I currently have linux 6.14.1.arch1-1, linux-zen 6.14.1.zen1-1, 6.14.2-1-cachyos, and linux-lts 6.12.22-1 installed on my system. I use the Zen kernel as my daily driver.
If/when any issues arise on my EOS install, I log into the others to see if the problem persists with either of those.
So, how do you use the Linux kernel? Do you have just one set-it-and-forget-it kernel, or do you have multiple kernels installed?
I use “linux” and as a backup I have “linux-lts”. I have used “linux-zen” in the past but it really never noticed a difference between “linux” and “linux-zen”, so I ended up just switching back to the normal kernel.
I have linux and linux-lts installed. I usually run linux and keep the lts as a fallback.
I usually just use the latest current kernel. Set and forget usually but i do have the lts on my Nvidia desktop also. One never knows when Nvidia will cause some issue but it’s been good for quite sometime on Kde Wayland! 
According to this specific benchmarking https://www.phoronix.com/review/arch-linux-kernels-2023/8 the “regular” latest stable performs overall better than zen 
Anyways, since for my use case the performance enhancement that zen brings to the table is irrelevant so for me islinux and linux-lts as auxiliary.
I’ve never been that fond of zen. I usually just stick to the regular kernel, lts if I’m having any issues.
I have the same, except I removed the mainline kernel. On my Virtualbox/VMware systems I run the LTS kernel.
Just the regular Linux kernel, never had a issue yet, but I know logically I should probably install the LTS as a just in case as a backup… Maybe when I get around to installing on my Win 10 desktop which is more important to me that it’s working, I’ll install both to be safe
Simply linux normally, and linux-lts as backup.
Works for me.
Depends on the definition of “performs”. Phoronix exclusively tests throughput, and you would expect a kernel that is tuned for responsiveness perform worse on throughput.
That’s why I wanted to save my skin by adding:
Somewhat apologetically on my part, Larabel adds:
… the Zen kernel performing well in gaming benchmarks and other select scenarios.
I usually install linux-lts as fallback, then a current kernel from the distro (linux-zen), and on the desktop a linux-tkg build to specifically get me all the bells and whistles for that CPU.
I also have lts, arch1 and zen installed and use zen mostly.
I also use Linux and Linux-LTS kernels. I’m glad as of today for this too, my LTS kernel got messed up somehow during an update from this morning, was able to boot into the standard Linux kernel, address the issue relatively quickly, and keep enjoying the rest of my day. It would’ve taken more time if I didn’t have another kernel installed.
It seems to depend on the game and the workload and the last benchmarks I found are quite old .
Currently running a zen kernel and installed lts as the backup. Used to compile a tkg kernel (PDS/Bore) to optimize for my hardware, but stopped doing that because of the long (and more and more frequent) compiling process and tons of disk I/O.
It always depends on so many factors when benchmarking hardware, systems etc. therfore no single statement should be taken at face value. The express perhaps more broadly a “tendency”.
It’s the same throughput fallacy. Phoronix is useless when it comes to these benchmarks.
Of course people like “high fps” in games. But what are the 1% lows for those fps, does it show a huge stutter every two seconds? Can you still game with a constant framerate while you have a browser, OBS and discord open in the background?
Even the scx gaming schedulers like lavd have no problem sacrificing a few average fps if it improves the 1% lows.
My normal Linux Kernel Usage is Daily 