Announcement

Collapse
No announcement yet.

A Fix Is On The Way For A Previously-Reported Linux 5.18 Performance Regression

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

  • A Fix Is On The Way For A Previously-Reported Linux 5.18 Performance Regression

    Phoronix: A Fix Is On The Way For A Previously-Reported Linux 5.18 Performance Regression

    Back during the Linux 5.18 merge window in March I wrote about a big NUMA benchmark performance regression I noticed and bisected. It turns out there has been a fix for it in patch form albeit I only noticed this weekend and now was able to successfully test and verify the fix. That fix is now working its way to the mainline kernel...

    https://www.phoronix.com/scan.php?pa...Regression-Fix

  • #2
    There's a reason games like MM are known as NES hard

    Seriously though, I hope the Linux Foundation or RHEL or some high level Linux organization is funding you for all your kernel regression finds. I know that I feel bad that my broke self hasn't paid for a Phoronix membership after all of these years I've been coming here in appreciation of all these finds.

    Comment


    • #3
      Originally posted by skeevy420 View Post
      Seriously though, I hope the Linux Foundation or RHEL or some high level Linux organization is funding you for all your kernel regression finds.
      Unfortunately, not...

      Separately, but speaking of RHEL, this week/end had found huge regression in RHEL9 / upstream... Though seems to be isolated to some particular EPYC CPUs in wonky state....In some workloads like a many times crazy slower regression. At least the RHEL engineers involved were quick to work through it after roping it in through AMD. Still have to find the time to write more about that one...

      https://lore.kernel.org/lkml/[email protected]/
      Michael Larabel
      https://www.michaellarabel.com/

      Comment


      • #4
        Originally posted by Michael View Post

        Unfortunately, not...
        That's sad to hear, but that's the assumption I had and why I gave you the heads up the other day about AMD wanting to hire someone to benchmark AOCC when I came across that post looking for an entry level AMD job (I don't those exist remotely). I assume that you have all the bits and pieces in place to run better tests than even AMD.

        Separately, but speaking of RHEL, this week/end had found huge regression in RHEL9 / upstream... Though seems to be isolated to some particular EPYC CPUs in wonky state....In some workloads like a many times crazy slower regression. At least the RHEL engineers involved were quick to work through it after roping it in through AMD. Still have to find the time to write more about that one...

        https://lore.kernel.org/lkml/[email protected]/
        I might not use their OS but I appreciate all that they do.

        Speaking of time, I'll just throw this out here for your consideration. If you have need of an editor, writer, forum moderator, graphical content generation, and someone who knows enough Bash to occasionally assist with the PTS I only need a hair over $1300 a month after taxes to make ends meet; that's $1760 a month pretax( $440/wk) which, coincidentally, is 40/hrs a week at minimum wage in my state (11/hr). While I don't know your revenue or what you can afford to pay, I do know the bare minimum I need to live on while focusing on Phoronix 100% and can offer that to you.

        Also in regards to time I totally understand, only from a different point of view. I don't do something like 80% of the things I'd like to do or comment about doing due to working low paying construction jobs and being too tired to think properly. Now that I'm 37, 38 in 5 months, I'm finding myself too tired for critical thinking until my 2nd or 3rd day off after doing labor and welding in the sun. That's why I haven't looked into submitting 8bitdo firmware to FWUPD yet or got around to forking Holoiso just to replace BTRFS+Ext4 with ZFS.

        FWIW, even I'm LOLing about the thought of someone like me who has to be in the top 10 for moderated posts asking for a job that could include being a moderator. For professionalism's sake I know that I'd have to retire this username because it is not an acceptable username (unless you are running a black market drug site). I still use it because 18 years ago my best friend and I picked it out to use for stuff and 11 years ago he killed himself and I've found it hard to retire it since that happened....in case any of y'all were wondering why I'm so old and am using that bad of a username.

        Comment

        Working...
        X