Am I the only one doing it?
Posted by Alhumamjaddoa0@reddit | linux | View on Reddit | 12 comments
So.. I was looking at some people comparing Distros between each other, and they always show the benchmark scores or whatsoever. But I got used to use Blender first up whenever I try (live test, no WM) a new distro and compare a lot of stuff : material (if it's a different PC), how much the distro use CPU/GPU/(V)RAM/FPS on start and so on. Then, I go to Blender and subdivide the default cube (it's laggy for some reason, so perfect for a stress test) and move the cursor/viewport/subdivided cube all around until it starts getting laggy with the real time rendering. I then look at how much triangles I'm rendering in real time and how much has changed with the material usage (RAM/CPU/GPU/etc.) This is a stress test I do based on my feeling (Am I fine being this slow after calculating so much?). I know it's not a scientific looking benchmark with quantifiable numbers, but at least, it's quick and easy.
By the way, if you find some mistakes in this long text, feel free to correct me. English is not my first language.
StmpunkistheWay@reddit
I've tried a number of different distros and for anyone seeing this that would like to try different distros, should check out Distrosea.com as an awesome tool for the Linux community as a whole since you can test drive many different distros without installing them on your own box. It helped me find what I like and donations are needed to keep it running at https://www.patreon.com/c/DistroSea/posts at any rate, I've found that it really comes down to 3 or 4 main distros. Ubuntu/Mint (I'm using Mint on my main box as I think it does many things better than Ubuntu but it's still based off of it), Fedora (running this on my laptop as it just seemed to have better tools right out of the box), and OpenSuse, maybe one other one. The rest all seem to have different flavors/packages of those three, at least for OS's using newer hardware, meaning 9th gen Intel/AMD or better.
I think when you're using older hardware, picking the right OS is much more important in how your system is going to run and there are a lot of options out there for this reason which is really a great thing to have.
But anyway, check out that distrosea site to check out the different distros, it's kind of a cool site to have for the community as a whole.
vivAnicc@reddit
The problem is that there is not much difference between distros, they are sll linux after all.
Sure, ubuntu will have more background processes running than vanilla arch, or GNOME might take more ram then a simple window manager, but it will not matter much.
q5sys@reddit
> The problem is that there is not much difference between distros, they are sll linux after all.
Sure if you ignore all the compile time options that different distros decide to use. There's a reason you cant take a debian package and run it natively on OpenSuse, or take a slackware package and run it natively on Ubuntu.
There are major differences under the hood. However most linux users never peer under the hood enough to recognize them. If you look at Debian vs Fedora, there are tons of massive differences right down to the userland tools and how to sysadmin the system. There's a lot that's similar... but if you've ever set up a LAMP stack on Debian and the gone and done it on Fedora, you'd realize that they're a lot that's not the same.
BigHeadTonyT@reddit
The only difference is what the package manager is, what they have in their repos, DEs, WMs, apps. How new the packages are, how well tested. What the distro installs out of the box, customizations, if any. Different philosophies, differing levels of utilities to help users. Or none at all. The Init system. Systemd, OpenRC, Dinit etc. The commands you run in terminal. I can't use Fedora commands in Debian. And so on.
Nothing different...the only common thing is the Kernel. And even that is customized most often.
SmileyBMM@reddit
Yeah I hate the people who say Distro doesn't matter, it absolutely does. Something like Void Linux is vastly different from something like Linux Mint.
KnowZeroX@reddit
It depends? For example, snaps can have overhead due to use of containers. Compiling in gentoo or using a version that has optimizations can also lead to performance differences. Then of course things like compositors, driver versions, kernel versions and default filesystems can make a difference.
MarzipanEven7336@reddit
You realize containers are no different than any other process, right? A container is defined literally by a single property on the process called CGroup. Aka there is zero fucking overhead whatsoever. Wait til you find out your whole system runs on containers, in the process group sense.
cgoldberg@reddit
You're running an arbitrary workload and interpreting the results based on inaccurate observations. If that makes you happy, that's great... but it is in no way a useful performance benchmark representative of any real world usage.
Besides, you aren't going to find any wild variation in performance between distros assuming you have them configured similarly.
Alhumamjaddoa0@reddit (OP)
I know that it's inaccurate and slightly arbitrary, but with my extra old laptop, there are some major differences from a distro to another.
JagerAntlerite7@reddit
Yes.
RoomyRoots@reddit
Yes, you are unique. Congratulations, or as the Japanese say one man dem toe.
blami@reddit
I don’t benchmark. I use my computer. I feel without Internet being crazy and making some webpages computationally complex as space rocket launch I would be fine with what we had in 00s performance wise.