Announcement

Collapse
No announcement yet.

BFS Scheduler Benchmarks

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

  • #51
    I think the kernel devs should take inspiration from Apple and devote just one release to code auditing, bug fixing, and performance improvements.

    Comment


    • #52
      Originally posted by kebabbert View Post
      "[Linux] is terrible," De Raadt says. "Everyone is using it, and they don't realize how bad it is. And the Linux people will just stick with it and add to it rather than stepping back and saying, 'This is garbage and we should fix it.'"
      Sigh...

      If we don't realize how bad it is then it cannot be that bad, can it?

      Same article, IT company CEO, PhD Comp Sci: "You know what I found? Right in the [Linux] kernel, in the heart of the operating system, I found a developer's comment that said, 'Does this belong here?' "Lok says. "What kind of confidence does that inspire? Right then I knew it was time to switch [OS]."
      I wonder how this guy got his PhD in computer science. Oh right, that's "Computer Science" not "Software Engineering". Because he obviously has no idea about software engineering at all.

      He might as well be saying that doctors use knives to operate on their patients (what kind of confidence does that inspire?), so he'd rather pray instead of going to the doctor.

      Read the "Old New Thing" for some juicy tidbits on the windows source code and the kinds of comments *that* contains. If you don't have any softeng experience, you'd run away screaming.

      Comment


      • #53
        Originally posted by BlackStar View Post
        If we don't realize how bad it[Linux] is then it cannot be that bad, can it?
        Yes you are right. Linux can not be bad nor bloated, despite Linus T says so. Even Linus T is wrong on this? Cool.

        And Windows can not be bad, nor bloated because so many people uses Windows all the time and thinks it is the best OS in the world. 90% of all computers run Windows, then Windows can not be bad, nor bloated.

        11 millions line of code for one simple KERNEL. That is huge. It is funny you dont agree with Linus T nor with me. Entire Windows NT was 10 millions LoC. Solaris is much smaller than bloated Linux. Every OS is much smaller than Linux. Maybe except Vista.

        Linux gets slower for each release. They have measured it, and they see that Linux gets a few percent slower for each new release. You want to read about this? Shall I post links, so you can see I am speaking the truth?


        Someone just wrote recently in another forum, when we discussed this claim from Linus T, that Linux is bloated:
        "I have been grumbling for the last week about breaking compatibility. That needs to be avoided. Just last saturday, I had to roll back a new Centos environment from 2.6.30 to 2.6.24 because the kernel dropped a function call after 2.6.27 that a library I use needs. The developers of the library have not made an upgrade available, and I can't switch at this time.

        I personally have not noticed any significant improvements in the kernel for quite a few releases now. I actually try to avoid upgrading my kernel on my workstation, but my package manager makes this into a royal PITA because I don't have an option to say "don't bug me about this...I don't want to."


        Someone else writes:
        "It's true, Linux is in danger of becoming not just bloated but unstable and unusable (far worse than just bloated)."


        Unix_epoch writes something good. He thinks Linus T should release versions where they fix problems. That would be very good for Linux.
        Last edited by kebabbert; 24 September 2009, 05:32 AM.

        Comment


        • #54
          Originally posted by kebabbert View Post
          Yes you are right. Linux can not be bad nor bloated, despite Linus T says so. Even Linus T is wrong on this? Cool.
          Linus' definition of "bad" and "bloaty" comes from a software engineering perspective and is grounded on the technical merits of the kernel. Yours or mine definition of bloat comes from a user experience perspective, simply because we don't have the technical knowledge to judge the 11MLoC kernel.

          Do you see the difference there? If the OS works for a user, is stable and fast, it cannot be bad by definition. This is completely orthogonal as to whether the kernel (or any other piece of software) could be better engineered.

          11 millions line of code for one simple KERNEL. That is huge. It is funny you dont agree with Linus T nor with me. Entire Windows NT was 10 millions LoC. Solaris is much smaller than bloated Linux. Every OS is much smaller than Linux. Maybe except Vista.
          And yet I can install Linux on a 2GB USB stick, carry it with me and boot it on whichever computer supports USB boot. Neither Solaris nor Windows can do that.

          Even worse, Windows and Solaris don't support my graphics card, TV card or even my freaking sound card out of the box. Windows doesn't run on ARM, MIPS, PowerPC and a host of other architectures supported by Linux. Do you think all this support comes out of thin air?

          Newsflash: all these features take raw, hard code. In fact, they take 11MLoC which provide the best device support bar none.

          Simply put, you don't have the necessary background to judge the Linux kernel on its technical merits. Neither do I and nor do 99.9999% of the people on this planet (if I counted the 9s correctly). Linus is one of those people who has the background - for us, it is exceedingly easy to misinterpret the meaning of what he said.

          Linux gets slower for each release. They have measured it, and they see that Linux gets a few percent slower for each new release. You want to read about this? Shall I post links, so you can see I am speaking the truth?
          Yet my user experience is getting better with every release. Faster boot times, improved hard disk performance (EXT4), better video drivers, better sound support, KMS.

