In case you haven’t seen it, the sibling project to the linux-zen kernel has just been updated from 5.7.19 to 5.8.5 with the added bonus of patches from Clear Linux:
After running it for the past hour, this kernel appears to have issues on my system. I was getting a lot of small freezes (which seemed to be related to having an NFS share mounted), and eventually network access stopped and then the system hard locked. So, it might have exposed a bug in the network driver, or some other kernel patch or config tweak has affected the network stack.
Although, it did also start freezing after I opened (and closed) VirtualBox, so it might be that. More testing needed.
What would you say is the main selling point over this kernel vs say the default Arch kernel and the zen kernel? I tried out zen and didn’t notice any appreciable differences so I just went back to default
Clear Linux patches are supposed to offer significant performance gains on modern hardware (check e.g. Phoronix tests). MuQSS scheduler is supposed to offer a fairer and more responsive desktop CPU scheduler than the kernel default.
However, the Clear Linux and MuQSS patches are new (and pretty significant).
So, the latest Liquorix is essentially “Zen on steroids”.
Not to linux unless they are included in mainline. As for mainline, probably not. Reading the latest entry on CK’s blog it looks like they have a very different approach:
According to Linus, the kernel will only ever have one single CPU scheduler.
The Clear Linux patches IMO only make sense on Intel machines with a few exceptions.
Personally I prefer PDS/ProjectC over MuQSS, though I did have issues with both - wrong CPU utilization readings with MuQSS, and VBox crashes with PDS/ProjectC.
I’m now back to CFS with a few tweaks that Zen/Liquorix also have.
By the way, I’ll soon push 5.9-rc. Since people were not happy with the vd name, I renamed it to klitoris - “Kernel with Little Improvements and Tiny Optimisations Resulting in Increased Speed”.
Oh eff off, torvic.
That was actually a few years ago, it might long be fixed. Con said that it was a cosmetical issue only, but I found it quite annoying and haven’t touched MuQSS since.
Does anyone running a MuQSS kernel notice any strange CPU utilisation readings, e.g. if the system is completely idle, does it show ~ 0,x% utilisation?
I have just loaded the build of lqx - and while it could be my imagination - my system just upped itself a bit.
Nvidia is never going to be a friend of mine - they produce great hardware - but I’ll never be Nvidia evangelist.
Now my system i9-9900K pure Intel with this build of lqx - I tell myself I can see and feel the difference.
Firefox is up like a fly on a pile of …
My virtual windows is like wise-I have never seen my VM’s up so fast.
Thank you making me aware of this - it is the best thing happened the last couple of days where my system has bugged my days - I ripped my Nvidia out in frustration.