There are some differences. Most Linux distros come in one of two general flavors: "Debian" based or "Red Hat" based. Ubuntu is similar to Debian while Centos is basically the free version of Red Hat's commercial Linux. Most of the differences are really only visible to system administrators: some files are in different places, configuring the system is different, they use different tools to do basically the same tasks. For day to day usage and for programming the differences are fairly small.
In terms of performance: there really shouldn't be any significant difference.
The biggest differences between the two is that 'Debian's repositories are MUCH bigger than CentOS's. For many things that Debian has in its default repo's, you need third-party repo's for CentOS. Often CentOS software versions are even older than Debian Stable versions. When running on non-critical systems, I have always used Debian Testing (also because that's basically a rolling distro).
I have to admit that the last time I tried anything else but a Debian-based distro has been several years ago, so the above may have changed.
With regard to benchmarks and speed, they should be similar; any differences can probably be accounted to different kernels (or different compilations thereof) than to the distro per se.
(I'm sure Jon just omitted this because they're the minor players within the Linux landscape, but there are also distro's that are not Debian and not Red Hat, such as Arch and Gentoo, and the derivatives thereof. And then there are some even smaller than those.)
Did some of you guys do some benchmark Ubuntu vs Centos?
Wondering if there is a significant difference
This question doesn't make sense. It's like asking whether installing Microsoft Office or Libre Office on your computer make it faster for chess, when neither of these applications are open and running.
So you must a noob. Nothing wrong with that. We've all been there. Start with Ubuntu. Eventually, if you want to muscle up your Linux game, use Arch.
But, CentOs? Heaven forbid... Why would anyone *choose* to use that ?
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.
It looks like Clear Linux had the best benchmarks overall, but they are all sourcing basically the same kernel, so there are just details in terms of how it is compiled and configured. Your hardware makes much more difference to overall performance than the OS layer does.
Apparently, it is an Intel project. But I'm not really interested because it has its own package manager and other quirks. Personally I want something that produces stable releases and has a large user community and uses established standard tools. I am using the Ubuntu LTS releases but there are other reasonable choices.
Remember that the kernel performance only comes into play for programs that are actually using OS functions like file operations, thread generation or dynamic memory allocation. With file operations, that could be EGTBs, but the small performance difference won't yield Elo. The others shouldn't be done during engine search anyway. What remains is the distro configuration, i.e. how many potentially unnecessary background processes are running and competing for computing time.
lucasart wrote: ↑Sun Nov 15, 2020 1:50 pm
But, CentOs? Heaven forbid... Why would anyone *choose* to use that ?
For example because it's much more stable.
You seems to have never ran a server under heave workload.
For enthusiast workstation Ubuntu is ofc much more comfortable.