I'm sososo scheptic about the 10% latency reduction.

Probably on arch you would have a better latency cause you will use the exact same low-level stack (apart for some pointless "tuning") and you will probably avoid some of their crapware putted on-top wasting resources.

If they minimize the crapware, the expectation could be to have the same latency you would have with arch.

If they really understand what they are doing with the low level optimizations (so rare in the sea of downstream distribution and something I'm so scheptic about) the latency reduction obtained could be in the order of 1%.

10% less latency across different machines just because kernel magic sound really a bullshit claim to me.

But if it works its okay, and if you come from a debian based probably you will see real better performance, but it would be just because its arch and there are less bullshit ontop then zorin-like distros.

Reply to this note

Please Login to reply.

Discussion

it's a different scheduler in the cachyos kernel

i picked it after searching for recommendations for a distro that gets the best compatibility and performance out of AMD hardware, so it's most likely not relevant to an intel based system, or nvidia

it's the most geeky linux installer i have ever seen, it opens up terminals showing the output of all the installation processes, AND the installer's own logger as well

also, i'm not budging from being an AMD maxi , i hate nvidia and i hate intel, it's probably not for nothing that Apple ditched intel for some samsung subsidiary or whoever makes their chips, if it has to be x86_64 i want Zen and Radeon

if its good its good, maybe they are competent and really optimized for amd in a meaningful way compared to plain arch. Still I would choose indipendence being upstream and eliminating this big possible point of failure on the software supply chain.

yeah, for me it's just the endless parade of come and gone kernels with patches and tweaks... the vanilla mainline is problematic in some cases for me from time to time, hardware support, dodgy scheduling