Xbench disk test: 40, time for new hdd?

Discussion in 'PowerPC Macs' started by johannnn, Dec 23, 2011.

  1. johannnn macrumors 65816

    johannnn

    Joined:
    Nov 20, 2009
    Location:
    Sweden
    #1
    Hi everyone!

    I just got hold of a iMac G4, 1.25 GHz 768mb ram. It has Leopard installed.
    I bought an 1GB SO-DIMM ram and bumbed the total ram to 1.25GB, feels nice.
    I ran the Disk Test of Xbench 1.3 to see how well the drive is doing after all these years (System Profiler says it's a Seagate Barrecuda 7200rpm), and I'm getting a total of 40 (reproducible).

    Here's the stats of one run:

    Code:
    	Disk Test	40.48	
    		Sequential	73.32	
    			Uncached Write	69.84	42.88 MB/sec [4K blocks]
    			Uncached Write	77.53	43.86 MB/sec [256K blocks]
    			Uncached Read	58.17	17.02 MB/sec [4K blocks]
    			Uncached Read	98.52	49.52 MB/sec [256K blocks]
    		Random	27.96	
    			Uncached Write	9.46	1.00 MB/sec [4K blocks]
    			Uncached Write	63.75	20.41 MB/sec [256K blocks]
    			Uncached Read	79.05	0.56 MB/sec [4K blocks]
    			Uncached Read	110.46	20.50 MB/sec [256K blocks]
    This is the first time I run Xbench, but 9 in random uncached write doesn't seem too high. However, the Xbench database has been down now for 2 days since I ran the test so I can't really compare my scores.

    Can anyone say anything about this score? My wallet is thin but if the drive is a big bottleneck I'm very tempted to get a new IDE-drive, as I've planned to use this computer as my main machine at work for 1-2 years. A 2.5" drive is also tempting to reduce fan noice, but I think it's too complicated for me with adapters and stuff (I'm scared just to replacing the drive).

    Thanks!
     
  2. Hellhammer Moderator

    Hellhammer

    Staff Member

    Joined:
    Dec 10, 2008
    Location:
    Finland
    #2
    What is the capacity? That looks normal for an old low-capacity HD.
     
  3. mulo macrumors 68020

    mulo

    Joined:
    Aug 22, 2010
    Location:
    Behind you
    #3
    your drive is fine, IDE doesn't get much faster then that
     
  4. johannnn thread starter macrumors 65816

    johannnn

    Joined:
    Nov 20, 2009
    Location:
    Sweden
    #4
    80GB (http://www.seagate.com/ww/v/index.jsp?vgnextoid=66cc26a2c88ef010VgnVCM100000f5ee0a0aRCRD&locale=en-US&reqPage=Legacy).

    Do you have any idea of what a new HD gets? I've been looking at 80GB & 160GB IDE drives that a few stores sells.

    Ok thanks for information. Although I hoped I could easily bring some more juice to this machine :(
     
  5. Hrududu macrumors 68020

    Hrududu

    Joined:
    Jul 25, 2008
    Location:
    Central US
    #5
    Those are just not the greatest drives. They're pretty old technology, so you definitely can get much better results with a newer HDD in there. I ran Xbench on my 1GHz 17' iMac which has the original 80 Seagate Barracuda just like yours, and also on my G4 Cube which has a new Western Digital Caviar 160GB drive in it. Both are 7200rpm. First the iMac:


    Disk Test 33.27
    Sequential 64.34
    Uncached Write 53.39 32.78 MB/sec [4K blocks]
    Uncached Write 64.76 36.64 MB/sec [256K blocks]
    Uncached Read 67.71 19.81 MB/sec [4K blocks]
    Uncached Read 75.57 37.98 MB/sec [256K blocks]
    Random 22.44
    Uncached Write 7.14 0.76 MB/sec [4K blocks]
    Uncached Write 58.63 18.77 MB/sec [256K blocks]
    Uncached Read 87.29 0.62 MB/sec [4K blocks]
    Uncached Read 102.83 19.08 MB/sec [256K blocks]

    And now the Cube:


    Disk Test 50.44
    Sequential 57.76
    Uncached Write 49.64 30.48 MB/sec [4K blocks]
    Uncached Write 82.16 46.48 MB/sec [256K blocks]
    Uncached Read 43.01 12.59 MB/sec [4K blocks]
    Uncached Read 73.08 36.73 MB/sec [256K blocks]
    Random 44.77
    Uncached Write 16.73 1.77 MB/sec [4K blocks]
    Uncached Write 102.24 32.73 MB/sec [256K blocks]
    Uncached Read 87.53 0.62 MB/sec [4K blocks]
    Uncached Read 119.59 22.19 MB/sec [256K blocks]

    As you can see, the Cube score is much better even though its an older computer with a slower ATA bus.
     
