Announcement

Collapse
No announcement yet.

Linux hacker compares Solaris kernel code:

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

  • #31
    Originally posted by brosis View Post
    I use BFS on zenkernel in Debian. Maybe its time for a good benchmark of schedulers, not only in numbers but in application switch times etc, something that Michael can't do because it would require drifting outside of what PTS can measure.
    Something like this would be nice to see. When comes to PTS it can benchmark FPS in games which can be interesting as well.

    Comment


    • #32
      Originally posted by Pawlerson View Post
      Not true at all. Linux on desktop is actually using CFS. I remember Con's scheduler in the past and it was disaster. 3D games were unplayable and sound has stuttered a lot. Afaik CFS was introduced in 2.6.23 and it has changed my desktop experience.
      Here's what Linus had to say:

      "I realize that this comes as a shock to some of the SD people, but I'm told that there was a university group that did some double-blind testing of the different schedulers - old, SD and CFS - and that everybody agreed that both SD and CFS were better than the old, but that there was no significant difference between SD and CFS."

      So why was CFS picked instead of SD? SD was written by a hobby hacker (CK), who's an anesthesiologist in real life, while CFS was written by a professional hacker (IM), who's actually just that in real life. Keep in mind that by 2008, the times where kernel development was driven by people who'd code for it in their spare time without commercial interests was over for good. Was mattered most was having a maintainer who can work on it full-time.

      Personally, I can't imagine Linux without BFS anymore. I'm happy CK is still around and making my desktop better. I also find it amazing how people can be so hostile (as apposed to just criticize the work on technical merits) to someone who's providing something for free, in his spare time, on a "I hope this works for you too" basis. If you it doesn't work for you, don't use it. The world does not owe you anything.

      Comment


      • #33
        Originally posted by ChrisXY View Post
        Hm... Should I get into this?
        Please, dont.


        The last time I tried OpenSolaris (some time ago) it definitely had a higher memory footprint than an equivalent linux system and I think I noticed that it was a little bit slower. But all in all it was ok.
        OpenSolaris was a alfa version of Solaris 11. It was not even beta. Code that is in alfa stage is not yet done. Solaris 11 performs much better than OpenSolaris.


        What I have not seen from any of the trolls here is any actual data on how much slower (or unreliable or whatever they claim) it really is.

        Now I personally do prefer Linux and I think I prefer the GPL too. But I don't see how repeating unfounded claims over and over again helps your "cause" in any way.
        Here are some benchmarks between Solaris and RedHat. RedHat is what is used in production, no one runs bleeding edge Linux in production - that would be madness. For instance, all stock exchanges run well tested Linux distros. For instance, London Stock Exchange runs Linux + Solaris, but I dont know which Linux distro. I do know that NASDAQ runs RedHat.

        Solaris SPARC T5 server with 2 cpus, vs Red Hat x86 server with 8 cpus:

        Here we see that the SPARC cpu is much faster than the fastest x86 cpus. It can be explained because SPARC is a much faster cpu than x86, and partly because Solaris scales well and Linux does not scale.

        Here is Solaris on 2 cpu x86 server, vs Linux on the same server:

        We see that the Solaris machine runs much smoother. Linux stutters and has peaks all the time. Solaris has no peaks, it is smooth butter.

        Here is an older SAP benchmark, Solaris 10 vs Linux on similar hardware. Linux used slightly faster x86 cpus and RAM, and Solaris used slower cpus and slower RAM, but both servers used 8 socket x86 AMD opteron 6-core cpus.
        download.sap.com/download.epd?context=B1FEF26EB0CC34664FC7E80B933FC CAC80DD88CBFAF48C8D126FB65D80D09E988311DE75E0922A1 4
        vs

        We see that Linux got 45,000 SAP and Solaris 10 got 55.000 SAP. Linux used faster hardware and still scored lower. The reason is because Linux has a hard time scaling to 8 cpus.


        Soooo, please back up your claims or keep quiet. Thanks.
        It is very typical of Kraftman to spread negative rumours, and never back them up, and call people ugly names ("retard",etc). This is exactly what Pawlerson and OpenSLOWlaris does. These three persons are probably the same. Actually, Kraftman confessed he FUD in one post.


        Japp, min Sunray funkar fin-fint!

        Comment


        • #34
          Originally posted by kebabbert View Post
          OpenSolaris was a alfa version of Solaris 11. It was not even beta. Code that is in alfa stage is not yet done. Solaris 11 performs much better than OpenSolaris.
          Let's not forget the significant disadvantages of nonfreedom: "Production use of Oracle Solaris requires a support contract."

