Results 1 to 7 of 7

Thread: OpenSolaris 2009.06 Performance

  1. #1
    Join Date
    Jan 2007
    Posts
    14,350

    Default OpenSolaris 2009.06 Performance

    Phoronix: OpenSolaris 2009.06 Performance

    Sun Microsystems released OpenSolaris 2009.06 on Monday as a key update to this Solaris desktop operating system used on both servers and desktops. The 2009.06 release introduced better codec support, SPARC support, improved hardware support, numerous enhancements to the Image Packaging System, and plenty of other changes we talked about in our article on Monday. Today though we are here with some benchmarks from this new OpenSolaris 2009.06 x86 release as we compare the performance to its predecessor, OpenSolaris 2008.11.

    http://www.phoronix.com/vr.php?view=13913

  2. #2

    Default

    I wonder why compiler has impact on Threaded I/O Tester?

  3. #3
    Join Date
    Oct 2007
    Location
    UK
    Posts
    160

    Default

    with 2gb of ram?

    Still, interesting benchmark

  4. #4
    Join Date
    Jun 2009
    Posts
    1

    Default

    I find the results very strange with the bork metric. Its a pure java based
    metric, gcc wouldn't be used here at all atll. So why does it show almost
    a 4 second difference between the opensolaris 2009.06 gcc3.4 and gcc4.3
    runs ? How many runs were done and whats the standard deviation with them ?

    For benchmarks the number of runs done and the standard deviation are
    very important with comparing two or more samples like here.

  5. #5
    Join Date
    Feb 2009
    Posts
    49

    Default

    Some questions:

    1. For the Java tests (incl. Java2D, etc..) why is OpenSolaris 2009.06 + gcc 4.3 listed as an option ? I think it doesn't make sense if the compiler is not involved.

    2. Building the Threaded I/O Tester on OpenSolaris 2009.06 seems to lead to a huge regression in its latency performance when it comes to four threads of 64MB writes." Hmm, there is such a difference when building with gcc 4.3.2 vs 3.4.3 ?

    3. "The C-Ray ray-tracing performance had regressed when being built with GCC 4.3.2 -- it took 56% more time to build." More time to build ? Or to run ? I'm confused as to what's being measured here.

    4. "Shortly we will have a new round of Ubuntu Linux vs. OpenSolaris testing." On OpenSolaris 32bit apps are generated by default by the compiler so you have to set CFLAGS/CXXFLAGS to "-m64", as comparing 32 bit apps to 64 is not exactly fair.
    Last edited by etacarinae; 06-04-2009 at 07:19 PM.

  6. #6

    Default

    Quote Originally Posted by etacarinae View Post
    4. "Shortly we will have a new round of Ubuntu Linux vs. OpenSolaris testing." On OpenSolaris 32bit apps are generated by default by the compiler so you have to set CFLAGS/CXXFLAGS to "-m64", as comparing 32 bit apps to 64 is not exactly fair.
    Let's hope Michael will take this into consideration.

  7. #7
    Join Date
    Feb 2009
    Posts
    9

    Default

    Quote Originally Posted by kraftman View Post
    Let's hope Michael will take this into consideration.
    Indeed! That's also my hope! Otherwise comparing 64bit to 32bit is like comparing apples and oranges. :-)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •