intel ones. However, OWC Mercury SSD should outperform their Intel counterparts.
     
  6. philipma1957 macrumors 603

    philipma1957

    Joined:
    Apr 13, 2010
    Location:
    Howell, New Jersey
    #6
    the 229 number for your overall disk score is low. my numbers push 290. i have a new 2010 mac pro a quad 2.8 but having the newer machine should not affect disk scores too much.



    http://www.aja.com/products/software/ aja system test is a free download try that
     
  7. Transporteur macrumors 68030

    Joined:
    Nov 30, 2008
    Location:
    UK
    #7
    As you said, XBench is not reliable at all, so please don't use it.

    My single 160GB G2 has an overall score of 290 to 300, so it can't be possible that a RAID 0 out of two scores only 229.

    The software really is completely useless for SSDs. Use the AJA system test instead.
     
  8. Caesar_091 macrumors regular

    Caesar_091

    Joined:
    Jan 18, 2005
    Location:
    Italy
    #8
    Don't know if this will helps but here my results with my new OWC Mercury Extreme Pro SSD (250GB) installed in a Mac Pro 1,1 (3GHz):
    Code:
    Disk Test	368.78	
    		Sequential	226.57	
    			Uncached Write	276.43	169.72 MB/sec [4K blocks]
    			Uncached Write	307.91	174.21 MB/sec [256K blocks]
    			Uncached Read	118.63	34.72 MB/sec [4K blocks]
    			Uncached Read	423.79	212.99 MB/sec [256K blocks]
    		Random	990.32	
    			Uncached Write	1018.53	107.82 MB/sec [4K blocks]
    			Uncached Write	529.17	169.41 MB/sec [256K blocks]
    			Uncached Read	3625.66	25.69 MB/sec [4K blocks]
    			Uncached Read	1121.46	208.09 MB/sec [256K blocks]
    
    This test was made with the empty disk (just formatted to work on Mac). After I cloned my boot disk on the SSD drive (almost 140GB of stuff inside) I got these:
    Code:
    Disk Test	345.05	
    		Sequential	209.12	
    			Uncached Write	279.33	171.51 MB/sec [4K blocks]
    			Uncached Write	306.02	173.14 MB/sec [256K blocks]
    			Uncached Read	101.30	29.65 MB/sec [4K blocks]
    			Uncached Read	415.25	208.70 MB/sec [256K blocks]
    		Random	985.81	
    			Uncached Write	1052.24	111.39 MB/sec [4K blocks]
    			Uncached Write	528.28	169.12 MB/sec [256K blocks]
    			Uncached Read	3318.16	23.51 MB/sec [4K blocks]
    			Uncached Read	1095.41	203.26 MB/sec [256K blocks]
    
    But if I repeat the test randomly I always get different results (range from 360 to 340 global rating in disk test)... as always xBench is not that good bench app.

    Try AJA instead (see attaches)
     

    Attached Files:

  9. Caesar_091 macrumors regular

    Caesar_091

    Joined:
    Jan 18, 2005
    Location:
    Italy
    #9
    And these too.
     

    Attached Files:

  10. ildondeigiocchi thread starter macrumors 6502a

    ildondeigiocchi

    Joined:
    Dec 30, 2007
    Location:
    Montreal
    #10
    Wow AJA really did show a difference compared to Xbench. Guess the latter should just be dropped completely for unreliability. Check these results!
     

    Attached Files:

  11. Transporteur macrumors 68030

    Joined:
    Nov 30, 2008
    Location:
    UK
    #11
    That looks better. :D And yes, it is time to get rid of XBench! Unfortunately, too many people still use it and give credit to the results.

    Considering your write speed, I assume you're using the 80GB models, right?

    Edit: The reason why you don't notice any difference is the fact that the OS doesn't really benefit from higher sequential speeds, rather than access time, random speeds and IOPS, which don't really improve when you stripe two or more drives.
     
  12. philipma1957 macrumors 603

    philipma1957

    Joined:
    Apr 13, 2010
    Location:
    Howell, New Jersey
    #12
    you aja nmbers are good 150 write is double what a single drive does 70 or so and 498 read is more then double what a single drive does.
     
  13. freshface macrumors newbie

    freshface

    Joined:
    Oct 17, 2010
    #13
    Where can I notice a real speed gain if I RAID0 3xSSD instead of just using 1xSSD? In a boot partition? In booting OSX or launching apps? In work data disk in opening files?
     
  14. Einz macrumors regular

    Einz

    Joined:
    Feb 14, 2008
    Location:
    Miami
    #14
    This is result of a Intel 80G G2 after almost a year.

    Results 195.17
    System Info
    Xbench Version 1.3
    System Version 10.6.4
    Physical RAM 6144 MB
    Model MacPro3,1
    Drive Type INTEL SSDSA2M080G2GC
    Disk Test 195.17
    Sequential 170.76
    Uncached Write 120.56 74.02 MB/sec [4K blocks]
    Uncached Write 137.76 77.95 MB/sec [256K blocks]
    Uncached Read 173.03 50.64 MB/sec [4K blocks]
    Uncached Read 478.24 240.36 MB/sec [256K blocks]
    Random 227.71
    Uncached Write 185.61 19.65 MB/sec [4K blocks]
    Uncached Write 91.81 29.39 MB/sec [256K blocks]
    Uncached Read 2575.97 18.25 MB/sec [4K blocks]
    Uncached Read 1112.62 206.45 MB/sec [256K blocks]
     
  15. milo macrumors 604

    Joined:
    Sep 23, 2003
    #15
    While AJA may be more accurate, unfortunately it doesn't go into much detail.

    Is there an accurate OSX disk benchmarking app that splits out things like different block sizes?
     
  16. nanofrog macrumors G4

    Joined:
    May 6, 2008
    #16
    I don't think so, as this question comes up from time to time (directly), and more often, indirectly (i.e. RAID threads, even if not explicitly asked like this). :(
     

Share This Page