Announcement

Collapse
No announcement yet.

A Fix For The Severe Linux Performance Regression Spotted By Torvalds

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

  • #31
    Hmm... bug gets written, bug gets fixed, world moves on.

    Except on the website fora which are filled with software developers who would never have written the bug in the first place, and how "other people testing the software" should have found the issue.

    Comment


    • #32
      Originally posted by OneTimeShot View Post
      Hmm... bug gets written, bug gets fixed, world moves on.

      Except on the website fora which are filled with software developers who would never have written the bug in the first place, and how "other people testing the software" should have found the issue.
      What's funny it's a bug in development tree where bugs are something common and expected. It seems Linux haters can't stand Linux superiority. They are definitely not programmers or testers. If the bug was in stable (as it also happens) then a disappointment would be justified.
      Last edited by Volta; 15 January 2024, 05:17 AM.

      Comment


      • #33
        Originally posted by Danny3 View Post
        Too bad that the Linux foundation gets a lot of money from companies and yet don't care to make sure that the most important man in the kernel's development doesn't have the maximum reliability, like solar panels, wind mills, batteries, UPSes (in case the house batteries have some problems) a second fixed internet subscription, a 3G-4G modem, etc!

        Every time I'm wondering what they are doing with the tons of money as they clearly don't help Linux adoption, don't help the desktop environments, don't help Phoronix to have what it needs for continuous regressions testing and now I see that they don't help even Linus.
        It seems to me that this is another organization like Mozilla that just wastes money because of bad leadership.
        Here at least the market share of Linux is not going down like the one of Firefox as there are no good alternatives, but it's clearly not helping advance much.

        I would fire those people and hire new ones, at the top people that care more about Linux and open source software, let's say Michael from Phoronix and Richard Stallman from FSF or wherever he is now.

        I wonder if Microsoft knew about this when it decided to become a member of it.
        Because if it knew that this organization does mostly nothing and does nothing or very little to improve Linux's adoption, I think it finally makes sense why they would join it as it would not help their competition much, but it would be a great PR move making them look like the good guys.
        Linus Benedict literally has tens of millions of dollars (RedHat stock). Maybe he could have solved the issue himself.
        Last edited by avis; 15 January 2024, 07:30 AM.

        Comment


        • #34
          Originally posted by airminer View Post

          Oh come on now. A week's delay in the kernel merge cycle every couple of years isn't the kind of disaster to justify forcing Linus to turn his private home into a bunker.

          Not to mention the solutions you are proposing are entirely unsuited to handle prolonged outages caused by harsh winder storms.

          (Minimal sunlight in the northern winter + snowed-in rooftop panels, reduced battery capacity due to low temperature, a workstation running power-intensive tasks like kernel compilation constantly for regression testing etc.)

          If the occasional week's delay is such a deal-breaker, a starlink antenna + a diesel generator would suffice.
          A gennerac and a starlink backup would be really good for a large number of reasons. Not just for hitting merge windows, but also just for the health and safety of the man and his family. A good gennerac costs about $6000 (plus install) and is a one-time fee, until it kicks into use and uses natural gas to power itself. The Linux Foundation should be able to EASILY afford that upgrade to Linus's home.

          Comment


          • #35
            Originally posted by Volta View Post

            Only if mentioned people are lamers who have no clue about operating systems. I already provided you methodology. Try to challenge me if you want. Let's focus on broken by design windows, joke os macos and legacy bsd (slowlaris is already out of the game). I'm waiting. I hope you're aware there must be reference.
            It is not for me to prove a negative: it is up to you to show how "Linux is the cleanest and nearly perfectly designed operating system." - think of it as a student essay where you need to argue your case. Picking holes in other operating systems is not sufficient. I expect you to reference Dijkstra and Hoare, among others. You will need to define 'operating system', how you define 'cleanness', and your grading system that allows you to make a statement about proximity to perfection. Extraordinary statements require extraordinary proof.

