Dragon Fly BSD versus UBUNTU.

Discussion in 'all things UNIX' started by linuxforall, Nov 19, 2012.

Thread Status:
Not open for further replies.
  1. linuxforall

    linuxforall Registered Member

    Joined:
    Feb 6, 2010
    Posts:
    2,136
    http://www.phoronix.com/scan.php?page=article&item=dragonfly_linux_32&num=1


    For looking at the multi-core scaling performance, first up there were benchmarks done when the Core i7 3960X Intel CPU had 1, 2, 3, 4, 6, and 12 threads exposed to each operating system. The tested operating systems from this same hardware were DragonFlyBSD 3.0.3, DragonFlyBSD 3.2.1, and Ubuntu 12.10 Linux. All operating systems were in their default settings/packages aside from switching Ubuntu 12.10 to the GCC 4.4.7 compiler to match that of the DragonFly operating systems. Using the Phoronix Test Suite benchmarking software with OpenBenchmarking.org for result analytics, the results of the multi-core scaling / SMP tests were normalized against their single-core values to look purely at how well each OS is scaling with the same benchmark.
     
  2. Hmm. I don't know a whole lot about benchmarking, but are these benchmarks actually germane to the kind of workload Dragonfly is designed for?
     
  3. linuxforall

    linuxforall Registered Member

    Joined:
    Feb 6, 2010
    Posts:
    2,136

    Agreed but the main crux of this test was the claimed advantage of BSD in SMP multithreaded apps and that wasn't evident in these tests.
     
  4. That's kind of what I'm getting at though. GCC might show entirely different scaling characteristics from, say, Apache or MySQL, for all I know.

    Edit: my point is, when the Dragonfly people talk about higher performance, they are probably talking about it in a server role. Whereas I have no idea if the Phoronix benchmarks are more server-oriented, desktop-oriented, or pure-number-crunching-oriented.

    GCC and gzip look like workstation stuff. PHP could be more about servers I guess. I have no idea what the more benchmarky benchmarks (e.g. C-Ray) are about. What is this actually supposed to tell me about these OSes in a role where they could reasonably seen to be competing?
     
    Last edited by a moderator: Nov 19, 2012
  5. linuxforall

    linuxforall Registered Member

    Joined:
    Feb 6, 2010
    Posts:
    2,136
    What would have been more interesting is if they would have had a specific test to see how well either of the distros did multi cpu scaling. That is a boon in server as well as workstation environments.
     
  6. Not sure but I don't think that's possible.
     
Loading...
Thread Status:
Not open for further replies.