Monday, January 15, 2018

PDS 0.98i release

PDS 0.98i is released with the following changes


1. Revert "pds: Remove yield support."
2. Set default yeild_type to 1 and remove yield_to() support.
3. 16ms dispersed balance interval.

This is the last release on 4.14 kernel and the first release in 2018, which brings back the yield_type support, users needs to adjust the value to their own requirement, now default is 1. And there is balance overhead cut-off feature which will help in all kinds of workload.

Enjoy PDS 0.98i for v4.14 kernel, :)

code are available at
https://bitbucket.org/alfredchen/linux-gc/commits/branch/linux-4.14.y-pds
and also
https://github.com/cchalpha/linux-gc/commits/linux-4.14.y-pds

 All-in-one patch is available too.

7 comments:

  1. @all
    For whom have downloaded the all-in-one patch before this post, I'm very sorry that it was generated from a test branch and I have upload an up-to-date one, please re-download it.

    ReplyDelete
  2. @Alfred & @Oleksandr:
    Many thanks for providing this update!
    I'm running it from within Oleksandr's v4.14-pf9 patchset (4.14.14 based) which includes all the recent BFQ fixes as well.

    Works well for more than 27h. Please keep up this good work!

    BR, Manuel Krause

    ReplyDelete
  3. [OFF-TOPIC] Regarding Meltdown and Spectre:
    I really hope that these problems get fully resolved soon.
    ATM. there keep coming-in patches for 4.14.

    By coincidence I've found a tool to check for vulnerabilities (no wish/ need to advertise it):
    https://github.com/speed47/spectre-meltdown-checker
    and one probably wants to check the plain call vs. one with the --no-sysfs --verbose option, that shows prominent differences at least on my system.

    Regarding the compiler, although with "rolling-release" openSUSE Tumbleweed, I needed to go to an OBS Factory GCC, to get retpoline support within.
    Additionally, it seems like intel indeed didn't fix all affected cpus microcode, as reported by some media, so these would come later. But when?

    Worst way to interpret the results for my system would be: My Core2duo is not affected, and PTI would only produce a performance impact.

    INSTEAD we should keep our microcode/ ucode and GCC and kernel as fresh as possible.


    BR, Manuel Krause

    ReplyDelete
    Replies
    1. [OFF-TOPIC] Regarding Meltdown and Spectre:
      Weird practice, that INTEL withdrew their January 2018 microcode within the last few days. Last known good is now named 20171117.
      The Checker tool had been updated many times inbetween. It's purpose is to only show whether the fixes/ settings work.
      My gcc 7.2.1 20180122 [gcc-7-branch revision 256939] from openSUSE Factory now supports retpoline.

      Still, I don't get decided info, whether my system is affectable, at all.

      BR, Manuel Krause

      Delete
  4. For users still on 4.14.15/+, and have compilation issue, pls check out this issue and the fix patch provided by Holger Hoffstätte.

    ReplyDelete
    Replies
    1. The actual link is:
      https://bitbucket.org/alfredchen/linux-gc/downloads/pds-20180122-fix-delayacct-account-blkio-completion-on-the-correct-task.patch

      Thank you @Alfred, I had taken it from Oleksandr's first approach for kernel 4.15, what's in fact changing just the very same code, and it's working well for over 48h now. :-)

      BR, Manuel Krause

      Delete