From what I recall the completely fair scheduler (CFS) used by default on most Linux systems has a lower average latency than the RT kernel. The RT kernel just gives you more consistency, hence the CFS having lower latency “on average”

So honestly for opening Firefox it’ll probably depends more on your SSD data rate, but in theory it’ll open faster on a “regular” distribution most of the time.

Real time is good for things like audio processing where having better guarantees that a process will get its share of the CPU is a benefit.

Hey man, I don’t want to discourage you, but this is one of those things where if you have to ask how to do something you’re probably not experienced enough to do it. That being said, as a learning opportunity even if you don’t make it far you’ll still learn a lot about how GPUs work.

I’d start by looking at any existing drivers you can find and see if you can document or find documentation for the commands fed to the GPU. From there you can look at the Mesa project for examples of converting Vulkan to instructions for specific processors and see if you can get it to all fit together for your project.

For newer APIs the kernels documentation is good:

https://www.kernel.org/doc/html/latest/

As for getting help, the #kernelnewbies channel on the OFTC IRC network is a good place to ask questions.

Have fun learning!

I’ve done some kernel development. I’d suggest using a modern system. Most of the basic stuff in the book is still accurate and there’s great documentation available to help find newer APIs. The kernel is always refactoring so learning to find newer info is part of the skills you’ll need to develop anyway.

I don’t think you really do anymore. I’d consider myself an experienced Linux user. I’ve been using it as a my desktop OS for over 20 years. I’ve also used Linux heavily through my career and am completely comfortable with the command line.

With recent installs of Fedora the only thing I use the command line for is the initial setup of the multimedia codecs. After that I haven’t been required to touch it.

I used to consider a terminal required to keep your desktop Linux system running. Now I look at is as an optional install for programmers.

That was a good read. I’d not really been sure of the differences between libadawaita and GTK were. It sounds like this frees up GTK to focus on being a cross platform GUI library, perhaps competing more directly with Qt. Meanwhile, libadawaita allows GNOME developers to keep leveraging GTK and tune it to their design guidelines.

I’ve only seen positive things come out of recent GNOME apps, but I wonder if the downside of GTK no longer embedding GNOME’s design language would be apps choosing to use GTK directly instead of libadawaita for better cross platform support. Will we end up with a less cohesive GNOME environment in the future?

I’m curious where you’re having issues. I’ve been able to use the little GNOME widget for setting up wireless connections for years.

Do you have an edge usecase that makes you drop back to using nmcli or is there a missing feature forcing you back to the ip/ifconfig commands?

No sarcasm, I’d just be interested in understanding your frustration a little better.

I recently purchased a laptop from System76. I’ve been very happy with it. You can get many of their models with coreboot used as the system firmware which is unique. I have been disappointed that they usually recommend installing open source, but not in-tree modules for getting things like keyboard backlight working. It feels a bit like they’re not a Linux laptop company but instead that they’re a Pop!_ OS laptop company.

Linux distributions have definitely standardized over the years. You get a kernel, systemd, network manager, Firefox, etc from basically every distro targeting desktops. Most will have different spins for the popular desktop environments as well.

From a purely technical perspective the main difference of distributions today is the package manager. Are you using pacman, apt, dnf, or something else? We know as users that while some of these different package managers have advantages and disadvantages they are all doing the same thing. You can get basically all the equivalent packages on each major distribution. I sometimes feel sad thinking about all the volunteer effort working in parallel, but not together to package the same software using different package managers. In many ways it’s duplicate effort that I wish could be spent in better ways.

Even package managers are beginning to converge. Flatpack is becoming extremely popular and is my current preferred way to add software to my system.

Leaving technicals behind the only major difference I see between distro today is their philosophy on how frequently to update and what to exclude. Does every package get a new update immediately when it’s rolled out upstream like Arch? Are we going to stick to older packages and only apply security/bug fix updates like Debian, or do something in between? Do we want to bend over backwards to make it easy to install Nvidia drivers or tell users we don’t support closed source software? Do we want to make it as easy as possible to install codecs or leave it to the wider community to figure that out on their own?

I don’t think there’s any right or wrong answer. Use what you enjoy!

I’m glad to see open source video editors are still evolving. I know Davinci Resolve has become popular with Linux YouTubers for serious work.

I’d be curious to know a video editors opinion of what’s missing. Is it stability, GPU acceleration, UI or a combination of all of the above? What would be needed for you to switch to Kdenlive for example?

I use Hexchat. It’s a fine GUI a client, simple and reliable. I use a ZNC bouncer so no need to keep a CLI client running 24/7.

Anything is possible, but I’d be surprised if it’s a driver issue. Many server vendors use tg3 based NICs as the onboard NIC integrated into the motherboard. The install base for that driver is huge.