I did. Tbh I think there’s plenty of other people who do too. Once you’re comfy in an environment, you stick to it until you have enough reason to switch. And you’ll inevitably become a power-user if you solve enough problems, tinker enough, etc.
I’m on Arch, with Hyprland as my Window Manager. I use an RTX 3070.
For Wayland specifically, the driver was next to unusable for a while. I jumped ship from Windows in Sept. 2023. Beginning with driver 560 iirc, it got a lot better, plus their engineers pushed a lot of changes across the Wayland ecosystem to implement explicit sync support (a net positive, but before this, Nvidia was too stubborn to implement implicit sync, so bad screen tearing was unavoidable). Also there’s been a slow migration to using the GSP processor on newer cards. They claim it can improve performance, which may be true, but I also recently learned it helps them keep some more parts of their code closed-source, which is likely why it’s required to use the open source kernel modules.
At this point, though, it does feel very smooth and I can play games like The Finals at competitive framerates!
But relative to my performance under Windows, it’s still worse, mainly in average framerate. Like others have said, DX12 games seem to be hit hardest. I sometimes have to run lower settings to compensate. Also, if my VRAM gets filled, Xwayland apps all break, so I have to be very careful with higher quality texture quality especially.
Anyways, to answer your question, I think an average gamer doesn’t notice the degraded performance, without benchmarking or comparing framerates back to back— it still runs pretty smooth and framerates are still pretty high. If they aren’t happy with it, they’ll drop quality settings or resolution, just like they’d do under Windows.
You’re strawmanning their comment— I’d imagine they’d have the same, if not more, issues with snap.
Flatpak doesn’t integrate well with all systems. For me personally, on Arch, I have to update and store Flatpak versions of some dependencies, like proprietary Nvidia drivers, separately from the rest of my system and its package management system. And it does take up some space to store the runtime too.
Also Flatpaks may require some extra set up and/or workarounds due to their sandboxed environment. That’s not inherently bad and has some big security upsides, but it’s a consideration.
Also I don’t know how well it plays with immutable distros, but I’d imagine there may be similar integration issues there, too.
It’s still probably a lot easier for devs to have a consistent distribution format though, and they are typically more secure, so I’m not saying there’s not merits to only providing a Flatpak. Just pointing out that your reply here was misguided, imo.
I love Bluetuith! It’s my go-to for Bluetooth device management in my setup, since TUI applications tend to integrate really well into my existing theme, since I’ve set up my preferred terminal colors already.
I can imagine this would be a great option now for Windows users who already use the terminal frequently, like devs!
Did something similar to this with one of those “speed up silence” plugins. It was hard to notice cuz dialog was still 1x…
I’m sorry to hear you’re having trouble!
Regarding the vainfo thing, at least on Arch, the Nvidia vaapi driver is a separate package. I’m not sure what the equivalent is on bazzite, but that’s likely why vainfo doesn’t show anything. And you should still be able to use nvenc anyways, which sunshine will choose before vaapi.
Is moonlight able to see the computer on your network? And in sunshine’s web interface under the “troubleshoot” tab, there’s a log. Does it say anything that might help?
If you’re feeling adventurous, you could compile Marlin yourself. I did it to use a pin as an “off” signal for my fans, but ended up tinkering with options like linear advance. The set up isn’t too hard, but you need vscode.
I’ve played a small amount with it— I didn’t go far because I only have one graphics card. But my understanding is that you can set up the virtual machine to “capture” the graphics card when it is booted up, and “release” it when it’s done. That’s how some people can use a single graphics card, they just can’t see the Linux desktop while the vm is in use.
For switching games, make sure the drive the game is on is supported by both Windows and Linux. I imagine you’d need to unmount it on the host while the vm is open. I do this for some games with dual booting. Also, install the Windows version and force proton, otherwise you’ll have to install it twice.
For microphone, I assume there’s a solution like network: a way of forwarding the data to the VM in addition to being available on the host. I haven’t tinkered with this though.
For monitors, I’d look into Looking Glass. My understanding is that it can make a window to display the vm output. If you don’t mind the monitor being “tethered” to the gpu, though, I’d imagine any monitors attached on the gpu you pass would display the vm’s contents, while any on the host card will remain the same.
Best of luck!
That’s awesome! It’s really cool seeing low poly models realized.
I don’t think there’s a separate build with Nvidia patches now (at least on the AUR), as they’re included/fixed upstream. Still plenty of config and environment variables to set on Nvidia. And the new Nvidia beta driver, 555, should fix flickering in xwayland windows that was caused by a lack of explicit sync support. It’s been working well for me so far!
Linus Torvalds, the creator of the Linux kernel, owns the Linux trademark. He is currently the lead developer, and he approves and merges the code contributions from each of the “areas” (idk the right term for them) into the main branch for release versions.
The Linux kernel is distributed by The Linux Kernel Organization, a “California Public Benefit Corporation”. It’s “recognized by the IRS as a 501©3 private operating foundation”.
They are managed by The Linux Foundation which is “a 501©6 non-profit”.
It’s not “sold” to big corporations, because none of these companies are publicly traded.
But companies do use Linux massively. It has a massive market share in the server space. That’s why the Linux Foundation has a board of directors with many people from big tech corps.
Like other FOSS projects, Linux couldn’t really exist (the way it does) without sponsors and contributors. Companies sponsor its development to help improve and maintain the kernel, and in return they can use it for their needs. That’s how maintainers are compensated for their work. Many contributions are also from companies who add code to help adapt the kernel for their needs.
But because of the project’s licensing and leadership, privacy and transparency are protected. The kernel doesn’t collect or send user data to any company or server. If someone writes code that does that, leadership would probably not merge it anyways, but even if, any person could check (more frequently, “audit”) the code and remove it or choose to not use the project. Anyone is free to “fork” the kernel to make modifications or go a different direction, assuming they follow the license.
Now for desktop Linux distros, like Linux Mint, there’s other software that you should also think about its features and management, and whether that is “safe” too. Out of the box, things like Systemd (init process) or Cinnamon (desktop environment) are maintained and distributed by different groups. IMO for Linux Mint: it’s safe and protects privacy. But that’s subjective, it’s always good to be critical and to form your own opinions! :)