Does being an edge lord actually give you better performance

[Goddard]

Beyond the additional features(unless you are running Gnome Shell jk) and security fixes being on the edge of software development such as running Arch, does it actually mean better performance?

[oiaohm]

https://www.phoronix.com/scan.php?page=article&item=15-way-linux&num=3

If you read over benchmarks you will find the answer is sometimes. In this 2016 benchmarks Debian 8.3 that is stable is at times in performance defeating the newer Debian testing and arch.

The thing about security fixes is there is no such thing as a free lunch. So fixing a security fault more often than not is cause of a performance loss.

[chithanh]

With mesa I would definitely say that newer = faster.
With compilers and kernels, the results are mixed. But generally that is more often true than not.

But as oiaohm noted, many security features will slow down your system to varying degrees. Plus, build configuration can have a much bigger impact than compiler version (like position independent code, SSP, optimization options, enabling or disabling kernel debugging, etc.).

[duby229]

Beyond the additional features(unless you are running Gnome Shell jk) and security fixes being on the edge of software development such as running Arch, does it actually mean better performance?

Well, I think the answer is going to be not really, I mean by that, yes for things like the open source graphics drivers, where hardware features are still being implemented. But for most other things its not really.

Arch is pretty good about building their binaries with the most important dependencies, but not with every single possible dependency. I think the answer is, if you want the fattest possible operating system then choose Ubuntu or Fedora. But if you want a lighter operating system then choose Arch and if you want something you can intimately control then choose Gentoo. But performance between them won't be any more than a few percent.

[HenryM]

The recent 10-way comparison prompted me to try Antergos, and compared to my Ubuntu 17.10 install, gaming performance at least is consistently 2-10% faster on Antergos - out-of-the-box - on a 1st-gen i7 860 & AMD 5770 over Ubuntu in a number of Mesa and kernel combos.

comparing static scenes on empty maps i got numbers like these:
TF2 frame rate: 106 vs 103, 123 vs 119
CS:GO frame rate: 107 vs 102, 138 vs 128

Antergos has now replaced Ubuntu (+a bunch of third-party PPAs) as my main OS and its been very nice so far, whereas 17.10 gave me some trouble.

I don't know why Ubuntu was slower. one thing I forgot to test is whether turning desktop icons off did anything.