          Frankly, I don't care if .31 compiles Apache 1.1 seconds slower than .23 and neither should you.

          I personally have not noticed any significant improvements in the kernel for quite a few releases now. I actually try to avoid upgrading my kernel on my workstation, but my package manager makes this into a royal PITA because I don't have an option to say "don't bug me about this...I don't want to."
          Off-hand: KMS, EXT4, faster boot times.

          Someone else writes:
          "It's true, Linux is in danger of becoming not just bloated but unstable and unusable (far worse than just bloated)."
          Stop complaining and start coding then.

          Unix_epoch writes something good. He thinks Linus T should release versions where they fix problems. That would be very good for Linux.
          Every new release fixes problems. Damn, every *point* release fixes problems (2.6.28-14 vs 2.6.28-15).
          Last edited by BlackStar; 24 September 2009, 06:14 AM.

          Comment


          • #55
            Originally posted by kebabbert View Post
            Yes you are right. Linux can not be bad nor bloated, despite Linus T says so. Even Linus T is wrong on this? Cool.

            And Windows can not be bad, nor bloated because so many people uses Windows all the time and thinks it is the best OS in the world. 90% of all computers run Windows, then Windows can not be bad, nor bloated.

            11 millions line of code for one simple KERNEL. That is huge. It is funny you dont agree with Linus T nor with me. Entire Windows NT was 10 millions LoC. Solaris is much smaller than bloated Linux. Every OS is much smaller than Linux. Maybe except Vista.
            It's about archive size moron. Solaris is much slower and it's really bloated (no archive, but Slowlaris itself). They use Windows mainly, because MS has monopoly (and probably more then 60% users have no idea other systems exist). You're so incredibly stupid or just trolling. Passed mensa for idiots test? You should have master in this. Where such idiots came from? It is funny you're so stupid. Puppy Linux is much smaler then slowlaris, bsd*, Windowses, OS X...

            Linux gets slower for each release. They have measured it, and they see that Linux gets a few percent slower for each new release. You want to read about this? Shall I post links, so you can see I am speaking the truth?
            Intel measured this, but we know nothing about this (microbenchmark?). What we know is some new kernels are even 25% faster then previous. It seems nobody cared to measure some other OS, so who knows? 50% drop in some cases?

            Linus T? You sound like a child who feels inferior.

            Unix_epoch writes something good. He thinks Linus T should release versions where they fix problems. That would be very good for Linux.
            And they don't fix problems already? This is your logic? xd
            Last edited by kraftman; 24 September 2009, 01:58 PM.

            Comment


            • #56
              Originally posted by BlackStar View Post
              Linus' definition of "bad" and "bloaty" comes from a software engineering perspective and is grounded on the technical merits of the kernel. Yours or mine definition of bloat comes from a user experience perspective, simply because we don't have the technical knowledge to judge the 11MLoC kernel.
              Seriously, no one needs technical knowledge to know that 11 M LoC kernel is bloated, just common sense. Obviously some people disagree with me, Linus T and common sense.

              Originally posted by BlackStar View Post
              Do you see the difference there? If the OS works for a user, is stable and fast, it cannot be bad by definition. This is completely orthogonal as to whether the kernel (or any other piece of software) could be better engineered.
              But there are lots of testimonies contradicting this. Shall I post again? Maybe you didnt read?

              "I have been grumbling for the last week about breaking compatibility. That needs to be avoided. Just last saturday, I had to roll back a new Centos environment from 2.6.30 to 2.6.24 because the kernel dropped a function call after 2.6.27 that a library I use needs. The developers of the library have not made an upgrade available, and I can't switch at this time.

              I personally have not noticed any significant improvements in the kernel for quite a few releases now. I actually try to avoid upgrading my kernel on my workstation, but my package manager makes this into a royal PITA because I don't have an option to say "don't bug me about this...I don't want to."

              Besides, if the OS is good enough for a user, it doesnt prove that the OS is good at all. It just proves that the OS is good enough. Windows works good for many users, but it doesnt prove that Windows is good by definition. You have got the definition wrong.

              Originally posted by BlackStar View Post
              And yet I can install Linux on a 2GB USB stick, carry it with me and boot it on whichever computer supports USB boot. Neither Solaris nor Windows can do that.
              Ah! Then surely Linux must be the best OS, and there are no complaints! If you can install an OS to a 2GB USB stick and boot, then it is a proof it is good OS. (Besides you can do that with Solaris also, maybe you should check up your claims first? Post links supporting your claims, just as I do, upon request? Instead of making things up?)

              Originally posted by BlackStar View Post
              Even worse, Windows and Solaris don't support my graphics card, TV card or even my freaking sound card out of the box. Windows doesn't run on ARM, MIPS, PowerPC and a host of other architectures supported by Linux. Do you think all this support comes out of thin air?
              Ah, Windows/Solaris doesnt fully support your HW? Then surely Windows/Solaris must be bader than Linux: unstable, slow and buggy. Thanks for your convincing argument of Windows/Solaris being bader than Linux.

