News Intel axes Clear Linux, the fastest distribution on the market — company ends support, effective immediately

[Admin]

Intel shuts down Clear Linux OS as part of the broad global cost-cutting strategy, but it is set to continue upstreaming Linux innovations to the Linux kernel.

Intel axes Clear Linux, the fastest distribution on the market — company ends support, effective immediately : Read more

 

[Firestone]

Wow this is sad. I never used it but I had several colleagues who were strong enthusiasts of Clear Linux for a long time and I always enjoyed hearing about the new innovations it was bringing.

 

[bit_user]

It consistently set the standard for other distros to measure up against. I'm sure it also fed a significant number of performance patches to upstream projects. Some of the other distros were probably copying bits and pieces of Clear Linux' config, to bolster their own performance.

In other words, the Linux ecosystem will be much poorer for its loss.

 

[das_stig]

Fork it and use it as the basic for the official Sovereign Linux if it's that good?

 

[bit_user]

Fork it and use it as the basic for the official Sovereign Linux if it's that good?

As I said, it's not like people haven't been borrowing and adapting changes from it, in other distros. So, the benefit of forking it seems somewhat limited to me.

The value of Clear Linux is really one of having a focused & skilled team that's dedicated to continual performance analysis and tuning, especially for the latest server processors that many Linux developers don't have ready access to. It's the resources and expertise behind the project that was its main value.

 

[thestryker]

It's pretty clear that Tan doesn't care what gets cut to meet his goals. The part I find very frustrating is that a lot of the seemingly extraneous parts of Intel are why they got where they are. Having a good product is only part of the equation being able to support said product better than the competition is how you keep/expand business.

While Clear itself isn't mandatory for Intel contributing to Linux it's been a good way to get optimizations into that entire ecosystem. It's also obvious that Intel is letting software engineers go which likely will have an effect on overall Linux support/contribution.

 

[jg.millirem]

It's pretty clear that Tan doesn't care what gets cut to meet his goals. The part I find very frustrating is that a lot of the seemingly extraneous parts of Intel are why they got where they are. Having a good product is only part of the equation being able to support said product better than the competition is how you keep/expand business.

While Clear itself isn't mandatory for Intel contributing to Linux it's been a good way to get optimizations into that entire ecosystem. It's also obvious that Intel is letting software engineers go which likely will have an effect on overall Linux support/contribution.

An important mark of a good organization is not just how well it achieves its own bottom line, but how it contributes to the ecosystem, to society. Intel is retreating into itself with recent leaders, because share price. What a broken system.

 

[bit_user]

Intel is retreating into itself with recent leaders, because share price. What a broken system.

"recent leaders"? When did Gelsinger do something like this?

 

[FoxtrotMichael-1]

I already knew Tan was going to be a disaster as soon as he was made CEO. I took it as a sell signal and sold my entire position in Intel at a small profit (and I do mean small). I entered the position with the full faith that Gelsinger could turn the ship around and make Intel competitive again while making them a second real option to TSMC in the foundry space. His exit showed that the board, and not any particular CEO, should really be blamed for the downfall of this historic US innovator.

 

[bit_user]

I already knew Tan was going to be a disaster as soon as he was made CEO. I took it as a sell signal and sold my entire position in Intel at a small profit (and I do mean small).

Since you were trying to make a long-term investment, it was probably wise to get out when you did.

However, I think there's a chance the savings from Tan's cuts will benefit the share price before the cumulative impacts really start to drag the company down. By then, we'll have to look at what the Taiwan situation looks like and whether it seems like Intel is on track to have any viable node beyond 14A. If not, that's when I'd get out. If so, it might still be worth the risk of staying in. But, not because he will Make Intel Great Again. Rather, a gamble that Intel is still best-positioned to capitalize on TSMC being cut off.

 

[waltc3]

Interesting. As this article mentions that the OS runs well on AMD CPUs, as well, I wonder if the OS isn't serving Intel CPU sales as originally envisioned. These days, Intel is just trying to keep its head above water. Hardware companies traditionally shy away from OS development, though, as the competition is fierce and the investment substantial. (Says the ex-owner/user of three separate versions of OS/2...😉)

 

[bit_user]

Interesting. As this article mentions that the OS runs well on AMD CPUs, as well, I wonder if the OS isn't serving Intel CPU sales as originally envisioned.

But they didn't say that it benefitted AMD CPUs as much or more than Xeons. Furthermore, it started out, back when Intel was the only one with AVX-512. Nowadays, the big Intel Xeons still have AVX-512, but they also have AMX and like half a dozen other hardware engines that need to be catered for with special drivers, library support, and OS tuning.

Another thing to consider is that Intel has non-x86 competition to worry about. So, if Clear Linux gave it an edge over ARM, then it's still valuable even if it's helping to lift AMD, as well.