            If you can't do this, you have not proven your case.

            Comment


            • #36
              Originally posted by Danny3 View Post
              Too bad that the Linux foundation gets a lot of money from companies and yet don't care to make sure that the most important man in the kernel's development...
              The biggest problem, apparently, is that this is technically a one man show. I'm pretty sure there are plans if something happens to Mr. Torvalds, but it's obviously not well know to the world.

              Comment


              • #37
                Originally posted by sophisticles View Post

                You do know that the Linux Foundation, which Linus technically owns, pays Linux 1.5 million a year and that Linus has a net worth of about 50 million dollars?

                The real question is why he never bothered to get these things.
                You are quoting unsubstantiated rumours from the Internet. He does not own the Linux Foundation, he is not even among the executives nor on the board. And from the latest tax report from the foundation he was given $619K as a Fellow so far from $1.5M

                Comment


                • #38
                  Originally posted by Old Grouch View Post
                  blah, blah, blah
                  I already provided my methodology. However, I fully understand why you're ignoring it. I also have other tools for measure like time needed to fix a security hole and my favorite one: performance. However, I can show you some interesting numbers:





                  As shown in pasted links Linux quality is ways above FreeBSD. Linux quality is much better than average Open Source software while FreeBSD is exactly opposite. That's not all. Bellow is university paper from 2006 (please ignore Tanenbaum's microkernel propaganda):



                  the article states:

                  One study of software reliability showed that code contains 6-16 bugs per 1000 lines of executable code [1] while a different one [7] put the fault density at 2-75 bugs per 1000 lines of executable code, depending on module size. Using a conservative estimate of 6 bugs per 1000 lines of code the Linux kernel probably has something like 15,000 bugs; Windows has as at least double that. To make matters worse, typically about 70% of the operating system consists of device drivers, and they have error rates 3 to 7 times higher than ordinary code [2], so the above guesses are probably gross underestimates. Clearly, finding and correcting all these bugs is simply not feasible, and bug fixes frequently introduce new bugs.​
                  What can we learn from this? The current quality of Linux is much higher than Andrew's estimates from 2006. Another important thing: Linux drivers (especially GPU) are of the highest quality. Nowhere near Tanenbaum's estimate. How is this even possible when Linux is many times bigger than it was years ago? Rhetorical question. Oh, and Tanenbaum was clueless as usual.
                  Last edited by Volta; 15 January 2024, 09:21 AM.

                  Comment


                  • #39
                    Originally posted by Volta View Post
                    What can we learn from this? The current quality of Linux is much higher than Andrew's estimates from 2006. Another important thing: Linux drivers (especially GPU) are of the highest quality. Nowhere near Tanenbaum's estimate. How is this even possible when Linux is many times bigger than it was years ago? Rhetorical question. Oh, and Tanenbaum was clueless as usual.
                    I acknowledge your attempt at diversion. As an essay, I'd grade it at about 𝛾-. You've completely failed to support your assertion that "Linux is the cleanest and nearly perfectly designed operating system.". I'm jolly glad you are not one of my students. You probably are too.

                    Please don't waste people's time by making wild, unevidenced claims. 'Fewer bugs' does not mean 'better designed'.

                    Comment


                    • #40
                      Originally posted by Old Grouch View Post

                      I acknowledge your attempt at diversion. As an essay, I'd grade it at about 𝛾-. You've completely failed to support your assertion that "Linux is the cleanest and nearly perfectly designed operating system.". I'm jolly glad you are not one of my students. You probably are too.

                      Please don't waste people's time by making wild, unevidenced claims. 'Fewer bugs' does not mean 'better designed'.
                      I mean, the same kernel can be built to run anything from a phone to a supercomputer. What other kernel can boast the same track record?
                      (Yes, I know, that still doesn't touch on the design itself. But if it's that adaptable, the design can't suck, can it?)

                      Comment

                      Working...
                      X