              So, if I can find one piece of HW that Linux doesnt support, then it is convincing argument that Linux is bad: unstable, slow and buggy as well? Yes?

              Originally posted by BlackStar View Post
              Simply put, you don't have the necessary background to judge the Linux kernel on its technical merits. Neither do I and nor do 99.9999% of the people on this planet (if I counted the 9s correctly). Linus is one of those people who has the background - for us, it is exceedingly easy to misinterpret the meaning of what he said.
              I dont need the necessary knowledge or whatever, if Linus T says Linux is bloated, am I not allowed to claim that "Linux is bloated"? Do I lie when I say so? Should I stop saying to people that "Linux is bloated, even Linus T thinks so"?



              The funny thing is that lots of people has said this for years, and theyve been called Troll. And now Linus T says the same. Who is the Troll now? The ones who were correct all the time, or the ones falsely accusing others of being Troll? I say that the ones who falsely accuse others are Troll. And the correct ones are not troll. What do you say?

              Originally posted by BlackStar View Post
              Frankly, I don't care if .31 compiles Apache 1.1 seconds slower than .23 and neither should you.
              Fine if you dont care that performance gets slower, but I care about performance. Vista was slower than WinXP, and it got lots of attention on bloat. Linux does the same, and everyone defends it. When MS does bloat it is bad, when Linus T does it, it is good and a feature. Not a bug.

              Originally posted by BlackStar View Post
              Off-hand: KMS, EXT4, faster boot times.
              On-hand: slowing performance, more bloat, unstable ABIs, etc. How hot is Ext4 for instance? Not really. Why not invent some new smoking hot tech instead of just copying other OSes?


              Originally posted by BlackStar View Post
              Stop complaining and start coding then.
              I wouldnt want to code something bloated, would I? Soon my fine code would be swapped out for some buggy code. I prefer good projects with lean code.


              Originally posted by BlackStar View Post
              Every new release fixes problems. Damn, every *point* release fixes problems (2.6.28-14 vs 2.6.28-15).
              You missed the point. Epoch_Unix thinks that Linus T should try to release versions where just bugs are fixed, and no new func is added. That is a good idea. Even Andrew Morton says so, you want to read, so you see I speak the truth? Shall I post Andrew Mortons saying here?

              Comment


              • #57
                Originally posted by kebabbert View Post
                On-hand: slowing performance, more bloat, unstable ABIs, etc. How hot is Ext4 for instance? Not really. Why not invent some new smoking hot tech instead of just copying other OSes?
                Show us the facts and figures behind this assertion. Meaningful ones please, not ".31 IZ 0.0023 SECONDS SLOWER THAN .11 OMG!!2".

                Comment


                • #58
                  Any chances you guys go and benchmark BFS to get some more content to this thread?

                  Comment


                  • #59
                    "On-hand: slowing performance, more bloat, unstable ABIs, etc. How hot is Ext4 for instance? Not really. Why not invent some new smoking hot tech instead of just copying other OSes?"
                    Originally posted by Ant P. View Post
                    Show us the facts and figures behind this assertion. Meaningful ones please, not ".31 IZ 0.0023 SECONDS SLOWER THAN .11 OMG!!2".
                    I claim five things here. Slowing performance. More bloat. Unstable ABI. Ext4 is not hot. And that Linux just copies from other OSes, for instance from Solaris, and never invents new smoking hot tech

                    1. Slowing performance:
                    12% lower performance is a significant number.

                    "Citing an internal Intel study that tracked kernel releases, Bottomley said Linux performance had dropped about two per centage points at every release, for a cumulative drop of about 12 per cent over the last ten releases."


                    2. More bloat
                    Same link. Linus T says Linux gets more bloated.


                    3. Unstable ABI
                    Everyone knows this.


                    4. ext4 not hot
                    True. ZFS is hot. And BTRFS is hot (because it is a copy of something hot: ZFS)


                    5. Linux just copies and never invents new smoking hot tech.
                    BTRFS is a ZFS wannabe. Systemtap is a DTrace wannabe. All I see is wannabes. I havent heard of some new hot Linux tech that excites people.





                    VESA,
                    Yes you are true. Sorry for this off topic. I will stop this off topic now. (Too bad people will not believe that Linux has it's flaws until Linus T says so. And even then, people will not believe it. They continue to accuse others of being Troll, even when Linus T confirms. They never think themselves. As someone said:

                    "It never fails to amaze me how fanboyism can kill rational thought in otherwise intelligent human beings.")



                    On topic, this BFS scheduler seems really nice I think! I always liked Con Kolivas and his ideas. Come on, who can learn to kernel hack from themselves? It is mighty impressive, as he is trained as a doctor and not a programmer. I am impressed by him. Too bad others are not.

                    Comment


                    • #60
                      Great! Kebabbert troll is here. You can read his trolling on OSNews in every article that involves Linux.http://www.osnews.com/story/22207/Li...rocess_Working So please do not feed the troll. He has his head stuck so badly in Jonathan Schwartzs ass so he can't stop.

                      Comment

                      Working...
                      X