Revisit Debian? Stick with Tuxedo? Or look at Clear?
Hi all. As some might have read in a previous thread, my daily driver PC has gone flaky and it's in the shop. When it comes back I will need to reinstall an OS as I wiped the drive before sending it in (and I had decent backups). I may also find myself with a new system to be used for AI and video content development. Now that I'm reinstalling, I'm wondering if this is a time to re-evaluate what distro to use. I'm not a frequent distro-hopper but I'm not scared of it. In the last five years I've gone from Kubuntu to KDE Neon to TuxedoOS. So yes, the replacement needs to be KDE and I'm far more familiar with Debian/Ubuntu package management than Fedora/CentOS's. And I would prefer to have one that I can stick with for many years. I continue to hear people say great things about Debian+KDE, but my experiences with it were not good the last time I tried it a few years back. Too much hardware was unsupported, too many tweaks and workarounds just to get it to work. Is it worth a revisit? Which version? Should I stick with Ubuntu-based TuxedoOS? Or consider something wholly new like Clear Linix? The hardware to be used: - One system: High-end (strix halo) Ryzen CPU/NPU/GPU, Linux only - Another system: mid-range AMD CPU and RTX 3060 GPU, dual-boot - Realtek 8125B 2.5Gb Ethernet in both Notable software to be run: - OBS - Kdenlive - Handbrake - Firefox - Davinci Resolve (maybe) - Blender (maybe) - LMStudio - Ollama - CUDA on the Nvidia system, ROCm on the AMD one I don't want: - snapd (flatpacks are OK, I guess, for non-system stuff) - X11 (want Wayland as the default) - non-optimal drivers for any hardware (closed-source drivers are sad but OK) - old releases. I'm OK with a reasonable update frequency so long as it's not too bleeding edge. Advice appreciated. I'm especially curious to hear from anyone who has tried Clear Linux; it's supposed to be tuned for Intel, but both my CPUs are AMDs. Thanks! -- Evan Leibovitch, Toronto Canada @evanleibovitch / @el56
I was going to say Mint, but it doesn't do KDE. There isn't too many "KDE + Ubuntu/Debian - snap" distro. Since you already run Tuxedo, stick with it. On 2025-06-27 01:52, Evan Leibovitch via Talk wrote:
Hi all.
As some might have read in a previous thread, my daily driver PC has gone flaky and it's in the shop. When it comes back I will need to reinstall an OS as I wiped the drive before sending it in (and I had decent backups). I may also find myself with a new system to be used for AI and video content development.
Now that I'm reinstalling, I'm wondering if this is a time to re- evaluate what distro to use. I'm not a frequent distro-hopper but I'm not scared of it. In the last five years I've gone from Kubuntu to KDE Neon to TuxedoOS. So yes, the replacement needs to be KDE and I'm far more familiar with Debian/Ubuntu package management than Fedora/ CentOS's. And I would prefer to have one that I can stick with for many years.
I continue to hear people say great things about Debian+KDE, but my experiences with it were not good the last time I tried it a few years back. Too much hardware was unsupported, too many tweaks and workarounds just to get it to work.
Is it worth a revisit? Which version? Should I stick with Ubuntu-based TuxedoOS? Or consider something wholly new like Clear Linix?
The hardware to be used: - One system: High-end (strix halo) Ryzen CPU/NPU/GPU, Linux only - Another system: mid-range AMD CPU and RTX 3060 GPU, dual-boot - Realtek 8125B 2.5Gb Ethernet in both
Notable software to be run: - OBS - Kdenlive - Handbrake - Firefox - Davinci Resolve (maybe) - Blender (maybe) - LMStudio - Ollama - CUDA on the Nvidia system, ROCm on the AMD one
I don't want: - snapd (flatpacks are OK, I guess, for non-system stuff) - X11 (want Wayland as the default) - non-optimal drivers for any hardware (closed-source drivers are sad but OK) - old releases. I'm OK with a reasonable update frequency so long as it's not too bleeding edge.
Advice appreciated. I'm especially curious to hear from anyone who has tried Clear Linux; it's supposed to be tuned for Intel, but both my CPUs are AMDs.
Thanks!
-- Evan Leibovitch, Toronto Canada @evanleibovitch / @el56
------------------------------------ Description: GTALUG Talk Unsubscribe via Talk-unsubscribe@lists.gtalug.org Start a new thread: talk@lists.gtalug.org This message archived at https://lists.gtalug.org/archives/list/talk@lists.gtalug.org/message/LNN7FXR...
I just noticed CachyOS is #2 at distrowatch.com. Default desktop is KDE, and based on Arch which rubs me the wrong way. Maybe it goes the right way for you. On 2025-06-27 13:07, William Park via Talk wrote:
I was going to say Mint, but it doesn't do KDE. There isn't too many "KDE + Ubuntu/Debian - snap" distro. Since you already run Tuxedo, stick with it.
On 2025-06-27 01:52, Evan Leibovitch via Talk wrote:
Hi all.
As some might have read in a previous thread, my daily driver PC has gone flaky and it's in the shop. When it comes back I will need to reinstall an OS as I wiped the drive before sending it in (and I had decent backups). I may also find myself with a new system to be used for AI and video content development.
Now that I'm reinstalling, I'm wondering if this is a time to re- evaluate what distro to use. I'm not a frequent distro-hopper but I'm not scared of it. In the last five years I've gone from Kubuntu to KDE Neon to TuxedoOS. So yes, the replacement needs to be KDE and I'm far more familiar with Debian/Ubuntu package management than Fedora/ CentOS's. And I would prefer to have one that I can stick with for many years.
I continue to hear people say great things about Debian+KDE, but my experiences with it were not good the last time I tried it a few years back. Too much hardware was unsupported, too many tweaks and workarounds just to get it to work.
Is it worth a revisit? Which version? Should I stick with Ubuntu-based TuxedoOS? Or consider something wholly new like Clear Linix?
The hardware to be used: - One system: High-end (strix halo) Ryzen CPU/NPU/GPU, Linux only - Another system: mid-range AMD CPU and RTX 3060 GPU, dual-boot - Realtek 8125B 2.5Gb Ethernet in both
Notable software to be run: - OBS - Kdenlive - Handbrake - Firefox - Davinci Resolve (maybe) - Blender (maybe) - LMStudio - Ollama - CUDA on the Nvidia system, ROCm on the AMD one
I don't want: - snapd (flatpacks are OK, I guess, for non-system stuff) - X11 (want Wayland as the default) - non-optimal drivers for any hardware (closed-source drivers are sad but OK) - old releases. I'm OK with a reasonable update frequency so long as it's not too bleeding edge.
Advice appreciated. I'm especially curious to hear from anyone who has tried Clear Linux; it's supposed to be tuned for Intel, but both my CPUs are AMDs.
Thanks!
-- Evan Leibovitch, Toronto Canada @evanleibovitch / @el56
------------------------------------ Description: GTALUG Talk Unsubscribe via Talk-unsubscribe@lists.gtalug.org Start a new thread: talk@lists.gtalug.org This message archived at https://lists.gtalug.org/archives/list/ talk@lists.gtalug.org/message/LNN7FXRV6FQFX5WMBRTFJI4S32Q4XNSV/
------------------------------------ Description: GTALUG Talk Unsubscribe via Talk-unsubscribe@lists.gtalug.org Start a new thread: talk@lists.gtalug.org This message archived at https://lists.gtalug.org/archives/list/ talk@lists.gtalug.org/message/3XQH2VBGUESPO2HOAMNLXJBP672OAWFV/
William Park via Talk said on Mon, 30 Jun 2025 01:50:32 -0400
I just noticed CachyOS is #2 at distrowatch.com. Default desktop is KDE, and based on Arch which rubs me the wrong way. Maybe it goes the right way for you.
LOL, anything using systemd (Arch) rubs me the wrong way, with its unneeded overcomplexificationism. I go rid of all KDE for the same reason. Therefore, CachyOS isn't something I'd consider. SteveT Steve Litt http://444domains.com
Steve Litt via Talk wrote on 2025-06-30 17:02:
systemd (Arch) rubs me the wrong way, with its unneeded overcomplexificationism.
*tweeeeeet* Foul called! Opinion stated as fact, 10 yard penalty.
I go rid of all KDE for the same reason.
Again! You're about negative 20 miles by this point, Steve.
Ron wrote:
Opinion stated as fact, 10 yard penalty.
Penalty overturned by VAR. "Rubs me the wrong way" is as much opinion as can be. "Overcomplexificationism" seems enough of a coined term (that I like) to also qualify as subjective. And "uneeded" certainly sounds personal too. Independent of one's agreement or not, I see no claims of advice let alone authority.
I go rid of all KDE for the same reason. Again!
You're about negative 20 miles by this point, Steve.
Way too harsh. I like KDE and want it as my desktop, but "simplicity" is not in the top 20 of its feature list. At this level of refereeing you might consider applying to the WNBA. - Evan
Ron via Talk said on Mon, 30 Jun 2025 19:00:00 -0700
Steve Litt via Talk wrote on 2025-06-30 17:02:
systemd (Arch) rubs me the wrong way, with its unneeded overcomplexificationism.
*tweeeeeet*
Foul called!
Your opinion. https://troubleshooters.com/linux/systemd/lol_systemd.htm
Opinion stated as fact, 10 yard penalty.
Since when is "systemd rubs me the wrong way" an opinion? And if your penalty is for overcomplexificationism, compare systemd to runit. And if you say "but systemd does more", that's my point exactly: Elimination of interchangeable parts and test points. And if you say "but systemd gives its own diagnostics", that's like the fact that your car's ODB2 connection gives diagnostics, but in fact you need to get under the hood to get the true story.
I go rid of all KDE for the same reason.
Again!
You're about negative 20 miles by this point, Steve.
Not true, because long ago you got replaced by a referee not operating on the appeal to novelty fallacy. SteveT Steve Litt http://444domains.com
Should I stick with Ubuntu-based TuxedoOS?
What's the feature set of TuxedoOS, I'm not familiar with it?
Or consider something wholly new like Clear Linix?
Isn't that an Intel distro designed as a dev platform for building software for Intel chips? I might be way out in left field on both of those. I'm still using KDEneon and quite happy with it. I think I switched from snaps to deb packages for Thunderbird and Firefox because: 1) the profiles were copied into the snaps just fine, but I preferred them in ~/.{mozilla|thunderbird} 2) mailto: links weren't working within snap, possibly due to process isolation. Not commonly needed but when it didn't work that one time, I switched.
ron＠bclug.ca wrote:
Or consider something wholly new like Clear Linix? Isn't that an Intel distro designed as a dev platform for building software for Intel chips?
My main systems are AMD CPUs so Intel fine tuning won' t help me. I thought Clear was optimized for 64bit x86 systems but yeah it might just be something on which to run Intel benchmarks. My install has to be Ubuntu-derived because most of the Linux AI stuff seems to prefer it. I have no doubt that the Arch and Fedora communities have figured out how to do these things; but I'm interested into minimizing OS steps so I can spend more time playing with Ollama and friends,
I'm still using KDEneon and quite happy with it.