Hardware companies traditionally shy away from OS development, though,

One of their former CEOs said, a couple decades ago, that Intel had made the mistake of depending on Microsoft to catering for its hardware. It was made in the context of Intel starting to invest in Linux support, where the idea might not have been so much to undermine Microsoft, but at least to show what's possible and perhaps to serve as a better test vehicle for its new chips.

In the end, such optimization efforts probably helped Linux dominate HPC and the cloud, both of which had subsequently been very profitable for Intel. Who knows: if Intel hadn't played an active role in optimizing Linux for x86, maybe POWER would've eaten more of their server marketshare.

(Says the ex-owner/user of three separate versions of OS/2...😉)

That's why Linux is brilliant. No one company is shouldering the whole burden of it.

Also, IBM sold OS/2 as a software product that you could install and run on any PC, regardless of whether it was IBM-branded. So, that was really a pure software play they tried to make. It almost worked, until they lost the court case over whether they could provide binary compatibility with Windows. IBM simply got out-lawyered - OS/2 didn't lose on merit.

 

[EduApps]

This is precisely why I chose Ubuntu over newer Linux distros. It is maintained for more than 20 years so I can safely say it will not end it's support by tomorrow.

 

[bit_user]

This is precisely why I chose Ubuntu over newer Linux distros. It is maintained for more than 20 years so I can safely say it will not end it's support by tomorrow.

Clear Linux is hardly a new kid on the block. It's been around for about a decade.

It's more telling to look at the financial health of the company behind it. As long as Canonical is doing well, Ubuntu should be alright. However, even then, you could end up in a CentOS type situation, which was an open source distro mirroring RHEL (Redhat Enterprise Linux) that Redhat killed off.

Something similar could happen with Ubuntu, if Canonical decided to make the LTS releases non-free, thereby forcing free users to upgrade every 6 months or so. Ubuntu already does a thing where they have a premium support tier. I installed it on a ARM board I have and discovered that they want you to buy a paid subscription to get critical security fixes.

https://ubuntu.com/pro/devices
​

 

[EduApps]

Clear Linux is hardly a new kid on the block. It's been around for about a decade.

Wow, I'm getting old. In my head it wasn't that long ago.

Ubuntu already does a thing where they have a premium support tier. I installed it on a ARM board I have and discovered that they want you to buy a paid subscription to get critical security fixes.

Ubuntu Pro isn't needed if you update it regularly. You can also subscribe for free up to 5 devices.

 

[skedge]

Tan works for the PRC. The future is closed source max deepseek censorship.

 

[bit_user]

Tan works for the PRC.

Just because we don't like what he's doing is no reason to demonize the man, himself. He will have presented his plans to Intel's board of directors and they must've signed off on this (or some broad outline, thereof), or else they'd just fire him and find someone else.

The future is closed source max deepseek censorship.

That's quite an overstatement. Through its architecture and its licensing, Linux is designed to strongly encourage open source drivers. Furthermore, big cloud operators don't want to deploy closed-source solutions. So, as long as Linux remains the OS of choice for many, we will continue to have a compelling open source option.

As for AI models and much of the rest of it, it's indeed proprietary and there's no reason to think that will change. The big AI companies and hyperscalers are at the top of the food chain. Whoever sits in that position of dominance tends to keep their own stuff closed, while pushing for their suppliers and vendors to embrace open source. It's hypocritical, but not illogical.

 

[abufrejoval]

I've never quite understood why Intel made this distro.

Or rather, I've always seen the existence of this distro as a sign that something wasn't working as it should in the Linux universe.

It harked of Intel pushing its own C-compiler, which for at least a time actively bypassed optimizations and ISA extensions also suitable or present in modern AMD CPUs in a move that seemed somewhat sinister.

But perhaps it also exposed that the other kernel maintainers simply weren't putting enough effort into exploiting what compilers could deliver.

In a better world, all of Intel's engineering efforts should have gone upstream and become an inherent part of all Linux x86 distros. Or at least into alternate kernels for the other distros, when backward compatibility with older hardware prevented mainlining everything.

And then there was the better and lighter CloudHypervisor, which also felt a bit odd as an Intel product, while it seemed very sensible as a general initiative. But I guess that one has gone and become independent of Intel?

In both projects I guess it's the hyperscaler who profited most from the work of these Intel engineers. And it's who should on one hand continue the work and on the other make sure it stayes open source and goes upstream.

Only... they tend to only go open source when it can be used as a torpedo against the competition e.g. Kubernetes vs. all other container variants.

Perhaps Valve could take over a significant portion of the work here: they seem very interested in pushing some parts of the performance envelope.

Are they working on a server equivalent, too, to host game cloud servers? It sound somewhat logical (and a bit frightening) to me..

 

[bit_user]

I've never quite understood why Intel made this distro.

It showcased what was possible, with a server-oriented distro tuned for maximum performance on their CPUs.

Or rather, I've always seen the existence of this distro as a sign that something wasn't working as it should in the Linux universe.

