Announcement

Collapse
No announcement yet.

Oracle Might Be Canning Solaris

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

  • #91
    Originally posted by ldo17 View Post

    Funny how you compare sources (which include all the options) rather than binaries (which include only chosen options) ...
    Have you and mr.Pawlerson compared the relative sizes and performance of the equivalent binaries between Linux and Solaris, identical CPU arch? Idle curiosity.. Because dude is throwing such self-assured statements. Leaving the binaries asides..

    Once you start comparing the sources of the respective kernels, you will see why Linus is throwing his temper tantrums..

    Comment


    • #92
      Originally posted by aht0 View Post

      Have you and mr.Pawlerson compared the relative sizes and performance of the equivalent binaries between Linux and Solaris, identical CPU arch?
      That’s been done since the early days:

      After a couple of weeks tunning [sic] the port, Linux/SPARC was able to keep
      up against SunOS and Solaris on the same hardware and in some cases
      outperforming them.

      The lmbench results are more interesting than they may appear at first
      sight: They do not only reflect that Linux is a great operating
      system, but most sadly it reflects the fact that corporate operating
      systems are sometimes bloated and slow.

      Comment


      • #93
        Originally posted by ldo17 View Post

        That’s been done since the early days:

        After a couple of weeks tuning [sic] the port, Linux/SPARC was able to keep
        up against SunOS and Solaris on the same hardware and in some cases
        outperforming them.
        The lmbench results are more interesting than they may appear at first
        sight: They do not only reflect that Linux is a great operating
        system, but most sadly it reflects the fact that corporate operating
        systems are sometimes bloated and slow.
        Does not tell on which hardware they "tunned" it for 2 weeks. Could have been "pizzabox" workstation..

        Now, in this light explain, if it's so easy and jolly to "best Solaris"..
        -why Solaris scales much better on multisocket system starting from certain socket count.
        -why there are more kernel schedulers around for Linux than my dog has fleas but still people seem unhappy with them.

        Comment


        • #94
          Originally posted by aht0 View Post
          Does not tell on which hardware they "tunned" it for 2 weeks. Could have been "pizzabox" workstation..
          You were trying to say Linux was bloated, when in fact it is Solaris that is bloated.

          Comment


          • #95
            my post was specifically tailored to be as meaningless as was your argument.

            Apply your brain and use logic. And try to get over the "I must defend Linux for all cost - unto the breach, unto the breach". I do criticize BSDs here and there, though they are my favorite family of OSes.

            Solaris was developed by single company with very specific goals in mind. It's sources are neat, well thought-out and concise. Whole operating system and it's tools are purposefully engineered. Linux still does not have some of the tools Solaris has, or if it has their equivalents are not working quite as well.

            Linux has.. evolved. It has issues with NIH syndrome, it has rather serious issues with a feature creep. Worse, getting "new feature" out in next announcement seems to be far more important than actually finishing it later. It has serious issues with reinventing the wheel the n'th time hoping it's perhaps rounder than the circle itself.
            It's kernel devs also seem to occassionally lapse into coding the most obscure way possible or worse, even flat-out trying to suppress errors, though their code seems to work or trying to suppress errors being produced right because of it "seems to work" otherwise..

            Shit like this would not be possible in a profit-oriented company. Such developers would get his/her asses fired.

            By it's design and nature, bloat in Linux is inevitable. It's a patchwork of an OS, like ducktaped GodZilla. Yeah, it gets it's job done within it's limitations, even rather well when specific efforts have been made. But you can never compare it's "bloat-factor" with something like Solaris which simply does not have much in a way of excess code. It does not have had herd of tinkerers being attached to it since it's inception, doing whatever the hell they liked, how they liked and when they liked. Solaris was produced by engineers with predefined set of goals under unified rule.

            Getting Linux work 'properly' (measurable as enterprise level needs) demands reverse effort nearly as large, removal/disabling/fixing of the excess that may make it unstable/unreliable and endless backporting become necessity. It's not "engineered", it's more akin to selective breeding. Pick animal (kernel version) that seems healthy enough, old enough and start fixin' it by crossing it with newer model hoping that positive traits of a "new animal" do carry over..
            Last edited by aht0; 05 January 2017, 07:22 AM.

            Comment


            • #96
              Originally posted by aht0 View Post
              By it's design and nature, bloat in Linux is inevitable.
              And yet it has not happened: it remains more efficient and versatile than any of its proprietary rivals. It works better on hardware designed for Windows than Windows does. It works better on SPARC than Solaris does. It works better on POWER than AIX does. It works better on IBM mainframes than IBM’s own mainframe OSes do. When somebody develops a new processor architecture nowadays, the first OS they boot on it is Linux.

              Linux rules, because it is the best.

              Comment


              • #97
                Originally posted by ldo17 View Post
                And yet it has not happened: it remains more efficient and versatile than any of its proprietary rivals. It works better on hardware designed for Windows than Windows does. It works better on SPARC than Solaris does. It works better on POWER than AIX does. It works better on IBM mainframes than IBM’s own mainframe OSes do. When somebody develops a new processor architecture nowadays, the first OS they boot on it is Linux.

                Linux rules, because it is the best.



                (genuine applause)

                Comment


                • #98
                  Originally posted by ldo17 View Post

                  And yet it has not happened: it remains more efficient and versatile than any of its proprietary rivals. (1)It works better on hardware designed for Windows than Windows does. (2) It works better on SPARC than Solaris does. It works better on POWER than AIX does. It works better on IBM mainframes than IBM’s own mainframe OSes do. (3)When somebody develops a new processor architecture nowadays, the first OS they boot on it is Linux.

                  Linux rules, because it is the best.
                  (1)lolz, no. It works much worse on Linux than on Windows. In fact, I haven't even gotten certain combinations of dual-MGPU (both Radeons) to even display picture in Linux without lot's of swearing.

                  (2)I'm refraining from personal comment on SPARC, no hardware to have tested it personally.
                  Just something for you to digest. Do you see Linux there? For some reason Oracle used Solaris for achieving it's record. Not Linux, although it "works better" in your opinion. Comparing it with a Windows.
                  http://www.oracle.com/us/solutions/p...p-3234897.html
                  (3) can't decide if you sound like a fanboy or some marketing department enthusiastically selling real estate on Western-Bumfuck.

                  Comment

                  Working...
                  X