I moved off of Neon because IMO they were too frequent with the KDE updates and too infrequent with the Ubuntu ones. Tuxedo seems to do an extra bit of testing -- so that its KDE is just a step behind Neon's but seems to be a little more stable. I could come back to Neon but an happy with Tuxedo and will probably be sticking with it.
I think I switched from snaps to deb packages for Thunderbird and Firefox because:
I refuse to use snaps for anything. I wouldn't use flatpack for browsers but can be persuaded to use them for other things. - Evan
From: Evan Leibovitch via Talk <talk@lists.gtalug.org>
There are two kinds of distros: ones that start with another distro and mod it, and ones that start with the original open source code and package it. I tend to like the second kind because they are in control of their timetable and destiny and because the project has, by necessity, more expertise. Often they also have more users, which is a good thing (more testing, more contributors). Ubuntu is an exception. Although it is derived from debian, it has way more users, mind share, and perhaps engineers. I'm not sure why Ubuntu doesn't cut the tie. Its innovations have just as often been bad as good. I do like their two update and support cycles (every six months and every two years). It is the default distro for many software projects. I've not used debian much. I very much respect the project's structure and goals. I have been unhappy with its slow pace of package bug fixes and updates -- their freezes don't seem to match the nature of software change these days. Backporting fixes seems like the wrong approach once change gets high-volume. Lennart seems to thrive with debian, and his opinion carries a lot of weight with me. I like Fedora. Mostly. It has been very good for my uses. I wish that there were a more stable version for some of my systems. There's too much of a discontinuity between RHEL (and clones) and Fedora. It does adopt new versions of packages quite quickly. Fedora updates come fast and furious. I like the open source purity of Fedora but sometimes that is uncomfortable -- much less these days. I used to use CentOS for some systems but the version update process was horrible: the steps were a lot larger and I tended to put them off too long. The drama of Red Hat fighting cloners and damaging users was too much for me. I use Android or ChromeOS, mostly because I don't really have a choice on phones and tablets. They are Linux distros. You can even install Android on a PC. I don't like Snaps or Flatpaks. For two reasons: library bugs should be fixed in one package, not dozens; library resources should be expended once, not dozens of times. The time when Flatpak's are justified is for really big things shared by many distros. Think of Firefox. The individual distros probably add no value to it -- too complex. The Firefox developers probably want to handle bug reports without needing to have every distro at hand. The extra cost of duplicated libraries is probably a minor percentage of the resources used by FireFox. Bonus: perhaps whole-program optimization can make a difference (Link Time Optimization is a start). Clear Linux, as I understand it (not necessarily accurate), was a showcase for optimizations, both micro (using particular Intel instruction set additions, compiling with Intel-centric optimizations) and macro (getting rid of cruft that irritated the developers). The Wikipedia article is probably more useful than my opinion. <https://en.wikipedia.org/wiki/Clear_Linux_OS>. Statelessness is a nice goal. I don't know how packaging stacks would play out in my world. Personally, I don't find it rewarding to switch distros. I have enough other challenges. I will do it when my current distro become sufficiently uncomfortable. I admit that the last time I switched was when RHL split into RHEL and Fedora Core.
On Tue, 1 Jul 2025 at 10:24, D. Hugh Redelmeier via Talk <talk@lists.gtalug.org> wrote:
Clear Linux, as I understand it (not necessarily accurate), was a showcase for optimizations
I think Gentoo would also fall into that category. I used Gentoo on a laptop many years ago but found that I spent a fair amount of time deciding which "switches" to set for optimising the system and individual packages. (I suppose I could have just taken the defaults and left it at that.) Also, I think the CPU spent far more cycles "recompiling the world" for updates that affected large portions of the system than it did doing actual useful work :-) Though for the effort, you can end up with a very lean and efficient system, tailored to your liking. https://www.gentoo.org/get-started/about/ -- Scott
On Tue, Jul 01, 2025 at 10:24:25AM -0400, D. Hugh Redelmeier via Talk wrote:
Ubuntu is an exception. Although it is derived from debian, it has way more users, mind share, and perhaps engineers. I'm not sure why Ubuntu doesn't cut the tie. Its innovations have just as often been bad as good. I do like their two update and support cycles (every six months and every two years). It is the default distro for many software projects.
Ubuntu's fixed release cycle is one of the things I don't like. If things are not ready, don't release them. They have had some very broken releases over the years due to rushing to get things out for the release cycle. And I think being based on Debian has helped them a lot of times over the years. I think it was last year Ubuntu decided to try changing the dependency resolver in apt. It broke. Badly. It was the main reason the new LTS release was 3 or 4 months late. The fix was to revert to the Debian code that worked. Debian takes time to do things right with an interest in long term sustainability. Debian took a few years to decide on the right way to handle multiarch. Redhat barged ahead and did some quick way to handle it, which has a lot of problems and isn't flexible and extensible. Debian's solution works very well, and now allows an arbitrary number of architectures at once, including installing things that don't even natively work on the CPU which it can then handle using qemu and the binfmt handler. It took time but it really worked in the end. Redhat is stuck with their half baked solution now. Of course now that x86 has mostly switched to 64 bit they probably don't even care anymore.
I've not used debian much. I very much respect the project's structure and goals. I have been unhappy with its slow pace of package bug fixes and updates -- their freezes don't seem to match the nature of software change these days. Backporting fixes seems like the wrong approach once change gets high-volume. Lennart seems to thrive with debian, and his opinion carries a lot of weight with me.
In most cases bug fixes are quite fast, but of course Debian is entirely a volunteer project, unlike fedora and ubuntu and such that have corporatians behind them that can dictate what must be worked on. Backporting fixes is very much the right things to do. You fix the known bugs while usually avoiding introducing new bugs that are added as new features are being worked on.
I like Fedora. Mostly. It has been very good for my uses. I wish that there were a more stable version for some of my systems. There's too much of a discontinuity between RHEL (and clones) and Fedora. It does adopt new versions of packages quite quickly. Fedora updates come fast and furious. I like the open source purity of Fedora but sometimes that is uncomfortable -- much less these days.
I used to use redhat for a number of years, back when they had a desktop version (long before fedora was invented). I eventually gave up on it because the bugs at the time were simply too much, and they had apparently no interest in fixing them and not even any interest in bug reports from users if they weren't enterprise paying users. Buying their CD sets didn't count for anything. Of course once I started using Debian and learned how dependencies and package management could actually be done, and especially the difference in how the source packages were structured, I lost all interest in anything rpm based. The structire of an rpm is simplistic junk compared to how clean and organized the debian format is. There is no consistency in source rpms, they are largely just a pile of scripts that get run on top of some patches. The use of debhelper in debian makes for a much much more consistent package design, that is much easier to work with, and so much more is automated which results in a more consistent result (rpm's often have had missing dependencies because it doesn't do nearly as good a job figuring out the needed dependencies automatically).
I used to use CentOS for some systems but the version update process was horrible: the steps were a lot larger and I tended to put them off too long. The drama of Red Hat fighting cloners and damaging users was too much for me.
Yeah the way redhat has treated centos is really sad.
I use Android or ChromeOS, mostly because I don't really have a choice on phones and tablets. They are Linux distros. You can even install Android on a PC.
I don't like Snaps or Flatpaks. For two reasons: library bugs should be fixed in one package, not dozens; library resources should be expended once, not dozens of times.
Yeah I don't use them either.
The time when Flatpak's are justified is for really big things shared by many distros. Think of Firefox. The individual distros probably add no value to it -- too complex. The Firefox developers probably want to handle bug reports without needing to have every distro at hand. The extra cost of duplicated libraries is probably a minor percentage of the resources used by FireFox. Bonus: perhaps whole-program optimization can make a difference (Link Time Optimization is a start).
I will stick with proper packages.
Clear Linux, as I understand it (not necessarily accurate), was a showcase for optimizations, both micro (using particular Intel instruction set additions, compiling with Intel-centric optimizations) and macro (getting rid of cruft that irritated the developers). The Wikipedia article is probably more useful than my opinion. <https://en.wikipedia.org/wiki/Clear_Linux_OS>. Statelessness is a nice goal. I don't know how packaging stacks would play out in my world.
Personally, I don't find it rewarding to switch distros. I have enough other challenges. I will do it when my current distro become sufficiently uncomfortable. I admit that the last time I switched was when RHL split into RHEL and Fedora Core.
I switched a few times when I discovered something that was much better than what I was using. The last time I found that was a but over 25 years ago. So it was SLS, Slackware, Redhat, Debian. I have looked at mandrake, yggdrasil, suse (We use that at work, I don't know why anyone ever cared about it to be honest), ubuntu, mint and probably a few others. Mint makes a lot of sense if you want something easier to deal with than Debian. Some distributions I haven't bothered to even look at given their design goal was to be simply wrong (Ie gentoo, arch, and a number of others). -- Len Sorensen
Thanks for the discussion. Very helpful. On Wed, Jul 2, 2025 at 3:18 PM Lennart Sorensen via Talk < talk@lists.gtalug.org> wrote:
On Tue, Jul 01, 2025 at 10:24:25AM -0400, D. Hugh Redelmeier via Talk wrote:
Ubuntu is an exception. Although it is derived from debian, it has way more users, mind share, and perhaps engineers. I'm not sure why Ubuntu doesn't cut the tie. Its innovations have just as often been bad as good. I do like their two update and support cycles (every six months and every two years). It is the default distro for many software projects.
Unfortunately that last point is a compelling factor. As I look around the AI-on-Linux world, it seems that developers have concentrated on "official" support for Fedora and Ubuntu and little else. I note that we now have a third type of distro based on Hugh's categorization; second order derivatives, those based on Ubuntu which is itself based on Debian. Two of the distros on my short list (TuxedoOS and KDE Neon) fall in that category, replacing or repackaging specific components (and often replacing snaps with flatpaks) but otherwise leaving the heavy lifting to Canonical & friends for better or worse. Lazy or efficient? You decide. I have already encountered a situation (the AMD proprietary drivers) in which the package won't immediately install on Tuxedo but can be faked out by temporarily changing /etc/issue or the install script. Even then one must hope that any driver/kernel "enhancements" from the derivative distro do not impact installations that expect naked Ubuntu. I'm quite sure that SUSE, Arch and other distro communities have figured out workarounds to install many of these apps, but to me it's just one more PITA and excuse for things not working.
I like Fedora. Mostly. It has been very good for my uses. I wish that
there were a more stable version for some of my systems. There's too much of a discontinuity between RHEL (and clones) and Fedora.
If continuity with RHEL is important ... weren't Rocky and Alma developed for just that reason? Question: besides the package management and look-and-feel uses, are there any big technical or cultural issues moving from Kubuntu to Fedora/KDE?
I like the open source purity of Fedora but sometimes that is
uncomfortable -- much less these days.
That's exactly the same reaction I had to Debian when I tried it some years ago. I don't have an option; I have some bleeding-edge hardware that *needs* the proprietary drivers. And, much as I utterly despise them, I have external needs that demand I have proprietary chat apps. And there's also Steam, which brings a whole new world to my desktop. I applaud any project that really really really wants to keep their own stuff FOSS. But I draw the line between encouraging users to choose FOSS and being an obstacle to the proprietary stuff. By all means suggest IRC as an alternative to Whatsapp but don' t try to FORCE me down that path.
The time when Flatpak's are justified is for really big things shared by many distros. Think of Firefox. The individual distros probably add no value to it -- too complex. The Firefox developers probably want to handle bug reports without needing to have every distro at hand. The
extra cost of duplicated libraries is probably a minor percentage of the resources used by FireFox. Bonus: perhaps whole-program optimization can make a difference (Link Time Optimization is a start).
I will stick with proper packages.
I insist that my browser be a regular package ... it's such a vital component that I consider it a core system part. Plus, browsers may need to access or provide data elsewhere on the system where sandboxing may get in the way. For other apps I prefer regular packages but am not dogmatically against flatpaks. Snaps, OTOH, I refuse. -- Evan Leibovitch, Toronto Canada @evanleibovitch / @el56
From: Evan Leibovitch via Talk <talk@lists.gtalug.org>
On Wed, Jul 2, 2025 at 3:18 PM Lennart Sorensen via Talk < talk@lists.gtalug.org> wrote:
On Tue, Jul 01, 2025 at 10:24:25AM -0400, D. Hugh Redelmeier via Talk wrote:
Ubuntu is an exception. Although it is derived from debian, it has way more users, mind share, and perhaps engineers. I'm not sure why Ubuntu doesn't cut the tie. Its innovations have just as often been bad as good. I do like their two update and support cycles (every six months and every two years). It is the default distro for many software projects.
Unfortunately that last point is a compelling factor. As I look around the AI-on-Linux world, it seems that developers have concentrated on "official" support for Fedora and Ubuntu and little else.
In the past, "official" support has been for RHEL and Ubuntu LTS. Vendors of kludgy stuff (not open source) wanted platform stability. That's not a phrase that I'd use to describe Fedora.
I have already encountered a situation (the AMD proprietary drivers) in which the package won't immediately install on Tuxedo but can be faked out by temporarily changing /etc/issue or the install script. Even then one must hope that any driver/kernel "enhancements" from the derivative distro do not impact installations that expect naked Ubuntu.
Those proprietary things are a kludge on the Linux distro model. No wonder there are problems. How can you (or the vendor) be sure you added kludges are reliable? They really want to run on Windows. Windows has ABIs that are stable until they are not.
I'm quite sure that SUSE, Arch and other distro communities have figured out workarounds to install many of these apps, but to me it's just one more PITA and excuse for things not working.
Another sad case of workarounds: *BSD variants. They resort to emulating Linux to get certain desktop capabilities. Note: *BSD systems are respectable but they don't have enough resources to make comfortable desktops.
I like Fedora. Mostly. It has been very good for my uses. I wish that
there were a more stable version for some of my systems. There's too much of a discontinuity between RHEL (and clones) and Fedora.
If continuity with RHEL is important ... weren't Rocky and Alma developed for just that reason?
Sorry: I meant continuity between versions of RHEL. I didn't say it, but another problem with RHEL and its clones is the lack of even a pretense of contributing code to them. I did make a few contributions through the Libreswan project. In fact, I even helped RH with customer problems (not often). I do make bug reports. Doing that right takes effort and I hope it improves the quality of distros. In the original CentOS days (before RH took them over) but reports to CentOS were pointless: all change had to be upstream. After RH took over CentOS, you could report bugs to RH, but they were really about RHEL. I have no idea about bug reports against Rocky or Alma.
I like the open source purity of Fedora but sometimes that is
uncomfortable -- much less these days.
That's exactly the same reaction I had to Debian when I tried it some years ago. I don't have an option; I have some bleeding-edge hardware that *needs* the proprietary drivers.
You had a choice: you could have not used that hardware (sending a message through the marketplace), or use that hardware on a dedicated machine. Do you think big compute farms running piles GPUs use Windows? I'm pretty sure they run Linux. And the GPU makers cater to that.
And, much as I utterly despise them, I have external needs that demand I have proprietary chat apps.
You could use that chat app in a virtual machine (Android or Windows, I guess). Bonus: that would sandbox the app.
And there's also Steam, which brings a whole new world to my desktop.
It comes with its own Linux Distro. I understand that they are moving to a more distro-agnostic approach.
I applaud any project that really really really wants to keep their own stuff FOSS. But I draw the line between encouraging users to choose FOSS and being an obstacle to the proprietary stuff. By all means suggest IRC as an alternative to Whatsapp but don' t try to FORCE me down that path.
You are reading that the wrong way. It's Whatsapp that is doing the forcing, not the distro. Or, as you did for a while, run Windows on your desktop with an enclave for Linux when useful.
On Fri, Jul 4, 2025 at 10:32 AM D. Hugh Redelmeier via Talk < talk@lists.gtalug.org> wrote:
In the past, "official" support has been for RHEL and Ubuntu LTS. Vendors of kludgy stuff (not open source) wanted platform stability. That's not a phrase that I'd use to describe Fedora.
The AI world is moving faster than even Fedora, so that's less of a factor here.
I like the open source purity of Fedora but sometimes that is
uncomfortable -- much less these days.