I'm sure they did, but you can only get so far by pushing. You also can't control what distros ultimately do, in the realm of how they build their kernel & packages and what tuning parameters they ship with. So, creating your own distro shows other maintainers and customers what's possible. This can create "pull" factors, motivating other distros to adopt some of your practices.

It also gives you an option that enables better service and support for key customers.

It harked of Intel pushing its own C-compiler, which for at least a time actively bypassed optimizations and ISA extensions also suitable or present in modern AMD CPUs in a move that seemed somewhat sinister.

Except they didn't, because Clear Linux also improved performance on AMD CPUs, compared to standard distros. So, if they tucked anything in there to undermine performance on AMD systems, it was neither very major nor very effective.

In a better world, all of Intel's engineering efforts should have gone upstream and become an inherent part of all Linux x86 distros.

They did upstream patches, but not all of their patches were suitable for distros or packages trying to support a wider range of CPUs or balance a wider variety of concerns.

 

[DS426]

I've never quite understood why Intel made this distro.

...

Marketing. It put Intel's name out there even more, did it not? It also lended them additionality credibility and possibly respect for contributing to Linux whereas Intel had been previously seen as exclusively in bed with Microsoft. Linux was gaining server market share back then, so it was the hot new thing to get into. Just look at how many distros there are today!

Intel probably wanted businesses to partner more directly with them rather than thru another vendor like Red Hat, SUSE, etc., even if it was unconventional for hardware manufacturers to directly develop an OS. They obviously did really well on partnerships and loyalty as even when AMD had better CPU's at times, they weren't able to keep those wins going long enough to take significant market share for long (Opteron peaked at about 26% for example, far less than EPYC today).

IBM is a notable exception, but they also ended up in a carve-out of the market where POWER and other platforms make the most sense given their strengths. My employer uses an ERP system based on POWER and IBM i that's shifted away from AS/400 emulation to using modern web apps. Looks like it won't be dead any time soon:

https://www.fortra.com/blog/as400-dead

 

[TerryLaze]

It's pretty clear that Tan doesn't care what gets cut to meet his goals. The part I find very frustrating is that a lot of the seemingly extraneous parts of Intel are why they got where they are. Having a good product is only part of the equation being able to support said product better than the competition is how you keep/expand business.

While Clear itself isn't mandatory for Intel contributing to Linux it's been a good way to get optimizations into that entire ecosystem. It's also obvious that Intel is letting software engineers go which likely will have an effect on overall Linux support/contribution.

It's such a shame that projects that get stopped can't be resumed when and if the company is at a better financial position again..................oh, wait!

 

[DSzymborski]

It's disappointing, but this is pretty typical behavior for a company in financial free fall. They're in triage mode; things have to basically part of their core competency *and* something that can conceivably be monetized effectively to survive at this point. They just cut Automotive a few weeks ago. I'm not sure there was ever going to be a real financial payoff, from their point of view, from efforts in Clear Linux. As marketing, it feels horribly inefficient.

 

[bit_user]

Marketing. It put Intel's name out there even more, did it not?

Not worth all the trouble, just for that. It should've boiled down to some combination of priming the open source ecosystem with optimizations for Intel products, maybe giving them an advantageous platform to compete for SPEC-marks, and trying to compete for big hardware purchases from cloudy customers.

I really wonder how many people used it as a deployment platform, and to what extent that was really ever even the goal.

even if it was unconventional for hardware manufacturers to directly develop an OS.

If you go back far enough, it's the way you got your OS. The idea of buying hardware from one company and an OS from another is largely a PC concept. Even today, Apple and IBM still do this, with IBM still having at least two of their own operating systems that have nothing to do with Linux (as you mentioned). As recently as the 1990's, Amiga and all of the non-x86 server/workstation vendors had their own operating systems (HP-UX, Solaris, IRIX, VMS, DEC UNIX, etc.).

even when AMD had better CPU's at times, they weren't able to keep those wins going long enough to take significant market share for long (Opteron peaked at about 26% for example, far less than EPYC today).

That's because Intel caught up to Opteron and passed it. They haven't managed to supersede EPYC. Opteron got a huge boost with 64-bit and its integrated memory controller, but also being more efficient.

 

[logainofhades]

It's such a shame that projects that get stopped can't be resumed when and if the company is at a better financial position again..................oh, wait!

You have a very optimistic attitude if you think that is going to ever happen without getting a GM style bailout. They should not have forced Pat out, and let him complete his vision. Instead, they could only think about short term $$$.

 

[bit_user]

It's such a shame that projects that get stopped can't be resumed when and if the company is at a better financial position again..................oh, wait!

If that were the plan, then they'd have just scaled it back but not completely ended it. Something like this is a major, long-term effort. You can't just start it up and shut it down as the business cycle ebbs and flows. Once it's gone, it's very unlikely ever to be coming back.