          Originally posted by kebabbert View Post
          Here are some benchmarks between Solaris and RedHat. RedHat is what is used in production, no one runs bleeding edge Linux in production - that would be madness.
          Would it be?

          Solaris may be more performant than Red Hat Enterprise Linux, but come on, they use ancient software like Kernel 2.6.32, Gnome 2.28, KDE SC 4.3...

          But thanks for being the only one arguing with actual data. I'll look at it later maybe.

          Originally posted by kebabbert View Post
          We see that Linux got 45,000 SAP and Solaris 10 got 55.000 SAP. Linux used faster hardware and still scored lower. The reason is because Linux has a hard time scaling to 8 cpus.
          So this is another point I don't get. Linux doesn't scale to 8 CPUs yet the supercomputer at my university with more than 100.000 cores (I believe 2 16 core CPUs per node) runs Linux. Sure, there may very well be other systems that scale better, but saying linux has a "hard time"? Come on!



          To the topic of BFS: I used it for a while on a netbook as part of the pf kernel (I wanted to try tux on ice): http://pf.natalenko.name/
          It was neither much better nor much worse than CFS.

          Comment


          • #35
            Don't bother replying to kebbart.



            That is what I will always think of whenever I see that handle.

            Comment


            • #36
              Originally posted by Pawlerson
              It's simple: slowlaris folks and fanboys like Kebabbert can't accept an obvious fact that Linux wiped this OS out from nearly every market. The reason of trolling is envy.
              You know, pride can be just as bad as envy. Not to say you shouldn't hold a certain amount of pride in your achievements, but you should keep it in check, along with your envy, lust, or whatever other of the seven you should choose to partake in.

              "Pride before the fall," and all that.

              Comment


              • #37
                Originally posted by kebabbert View Post
                I started to feel a little embarrassed by what we have as our own kernel.
                ...
                whereas ours looks like a garage job with duct tape by comparison.
                LOL. CK has been critical of the Linux scheduler for what, a decade now? This is probably less critical than he usually is.

                I'm not sure what the point of posting this was. Everyone knows he thinks BFS is better than anything else, and that linux is crap. Just like the linux developers felt that his code was crap.

                Comment


                • #38
                  Originally posted by Pawlerson View Post
                  It's enough to check server and HPC market share. They were dominated by UNIX, but Linux has wiped them out. It's such simple. The same when comes to stock exchanges which are one of the most demanding and critical workloads.
                  Nope, it's not enough. If you say it's slow or bloated, then the market share is not enough. The market share weighs a lot other factors too, mostly related to price. Then, you could argue "but they're all free". Except, support is not, and IT's are not. And it's far easier to find Linux IT's and Linux support companies than FreeBSD's or Solaris' ones. And easier usually implies cheaper, too. AND, it happens Solaris is not free for production use. So, it could have nothing to do with speed or memory footprint, as long as none of them are stupidly slow or bloated. So, facts, please.
                  I use Linux, mostly because I'm comfortable with it.

                  I don't have comments on anything else, just found this to be nonsense and had to answer.

                  Comment


                  • #39
                    Originally posted by Pawlerson View Post
                    I explained in another thread there are no antiBSD and antiSolaris trolls. It's impossible! It's like trolling against shit. Do you know any troll that trolls against shit? Furthermore, you're an antiIntelligence troll, is that ok? Why do you troll against intelligence?
                    I can't agree more. In fact there are a lot of people who agree with Pawlerson the BSD and Solaris = Shit:

                    http://aboutthebsds.wordpress.com/20.../bsd-vs-linux/

                    BTW, this blog is really interesting and enlightening. I'd recommend it to everyone, especially those pro-BSD fools.
                    Last edited by i386reaper; 07 May 2013, 08:29 AM.

                    Comment


                    • #40
                      Originally posted by i386reaper View Post
                      I can't agree more. In fact there are a lot of people who agree with Pawlerson the BSD and Solaris = Shit:

                      http://aboutthebsds.wordpress.com/20.../bsd-vs-linux/

                      BTW, this blog is really interesting and enlightening. I'd recommend it to everyone, especially those pro-BSD fools.
                      The situation becomes worse when BSD crashes while x11 is running. Once restarted, xorg hangs on startup and will continue to do so even when reinstalled. The only solution to fix this is to reformat the computer.
                      Yep, we have a really technologically minded person here.

                      Comment

                      Working...
                      X