That's exactly the same reaction I had to Debian when I tried it some years ago. I don't have an option; I have some bleeding-edge hardware that *needs* the proprietary drivers.
You had a choice: you could have not used that hardware (sending a message through the marketplace), or use that hardware on a dedicated machine.
Well, it's a choice, but I believe a safe one for the longer term though a pain in the short term. I intend to keep this box for a while, and what's bleeding edge today will be legacy before I can blink. FWIW, it's based on the AMD Ryzen AI MAX+ 395. Dumb name but quite the capability. I can assign 96GB to GPU memory which should let me run some interesting LLMs locally. It's probably a little slower than an RTX 5090, but that maxes out at 32GB and the GPU card costs more than my entire system. It's new enough that AMD is about to roll out ROCm for this architecture Real Soon Now. I'm confident they will since obviously they have a big stake in maximizing this chip's performance. And, yes, in a sense, I *did* want to send a message through the marketplace. AMD (for which I still have some sentiment because of ATI) should be encouraged to produce systems like this, which come close to the specs of the Nvidia DIGITS box (now called the DGX Spark) at substantially less cost. IMO Nvidia is treating its non-industrial customers like shit these days, and this is my way of saying with my choice that such arrogance has a cost. And if it means that I need to use the proprietary-driver kludge until the open source driver catches up, I can live with that. The relevance here is that this choice at least partially determines my choice of distro.
And, much as I utterly despise them, I have external needs that demand I
have proprietary chat apps.
You could use that chat app in a virtual machine (Android or Windows, I guess). Bonus: that would sandbox the app.
So do flatpaks, which is how most of them ship these days.
I applaud any project that really really really wants to keep their own
stuff FOSS. But I draw the line between encouraging users to choose FOSS and being an obstacle to the proprietary stuff. By all means suggest IRC as an alternative to Whatsapp but don' t try to FORCE me down that path.
You are reading that the wrong way. It's Whatsapp that is doing the forcing, not the distro.
They choose not to open source, I have no say in that decision. I have a choice not to use it but I am unwilling to forego the global connectivity that such apps offer. The people with whom I can communicate are forcing my hand and that's not going to change anytime soon. Or, as you did for a while, run Windows on your desktop with an enclave for
Linux when useful.
Indeed I did. But I've been researching and found that Windows Subsystem for Linux is sub-optimal for AI tasks -- especially on AMD. - Evan
From: Evan Leibovitch via Talk <talk@lists.gtalug.org>
Thanks for sharing your experience and insights. We really do learn from each others different experiences.
The AI world is moving faster than even Fedora, so that's less of a factor here.
I guess it depends on who is supplying what parts of the stack. Every supplier wants the stuff below them to stand still so they don't need to support many variants.
Well, it's a choice, but I believe a safe one for the longer term though a pain in the short term. I intend to keep this box for a while, and what's bleeding edge today will be legacy before I can blink.
Yeah. Ordinary computer hardware has been slow to change for the last decade or so. That reduced stresses on us. Hardware for AI is innovating like crazy. We're not used to that. It stresses all parts of the stack, including the OS.
FWIW, it's based on the AMD Ryzen AI MAX+ 395. Dumb name but quite the capability. I can assign 96GB to GPU memory which should let me run some interesting LLMs locally. It's probably a little slower than an RTX 5090, but that maxes out at 32GB and the GPU card costs more than my entire system.
Right. My (unearned) intuition that what matters is memory bandwidth and GPU's get amazing bandwidth via very-wide memory paths. You cannot do that with ordinary RAM on a PC. (The Mac gets partway there.) You are not trying to train models (the really compute-intensive stuff) so maybe the demands are not beyond what PC memory can handle.
It's new enough that AMD is about to roll out ROCm for this architecture Real Soon Now. I'm confident they will since obviously they have a big stake in maximizing this chip's performance.
I don't understand AMD's behaviour. They certainly have been slow and sporadic in their ROCm support for various cards. There were indications at some points that they only cared about the industrial AI cards.
And, yes, in a sense, I *did* want to send a message through the marketplace. AMD (for which I still have some sentiment because of ATI) should be encouraged to produce systems like this,
I've been encouraging AMD for decades. The results have been frustratingly mediocre.
Or, as you did for a while, run Windows on your desktop with an enclave for Linux when useful.
Indeed I did. But I've been researching and found that Windows Subsystem for Linux is sub-optimal for AI tasks -- especially on AMD.
Sometimes it makes sense to have different boxes for different universes. Mixing your everyday desktop and your compute monster might not be the right choice. Or it might be. - I want my desktop on all the time so I'd like its power requirements and noise to be minimal. Those might be limiting for AI (or might not). - I want a certain kind of stability on my desktop. Probably different from an AI box.
On Fri, Jul 4, 2025 at 11:51 AM D. Hugh Redelmeier via Talk < talk@lists.gtalug.org> wrote:
From: Evan Leibovitch via Talk <talk@lists.gtalug.org>
Thanks for sharing your experience and insights. We really do learn from each others different experiences.
Ditto. Every supplier wants the stuff below them to stand still so they don't need
to support many variants.
And in Linux AI it seems that this is why they've settled on one deb-based distro (Ubuntu) and one rpm-based one (Fedora). Everyone on something else can make their own kludges. Arch users, you're on your own. The choice of Fedora *appears* from my limited perspective to be based on strength, size and responsiveness of the community as being more important than stability. My guess of why Ubuntu was picked over Debian as a base is -- like Fedora -- a combination of popularity, community, and stable corporate sponsorship. (This might be why so many other popular distros such as Mint, Zorin and Pop_OS have gone that route too.) Hardware for AI is innovating like crazy. We're not used to that. It stresses
all parts of the stack, including the OS.
