Announcement

Collapse
No announcement yet.

Eight Great Features Of Linux 5.8

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

  • Eight Great Features Of Linux 5.8

    Phoronix: Eight Great Features Of Linux 5.8

    If all goes well the Linux 5.8 kernel will be released as stable this weekend. Linus Torvalds last weekend expressed some uncertainty whether an extra release candidate would be required, but so far this week the kernel Git activity is light, thus for the moment at least is looking like 5.8 will be christened on Sunday...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Typo and redundancy:

    Originally posted by phoronix View Post
    AMD Radeon hardware also shas now soft recovery support for Navi, better handling of critical thermal faults for Radeon GPUs, and other updates.

    Comment


    • #3
      Might be too early to tell but I think all the scheduler improvements might be worth mentioning as a 5.8 improvement. Hard to quantify this change or see which kinds of workloads benefit and if it's power-saving, performance, latency, or general fairness.

      I'm always excited about the little improvements too.
      Last edited by Mitch; 31 July 2020, 02:32 PM.

      Comment


      • #4
        Does "The "Improved power-savings for systems with PCIe to PCI/PCI-X bridges." mean that PCI card into motherboard will allow more power saving?

        Comment


        • #5
          I wonder what is Oracle waiting for ???
          Every time I upgrade the kernel to a new version RC or final, it breaks Virtualbox and I have to wait for them until they release a new version of Vitualbox with the proper support for that kernel.
          Are they so poor that they cannot put their driver into the kernel itself to avoid this madness.
          AMD and Intel have put huge GPU drivers into the kernel and they cannot put a small driver in ?
          What's the catch, are they lazy or just bad ?

          Comment


          • #6
            The power drivers and thermal sensors should be mandatory for the kernel to make it compatible with CPUs released. As of today I find Windows10 with AMD chipset driver is handling the one year old ryzen9 3900x relatively better in terms of power consumption and temperature than my Arch installation with kernel 5.7. Hopefully 5.8 should fix some of that but the powercap patch for amd is yet to be merged in so have to wait for next kernel for that I suppose.

            Comment


            • #7
              Originally posted by Azrael5 View Post
              Does "The "Improved power-savings for systems with PCIe to PCI/PCI-X bridges." mean that PCI card into motherboard will allow more power saving?
              It works for motherboards that don't have native PCI/PCI-X in the chipset but have PCI/PCI-X ports through a "bridge" chip connected to PCIe bus.

              Aka motherboards from 2010 and newer with PCI/PCI-X, older ones should usually have native PCI/PCI-X slots

              Comment


              • #8
                Originally posted by Danny3 View Post
                Oracle
                ...
                What's the catch, are they lazy or just bad ?
                Why not both?

                Meanwhile, for VMWare there is a guy maintaining up-to-date kernel modules on github so you can use VMWare Workstation/Player with latest kernel, I've been using that for more than a year on OpenSUSE Tumbleweed now https://github.com/mkubecek/vmware-host-modules

                Comment


                • #9
                  Originally posted by starshipeleven View Post
                  Why not both?

                  Meanwhile, for VMWare there is a guy maintaining up-to-date kernel modules on github so you can use VMWare Workstation/Player with latest kernel, I've been using that for more than a year on OpenSUSE Tumbleweed now https://github.com/mkubecek/vmware-host-modules
                  Maybe it's both and they should've left Virtualbox to Sun.
                  Probably that way it would've had more improvements, maybe even PCIE passthrough.
                  We already saw what they did with MySQL. It's a shame to see that Virtualbox is going in the same direction.

                  I'm not sure the VMWare situation is much better.
                  It's good that somebody is mainting that repository, but I don't know what to do with it or if it's a real improvement over waiting for about a month to a new Virtualbox release.
                  Anywy, thanks for the information, maybe one day I will change to VMWare.

                  Comment


                  • #10
                    Originally posted by Danny3 View Post
                    I wonder what is Oracle waiting for ???
                    Every time I upgrade the kernel to a new version RC or final, it breaks Virtualbox and I have to wait for them until they release a new version of Vitualbox with the proper support for that kernel.
                    Are they so poor that they cannot put their driver into the kernel itself to avoid this madness.
                    AFAIK the kernel already has many Virtualbox drivers. Don't know what is still missing, but there are plenty of Virtualbox drivers there.

                    Comment

                    Working...
                    X