Announcement

Collapse
No announcement yet.

Eight-Way BSD & Linux OS Comparison

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

  • Sergio
    replied
    Originally posted by oleid View Post
    The only thing he might be referring to is the "nobarrier" option. Yet, even then his statement makes no sense.
    Indeed, the nobarrier option. Sorry for using the word 'insecure'; what I wanted to say is that you can make ext4 ridiculously fast, and hence the benchmarks doesn't really compare operating system's performance: i.e benchmarks are useless.

    Leave a comment:


  • Sergio
    replied
    Originally posted by eLDST0RM View Post
    regarding ext4 being insecure:



    I don't get it. The link is to some tests that show that performance on a database improves if you turn off barriers on ext4. I don't understand how you can take that fact and use it to conclude that ext4 is inherently unsafe.
    You are right. What I meant is that you can make ext4 insanely fast at the cost of having an absolutely unreliable file system. Compared to the overhead implied by zfs or btrfs, this would provide Linux an unfair advantage. What I'm saying is that, ideally, to really compare operating system performance, you should use the same file system, or at least the same CLASS of file system.

    Leave a comment:


  • Yfrwlf
    replied
    Originally posted by droidhacker View Post
    Isn't it about time for you to get banned for being a stupid motherfucker?
    Seems to me that compared to other similar AGED distros, it won some, and lost some. Overall, about the same. THAT IS PERFORMANCE ONLY, which is far from the only consideration when selecting a distro. You know that some people choose their distribution philosophically, don't you? Their (canonical's) lack of upstream contribution is a great reason to avoid them. Also their crappy UI, and other attempts to break compatibility with GNU/Linux.
    Concerned about the compatibility thing too. It's in a distro's interest to fracture it so that software compiled and distributed for them will be exclusive to them and force Linux users to have to use their distro, like Android for example. Ubuntu with their Mir display server makes me think oh great, now I have to make my software compatible with Xorg (if it doesn't get completely replaced by Wayland), Wayland (if it becomes popular), and Mir? I'm all for improving Linux, but if there are no standards in place that the improvements are using then it is just fracturing. As if Linux needs MORE software ecosystem earthquakes at this point. Linux hasn't even solved universal software distribution issues yet so that a healthy ecosystem can exist. Way too many packages out there with big dependency requirement lists that you're on your own to find.

    What distro do you think is better, anyway? Mint is based on a lot of Ubuntu's stuff but may dump Mir when it comes. Debian seems to be a better choice but I thought it still had a lot of usability issues. Fedora seems slick but there are always problems with it being too bleeding edge and I feel like an alpha tester for Red Hat when I use it a lot of times, while other times it seems fairly stable. RPMs also seem much slower than DEBs to install.

    Leave a comment:


  • talvik
    replied
    Originally posted by LightBit View Post
    Benchmarks really shouldn't be preformed in VM.
    I tried the same benchmark Ubuntu ext4 in bare metal and got similar results, completely flat and capped performance. While running the resource usage is near zero IO, iowait and CPU.
    Then I ran wheezy ext4 inside vbox, the performance was 3x to 5x faster than on bare metal. 3x better than the results in that bench.

    There is clearly a bottleneck somewhere... and it's not in the VM.

    Leave a comment:


  • droidhacker
    replied
    Originally posted by CthuIhux
    Quoting Obscure, Untrusted sources written by liars aren't going to make your BSD FUDS more valid.
    The interesting thing is that those liars he is quoting, didn't actually say anything bad, SECURITY WISE **OR** DATA INTEGRITY WISE about ext4 (some people really should think about terminology before spouting crap though, "insecure" relates to security, not data reliability). The only bad thing the linked page had to say about it was related to PERFORMANCE. The dangerous thing they were talking about was INTENTIONAL DANGER that they introduced themselves, and ACKNOWLEDGED to have introduced themself. That would be the nobarrier option, which is ***NEITHER DEFAULT NOR RECOMMENDED***.

    Leave a comment:


  • LightBit
    replied
    Originally posted by talvik View Post
    Look at those curves, the test is flawed. BSD on ZFS and Linuxes on EXT4 are completely flat and capped at exactly the same value, this clearly indicates some limitation of his virtualization environment.
    And then the performance explodes on UFS, Hammer and EXTnobarrier. There must be some disk operation that is capped on the VM that only Ext4 and ZFS use.
    Benchmarks really shouldn't be preformed in VM.

    Leave a comment:


  • talvik
    replied
    Originally posted by Sergio View Post
    Look at those curves, the test is flawed. BSD on ZFS and Linuxes on EXT4 are completely flat and capped at exactly the same value, this clearly indicates some limitation of his virtualization environment.
    And then the performance explodes on UFS, Hammer and EXTnobarrier. There must be some disk operation that is capped on the VM that only Ext4 and ZFS use.

    Leave a comment:


  • oleid
    replied
    Originally posted by i386reaper
    Quite an amusing read, knowing that the page is full of irony...

    Leave a comment:


  • oleid
    replied
    Originally posted by CthuIhux
    Problem is BSD not just a millisecond slower then Linux, it has proven to be as much as 500 times slower then Linux in many areas not tested in this benchmarks. [...]
    *yawn*

    I know that you are the anti-BSD-troll here... but, well.... One whole page of senseless comments about how bad BSD is? Oh come on! That's so boring...
    At least confirm your statement from above, if you can -- which I doubt --, so that the people here don't fall asleep.

    Leave a comment:


  • oleid
    replied
    Originally posted by RahulSundaram View Post
    Can you be more specific? I don't see any security issues there. Only a db specific fs benchmark of 3.2 kernel. That isn't worth much really.
    The only thing he might be referring to is the "nobarrier" option. Yet, even then his statement makes no sense.

    Leave a comment:

Working...
X