It's even created totally new components, such as an NPU to coexist with the CPU and GPU. My (unearned) intuition that what matters is memory bandwidth and GPU's
get amazing bandwidth via very-wide memory paths. You cannot do that with ordinary RAM on a PC. (The Mac gets partway there.)
Bandwidth gets you speed, but GPU memory constrains the size (and thus the accuracy and utility) of the LLM you can easily work with. Both are relevant and these days we have tradeoffs. Expensive GPU boards have fast bandwidth but low amounts of memory. The new integrated CPU/NPU/GPU systems have less bandwidth but access to far more GPU memory. The Strix Halo (the name of the architecture, also not great but better than the chip's name) approach is interesting. Like in recent M4-based Macs and the announced-but-not-released Nvidia DGX Spark (formerly Project DIGITS), it has memory that can be used by either the CPU or GPU. In the case of my system the memory is soldered-in and can' t be upgraded. Prices are in USD and approximate: GMKtec EVO-X2 w/AMD Ryzen AI MAX+ 395 ($2000) 128GB shared memory, 256GB/sec bandwidth Nvidia DGX Spark ($3500 est) 128GB shared, 250GB/sec Nvidia RTX A6000 ($5300, card only) 48GB, 768GB/sec Nvidia RTX 5090 ($3000, card only) 32GB, 1792GB/sec Mac mini M4 Pro ($2000) 64GB shared, 273GB/sec Macbook M4 Max ($4700) 128GB shared, 546GB/sec
You are not trying to train models (the really compute-intensive stuff) so maybe the demands are not beyond what PC memory can handle.
This is why, to me, the tradeoff of slower speed but more GPU memory makes sense. I don't understand AMD's behaviour. They certainly have been slow and sporadic
in their ROCm support for various cards. There were indications at some points that they only cared about the industrial AI cards.
That is certainly Nvidia's behaviour. AMD is too new to the "consumer" GPU/NPU game to tell yet if they're following the same bad path. Sometimes it makes sense to have different boxes for different universes.
That may indeed be the solution. I have a suitable second system, which is currently down as its CPU is awaiting a warranty replacement. I've already started charting the various apps I want to use and what would run better (or only) on one system or the other.
Mixing your everyday desktop and your compute monster might not be the right choice. Or it might be.
There are clear benefits to doing desktop productivity and communications on a Windows box and leave the AI stuff to a native Linux system. The third category of work I want to do ... video production and content creation ... runs better (or only) on Windows but could benefit from the power of the AI system. So that might need to be divvied up. And this brings me back to the OP -- which distro? Thanks to this thread so far I've narrowed it to Fedora/KDE versus TuxedoOS (Ubuntu based, KDE, snap-free), but I haven't gone near RPM-world since my Mandrake days -- would there be much pain in a switch? - Evan
Or consider something wholly new like Clear Linix [sic]?
Interesting timing:
It's not a good time to be at Intel. They just shut down Clear Linux OS and have announced more layoffs with Linux Maintainers having to leave their projects. This is having a big impact on Linux and Intel and users of this operating system.
Another source: https://www.techpowerup.com/339074/intel-shuts-down-clear-linux-distribution...
Linux enthusiasts gather to hear the latest news: Intel is officially shutting down its Clear Linux distribution after ten years of development and optimizations. As many recall, Clear Linux is an Intel-optimized Linux distribution that serves as a high-performance, optimized OS designed to extract every last ounce of performance from Intel hardware, especially Intel Xeons.
arjan (Clear Linux Developer)
After years of innovation and community collaboration, we're ending support for Clear Linux OS. Effective immediately, Intel will no longer provide security patches, updates, or maintenance for Clear Linux OS, and the Clear Linux OS GitHub repository will be archived in read-only mode. So, if you're currently using Clear Linux OS, we strongly recommend planning your migration to another actively maintained Linux distribution as soon as possible to ensure ongoing security and stability.
Apologies for any formatting issues, had to reply in-browser as I couldn't find the thread in Thunderbird.
On Sun, Jul 20, 2025 at 1:02 AM Ron BC (me, superuser) via Talk < talk@lists.gtalug.org> wrote:
Or consider something wholly new like Clear Linix [sic]?
Interesting timing:
It's not a good time to be at Intel. They just shut down Clear Linux OS and have announced more layoffs with Linux Maintainers having to leave their projects. This is having a big impact on Linux and Intel and users of this operating system.
Another source:
https://www.techpowerup.com/339074/intel-shuts-down-clear-linux-distribution...
Linux enthusiasts gather to hear the latest news: Intel is officially shutting down its Clear Linux distribution after ten years of development and optimizations. As many recall, Clear Linux is an Intel-optimized Linux distribution that serves as a high-performance, optimized OS designed to extract every last ounce of performance from Intel hardware, especially Intel Xeons.
arjan (Clear Linux Developer)
After years of innovation and community collaboration, we're ending support for Clear Linux OS. Effective immediately, Intel will no longer provide security patches, updates, or maintenance for Clear Linux OS, and the Clear Linux OS GitHub repository will be archived in read-only mode. So, if you're currently using Clear Linux OS, we strongly recommend planning your migration to another actively maintained Linux distribution as soon as possible to ensure ongoing security and stability.
Apologies for any formatting issues, had to reply in-browser as I couldn't find the thread in Thunderbird.
