Announcement

Collapse
No announcement yet.

DragonFlyBSD Improves Performance Against Linux

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • staalmannen
    replied
    I am curious to know... would a DragonFly vkernel be able to run in user land of another OS? It is considered similar to UML but I wonder if it also could be used similarly to coLinux.

    A "coBSD" kernel (or a BSD syscall compatibility module for the Linux kernel) would be fun to play with on a Linux host (or perhaps am I rather alone in this) - especially the theoretical posibility to be able to run a BSD system in a chroot on a Linux host without full virtualization.

    Leave a comment:


  • enjolras
    replied
    Yes you're right. Linux rox, and has no scalability issue, it's well known. And postgresql scales easily on latest linux kernels. Oh wait !
    http://http://lwn.net/Articles/518329/

    Leave a comment:


  • dfcat
    replied
    Disclaimer: I am a df dev.

    """
    Originally posted by DaemonFC View Post
    So they came in second to an aging Enterprise Linux distribution and didn't even use the current version of that... but I'll let that slide because......FreeBSD....and NetBSD.. . ouch.
    """


    Scientific linux is based on RHEL - which also means that kernel/c library etc. versions are almost identical from release to release. So 6.2.x vs 6.3.x is practically irrelevant.

    And how is this "aging" since it is the latest one? Or are you saying that the test should have picked a non-enterprise distribution in general, like fedora or ubuntu or something?


    """
    I know their performance is bad (partially because they hate the GPL 3 imposing all of that horrible Freedom and don't want to use a recent version of GCC), but I didn't know it had decayed to this point.
    """

    This statement is just incorrect in so many ways. Due to the high level of parallelism in the test (many core machines), this is less
    of a indicator of 'performance decaying' as 'how far the different systems have advanced to take advantage of SMP & chip level SMT, etc'
    recall at one point, they did not support these things, and have all required major modifications (aka years of hard work) by all involved
    to get to their respective levels of advancement at present, with much to go for all involved.

    And being a database test, this has much more to do with scheduler algorithms, IO and virtual memory subsystems etc. than compiler optimization - eg. the test is system/IO/memory bound, and not compute bound, so compiler is less of a factor.

    The benchmarking was initially prompted by a major change in how postgres handles its memory structures - a change which
    has an extremely negative impact on how many unix virtual memory systems have traditionally been implemented - aka
    a questionable change to make - and some of the other non-dragonfly OS's have not been adapted to cope with this.
    This was a very recent change in postgres, so that is not a 'omg bsds are so old and not shiny' thing. So while it is good to
    see dragonfly competitive with Linux, this should be kept in mind. IIRC FreeBSD had a different workaround to avoid this problem
    but not 100%.

    R.e. GCC: firstly: dragonfly uses GCC4.4 - which is the same GCC major release as in RHEL6.. and 4.7 has been imported to switch
    over after this release. Also - performance tests indicate that clang (which FreeBSD will be switching to in 10.x) - is superior to gcc
    in some performance tests, but slower in others - so in other words, they are extermely competitive:

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite


    So really, I'm not sure what your compiler / gpl3 statements are in reference to here.

    Leave a comment:


  • DaemonFC
    replied
    So they came in second to an aging Enterprise Linux distribution and didn't even use the current version of that... but I'll let that slide because......FreeBSD....and NetBSD.. . ouch.

    I know their performance is bad (partially because they hate the GPL 3 imposing all of that horrible Freedom and don't want to use a recent version of GCC), but I didn't know it had decayed to this point.

    Leave a comment:


  • phoronix
    started a topic DragonFlyBSD Improves Performance Against Linux

    DragonFlyBSD Improves Performance Against Linux

    Phoronix: DragonFlyBSD Improves Performance Against Linux

    Benchmarks coming out of the BSD camp are showing that the soon-to-be-released DragonFlyBSD 3.2 is almost as fast as Scientific Linux (RHEL) 6.2 in at least one real-world workload...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite
Working...
X