How I crippled the SSD in my Macbook

Discussion in 'MacBook' started by Jethryn Freyman, Mar 8, 2013.

  1. Jethryn Freyman macrumors 68020

    Jethryn Freyman

    Joined:
    Aug 9, 2007
    Location:
    Australia
    #1
    So, it's a Penryn 4,1, 2008 Macbook. Bought it a year ago on ebay. Stuck a Corsair Force 3 120GB SSD in it. Read/write speeds maxed out on the SATA I bus at 130MB/sec. That was with Lion and Snow Leopard. Lions' disk encryption didn't have any noticeable effect, though I never benchmarked it.

    Anyway, fast forward to me doing a total wipe and installation of Snow Leopard, Quickbench is maxed out at 130MB/sec read/write speeds.

    Then I encrypted the drive with PGP Whole Disk Encryption. It now maxes out at about 35MB/sec read/write.

    On the bright side, for small [4-128KB random] read/writes, it still kicks ass, the the third-party disk encryption really throttles the SSD, probably because of the CPU requirements [not that I see any CPU usage spikes], and the fact I'm using the oldest version of PGP WDE that is Intel-worthy.

    In usage, however, I don't notice things being slower.

    Just a random FYI.

    <3
     
  2. benwiggy macrumors 68020

    Joined:
    Jun 15, 2012
    #2
    FileVault 2 works very well at full disk encryption with minimal effect on performance. It's one of the many useful features in Lion that nobody mentions, because Lion was just some iOS crap. :D
     
  3. Jethryn Freyman thread starter macrumors 68020

    Jethryn Freyman

    Joined:
    Aug 9, 2007
    Location:
    Australia
    #3
    Yeah it did work very well, it was actually the sole reason I initially installed Lion instead of SL
     
  4. flynz4 macrumors 68040

    Joined:
    Aug 9, 2009
    Location:
    Portland, OR
    #4
    This is a great example that in effect... large block I/O (measured in MB/s) is one of the most useless factors for measuring SSDs in client computers. Almost all of the performance issues of HDD based client computers is small random reads. Your post demonstrates that even with your severe reduction in "large block" data transfer rate... it still feels fast. That is because the only thing that really maters (first order affect) is small random reads.

    /Jim
     
  5. Jethryn Freyman thread starter macrumors 68020

    Jethryn Freyman

    Joined:
    Aug 9, 2007
    Location:
    Australia
    #5
    Yeah pretty much sums it up, when I ran the benchmarks on an eMac with a standard hard disk [hit 50MB/sec for large transfer], it still overall took several times long to complete the benchmarks because the small reads were just so damn slow
     

Share This Page