The only source that should matter :-) https://community.clearlinux.org/t/all-good-things-come-to-an-end-shutting-d... Dhaval
Dhaval Giani via Talk said on Sun, 20 Jul 2025 08:23:23 -0700
The only source that should matter :-) https://community.clearlinux.org/t/all-good-things-come-to-an-end-shutting-d...
When considering an escape route from your current distro, I suggest you consider and evaluate Void Linux, which I've used for 10 years now. Void is deVOID of the training wheels that most other distros either [help you with | saddle you with]. With Void, you are more in charge than with most other distros. Void Linux is a rolling release, meaning that your software is much newer than most, especially Devuan/Debian stable, or even unstable. If you've had bad experiences with rolling releases (Gentoo, Arch, etc), Void is *much* more reliable while updating, and if you get stuck, their #voidlinux IRC channel at Libera chat. Void Linux inits with the very simple and reliable parallel startup Runit init system, rather than the systemd atrocity or the sysvinit 30 year hoarder house full of junk. Although arguable, I think Runit is better than OpenRC and Busybox init. With Runit, your computer starts up clean as a spring breeze every time, in a timeframe comparable to systemd, and when it shuts down, it's much, much, MUCH faster than systemd. Naturally Void Linux doesn't package as much software as the big three (Redhat, Debian and Ubuntu), but it's very amenable to ./configure;make;make install , especially given that it always has very modern libraries. Even though Gnome is tightly coupled to systemd and there's no systemd package for systemd on Void Linux, the Void developers have somehow gotten Gnome to work on Void, both on X and Wayland. Getting back to the subject of training wheels. I define training wheels as GUI stuff that substitutes for editing a config file. I also define it as using a distro-proprietary declarative syntax config file to provide a "user friendly" layer above the shellscript that actually does the work. Here's a shortened version of my networking shellscript on Void Linux, where this shellscript is run from /etc/rc.local : ========================================= hostname -F /etc/hostname ip link set dev lo up ip link set dev enp3s6 down ip addr add 192.168.100.87/24 dev enp3s6 ip link set dev enp3s6 up ip route add default via 192.168.100.96 ========================================= The preceding is, to my knowledge, distro independent, as long as you shut off the distro's training wheel style networking. Training wheels they're good to have for the occasional user or the user using Linux as an appliance for one or a few programs. Training wheels are good for the user not yet confident of his shell scripting skills. They're very necessary for the new Linux user who is used to Windows or Mac. But eventually, they get in the way. Void Linux certainly isn't for everyone, but it's worth a look. SteveT Steve Litt http://444domains.com
As you said. Ron, good timing. Now, while Clear Linux as an ongoing sustained project is over, I wonder whether its existing optimizations may find their way into the mainstream kernel, libraries, compiler processes etc. so that all distros may benefit from what Intel has contributed. Even if nothing more has been added I wonder if what has been developed so far would be useful in other distros. - Evan On Sun, Jul 20, 2025 at 11:24 AM Dhaval Giani via Talk < talk@lists.gtalug.org> wrote:
On Sun, Jul 20, 2025 at 1:02 AM Ron BC (me, superuser) via Talk < talk@lists.gtalug.org> wrote:
Or consider something wholly new like Clear Linix [sic]?
Interesting timing:
It's not a good time to be at Intel. They just shut down Clear Linux OS and have announced more layoffs with Linux Maintainers having to leave their projects. This is having a big impact on Linux and Intel and users of this operating system.
Another source:
https://www.techpowerup.com/339074/intel-shuts-down-clear-linux-distribution...
Linux enthusiasts gather to hear the latest news: Intel is officially shutting down its Clear Linux distribution after ten years of development and optimizations. As many recall, Clear Linux is an Intel-optimized Linux distribution that serves as a high-performance, optimized OS designed to extract every last ounce of performance from Intel hardware, especially Intel Xeons.
arjan (Clear Linux Developer)
After years of innovation and community collaboration, we're ending support for Clear Linux OS. Effective immediately, Intel will no longer provide security patches, updates, or maintenance for Clear Linux OS, and the Clear Linux OS GitHub repository will be archived in read-only mode. So, if you're currently using Clear Linux OS, we strongly recommend planning your migration to another actively maintained Linux distribution as soon as possible to ensure ongoing security and stability.
Apologies for any formatting issues, had to reply in-browser as I couldn't find the thread in Thunderbird.
The only source that should matter :-)
https://community.clearlinux.org/t/all-good-things-come-to-an-end-shutting-d...
Dhaval ------------------------------------ Description: GTALUG Talk Unsubscribe via Talk-unsubscribe@lists.gtalug.org Start a new thread: talk@lists.gtalug.org This message archived at https://lists.gtalug.org/archives/list/talk@lists.gtalug.org/message/6WO2L4T...
-- Evan Leibovitch, Toronto Canada @evanleibovitch / @el56
On Sun, Jul 20, 2025 at 10:02 PM Karen Lewellen <klewellen@shellworld.net> wrote:
Interestingly enough on the Debian list there is a post who says he has run a specific Debian stable..for 20 years. They posted asking about Debian additions, but reading your post about clear Linux is why they decided to stop? Especially with a supportive community?
If I had to speculate, it is because Intel is going through a rough patch and that they are not in a position to support paying someone to work on this project. It is an open source project, so anyone who does care about it is free to fork it and continue maintaining it. Dhaval
participants (10)
-
D. Hugh Redelmeier -
Dhaval Giani -
Evan Leibovitch -
Evan Leibovitch -
Lennart Sorensen -
Ron -
ron＠bclug.ca -
Scott Allen -
Steve Litt -
William Park