Results 1 to 10 of 10

Thread: Linux 3.0, 3.4, 3.10, 3.12 AMD Kernel Benchmarks

  1. #1
    Join Date
    Jan 2007
    Posts
    14,559

    Default Linux 3.0, 3.4, 3.10, 3.12 AMD Kernel Benchmarks

    Phoronix: Linux 3.0, 3.4, 3.10, 3.12 AMD Kernel Benchmarks

    How does the Linux kernel performance compare if simply bouncing between some of the recent long-term kernel releases? Well, to try to answer that question on Phoronix today we have benchmarks of the Linux 3.0.101, 3.4.68, 3.10.18, and 3.12.0 kernel releases tested from an AMD Opteron system with Radeon graphics.

    http://www.phoronix.com/vr.php?view=19333

  2. #2
    Join Date
    Apr 2013
    Posts
    65

    Default

    So in conclusion, it can be said that, except rare exceptions, Linux dev team has succeeded to squeeze several new features and bug fixes while keeping the performance relatively stable.

    IMO, that is an excellent deal.

  3. #3
    Join Date
    Oct 2013
    Posts
    192

    Default

    Great benchmark, thanks!

  4. #4
    Join Date
    Nov 2012
    Location
    France
    Posts
    561

    Default

    Thanks for the benchmark. However, I noticed that there used to be MIN/MAX FPS on game benchmarks but this is no longer the case, it would be nice if we saw them back.

  5. #5

    Default

    Quote Originally Posted by Calinou View Post
    Thanks for the benchmark. However, I noticed that there used to be MIN/MAX FPS on game benchmarks but this is no longer the case, it would be nice if we saw them back.
    They're there, nothing has changed, it all depends upon the specific test profiles as whether min/max information can be obtained.

  6. #6
    Join Date
    Jan 2012
    Posts
    151

    Default

    Quote Originally Posted by Michael View Post
    They're there, nothing has changed, it all depends upon the specific test profiles as whether min/max information can be obtained.
    I don't think min/max is representative of the experience someone might have when playing the game using the given kernel/drivers. I think a better alternative would be to show the 1st and 3rd quartiles (or, really, just use box plots instead of bar charts).

  7. #7

    Default

    Quote Originally Posted by FourDMusic View Post
    I think a better alternative would be to show the 1st and 3rd quartiles (or, really, just use box plots instead of bar charts).
    Not many games/engines expose enough for min/max let alone enough to do quartiles or box plots of the data.

  8. #8
    Join Date
    Dec 2011
    Posts
    76

    Default

    I may be missing something, but why would you use 3 different configurations for the hard drives when doing a benchmark?

    Wouldn't it make more sense to have the same configuration for all kernels? (and then if you wanted, have one kernel with the different configurations on a different article...)

  9. #9
    Join Date
    Jul 2013
    Location
    Brasil
    Posts
    96

    Default

    More features, less bugs and about the same performance. Kernel developers rocks!
    Thanks for the benchs Michael.

  10. #10
    Join Date
    Feb 2008
    Location
    Linuxland
    Posts
    5,034

    Default

    Quote Originally Posted by Michael View Post
    Not many games/engines expose enough for min/max let alone enough to do quartiles or box plots of the data.
    Oi Michael, there's a way to get frame times for all GL games you know And all the stats that can be gathered from that then.

Posting Permissions

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