  6. johannnn thread starter macrumors 65816

    johannnn

    Joined:
    Nov 20, 2009
    Location:
    Sweden
    #6
    Thanks for posting your scores! Unfortunately 40 -> 50 says nothing to me, is it a big jump? Or hardly noticeably?
     
  7. Hrududu macrumors 68020

    Hrududu

    Joined:
    Jul 25, 2008
    Location:
    Central US
    #7
    The problem is the Cube is just a much slower computer all around than the iMac. Applications start up fast and boot time is good. If you get a new drive in your iMac, you'll see improvement for sure.
     
  8. ThunderSnake macrumors 6502

    Joined:
    Jul 23, 2010
    #8
    40 to 50 is a 25% performance increase.

    You can do better than what you have on your ATA100 machine. A WD5000AAKB would probably give you 60 to 70% better drive performance.

    Also, last year, someone posted their SSD results:

    http://forums.macrumors.com/archive/index.php/t-1003716.html
     
  9. johannnn thread starter macrumors 65816

    johannnn

    Joined:
    Nov 20, 2009
    Location:
    Sweden
    #9
    I found a Samsung 400GB drive (HD400LD) in my old room at parents house, didn't know I had one. Once I get a hold of some thermal paste I'll report how it performs!
     
  10. Jethryn Freyman macrumors 68020

    Jethryn Freyman

    Joined:
    Aug 9, 2007
    Location:
    Australia
    #10
    Xbench sucks for hard disk benching, I get wildly different results all the time.
     
  11. zen.state macrumors 68020

    zen.state

    Joined:
    Mar 13, 2005
    #11
    It sucks at everything it does to be honest. It's sad how many rely on it or think it's a worthy tool.
     
  12. castillo4141 macrumors member

    Joined:
    Sep 4, 2011
    #12
    And, why is that? What would you consider a reliable testing method for HDDs and SSDs?
     
  13. zen.state macrumors 68020

    zen.state

    Joined:
    Mar 13, 2005
    #13
    AJA System Test
     
  14. Jethryn Freyman macrumors 68020

    Jethryn Freyman

    Joined:
    Aug 9, 2007
    Location:
    Australia
    #14
    Haven't tried it yet, but it looks good. Will try.
     
  15. castillo4141 macrumors member

    Joined:
    Sep 4, 2011
    #15
    It's not a very deep test, it seems to only test sequential transfers, and doesn't even show if it is 256k or 4k. To be honest, I really don't see much difference on sequential transfers (since it's already high enough and I don't really perform large files transfers regularly), I'm more concerned on random read/writes (especially small files -> 4k) since that function is more useful for regularly daily tasks I perform (opening programs, saving/opening small files, web surfing, etc.).

    Is there any way to see that on the AJA System Test? Or any other test that does it better than XBench? I know is old, but still shows more info and is free; anyway it will be nice to use an up to date program instead.
     
  16. zen.state macrumors 68020

    zen.state

    Joined:
    Mar 13, 2005
    #16
    The point of it is it's very consistent. Xbench is far from consistent.

    A better and far more though HD test is QuickBench included in the SpeedTools software package. This is about $20 last time I looked. It also tends to come free with a lot of OWC storage products which is how I got it. AJA is simple but consistent and free so thats why I gave that link. Xbench can read one speed then 2 min later without doing any other task on the system the same test will be 10-20% up or down. You simply can't trust results like that.
     
  17. castillo4141 macrumors member

    Joined:
    Sep 4, 2011
  18. Nameci macrumors 68000

    Nameci

    Joined:
    Oct 29, 2010
    Location:
    The Philippines...

Share This Page