Intel: No Linux for You, Clover Trail

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

glubox

Distinguished
Apr 15, 2009
15
0
18,510
Hell no, it's wrong intel, very wrong. Intel makes the hardware why do they need to involve software vendors. Make the chip give the specs to MS and public and who makes the best OS let win.
 


Pretty much this.

Just because Intel says, "No!" does not mean it won't work.

It may work well, or may work with reduced functionality. But don't complain to Intel when you don't get *support* Same may well be the case with AMD, though they seem to be a bit better on the graphics side of things lately.

Just a guess, but I suspect the initial Win8 exclusivity with the new Tabs is just to give the platform a jump start over any competing ARM/*Nix product. It may come as a shock to some of the Win8'ators (Haters) out there, but they may have something special here in terms of connectivity and peripheral support.





 

Feldii

Distinguished
Nov 23, 2011
36
0
18,540
I think just about everyone is missing the point here. It's not that Intel is purposely building a system that does not support Linux, it's just hard right now to build an SoC that supports multiple operating systems. Can you think of any phone chip, for example, that supports multiple operating systems?

Intel is not designing every piece of CloverTrail itself, and some of the subparts do not have drivers for Linux. Due to IP issues, they probably are not going to have Linux drivers any time soon, unless someone wants to reverse engineer them. Intel is making SoCs that support Android though, and will eventually make SoCs that support both Windows and Linux I bet.

Disclaimer: I work for Intel, but I do not speak for Intel. I also do not work on CloverTrail. This is all just speculation on my part, based on public information.
 

jerm1027

Distinguished
Apr 20, 2011
404
0
18,810

I'm not sure if you've noticed, but Android runs a Linux kernel. And there quite a bit of SoC's that run multiple OSes, say the Rasberry Pi, which is capable of running Android, as well as Debian, and probably Lubuntu.
 
[citation][nom]Feldii[/nom]I think just about everyone is missing the point here. It's not that Intel is purposely building a system that does not support Linux, it's just hard right now to build an SoC that supports multiple operating systems. Can you think of any phone chip,[/citation](cough) Tegra 3 (cough)
[citation][nom]Feldii[/nom]for example, that supports multiple operating systems?[/citation]
You're kidding, right? That part where I "interrupted" you is literally when the words "Tegra 3" popped into my head. Unbidden. And I'm a pretty fast reader. Tegra 3 powers the HTC One X, as well as any number of Android tablets, including those made by Asus, who also chose the Tegra 3 to power their Windows 8 RT tablet.
[citation][nom]Feldii[/nom]Intel is not designing every piece of CloverTrail itself, and some of the subparts do not have drivers for Linux. Due to IP issues, they probably are not going to have Linux drivers any time soon, unless someone wants to reverse engineer them. Intel is making SoCs that support Android though, and will eventually make SoCs that support both Windows and Linux I bet.[/citation]
Yeah, and if that's really what's going on here, the bottom line is, if they had said, "Sorry, open source dudes, but there's this thing on our chip, we didn't design it, and the people we licensed it from won't release the driver source code..." there'd be a lot less outlash--granted we're talking a very small number of angry fan-people here, compared to, say, the population of the entire planet, but still...--against Intel and a lot more outlash against whomever designed "this thing".

Of course, that's the difference between Intel saying "Sorry, right now these guys do this better than we do, so we gotta go with them; you understand, right?" and Intel saying "Our $#!+ is just so awesome that we just can't be bothered to support Linux with it." Pride goeth before &c. &c...
[citation][nom]Feldii[/nom]Disclaimer: I work for Intel, but I do not speak for Intel. I also do not work on CloverTrail. This is all just speculation on my part, based on public information.[/citation]
Props to you for owning up to your affiliations, if you really do work for Intel...
 
[citation][nom]Feldii[/nom]I think just about everyone is missing the point here. It's not that Intel is purposely building a system that does not support Linux, it's just hard right now to build an SoC that supports multiple operating systems. Can you think of any phone chip, for example, that supports multiple operating systems? Intel is not designing every piece of CloverTrail itself, and some of the subparts do not have drivers for Linux. Due to IP issues, they probably are not going to have Linux drivers any time soon, unless someone wants to reverse engineer them. Intel is making SoCs that support Android though, and will eventually make SoCs that support both Windows and Linux I bet.Disclaimer: I work for Intel, but I do not speak for Intel. I also do not work on CloverTrail. This is all just speculation on my part, based on public information.[/citation]

There are several, if not several dozen, SoCs that support multiple platforms.
 

the_brute

Distinguished
Feb 2, 2009
131
0
18,680
darn ~2% of computer users will not use the new atom.

Intel is looking to keep their LARGEST market up. wintel. (this is the ~95% market)

This said yes it is a HUGE loss for Linux, Intel was a large supporter of them back in '06. But when you use a light linux or DSL you dont need anything faster than a 1st gen Atom and Intel cant sell you a new processor. This is a money thing not love/hate for any os.

I say this on a tri-booted Fedora(running)/Win7/kubuntu. and run through a linux firewall (1st gen atom I got for free).
 
[citation][nom]the_brute[/nom]darn ~2% of computer users will not use the new atom. Intel is looking to keep their LARGEST market up. wintel. (this is the ~95% market) This said yes it is a HUGE loss for Linux, Intel was a large supporter of them back in '06. But when you use a light linux or DSL you dont need anything faster than a 1st gen Atom and Intel cant sell you a new processor. This is a money thing not love/hate for any os. I say this on a tri-booted Fedora(running)/Win7/kubuntu. and run through a linux firewall (1st gen atom I got for free).[/citation]

It's technically not a normal Linux distro, but Android would have probably seen huge usage on devices using these CPUs.
 

Feldii

Distinguished
Nov 23, 2011
36
0
18,540

That is a good example. And it's the only one mentioned that I'd give full credit for. Yes, the Raspberri Pi and others SoCs support multiple OSes, but they all look like variants of Linux to me. And, I'm still not 100% sure its the exact same Tegra 3 SoC in a Windows RT tablet and an Android tablet. Note that the wikipedia page for Tegra 3 actually lists three variants of Tegra 3, and I bet more are coming. And I hope we can agree that Nvidia is #1 when it comes to confusing naming.

Yeah, and if that's really what's going on here, the bottom line is, if they had said, "Sorry, open source dudes, but there's this thing on our chip, we didn't design it, and the people we licensed it from won't release the driver source code..." there'd be a lot less outlash--granted we're talking a very small number of angry fan-people here, compared to, say, the population of the entire planet, but still...--against Intel and a lot more outlash against whomever designed "this thing".

Yes, I think that this is either a case of an Intel person talking out of turn, or the media only giving part of his answer. I certainly do not think Intel wanted to upset the Linux community.
 
[citation][nom]Feldii[/nom]That is a good example. And it's the only one mentioned that I'd give full credit for. Yes, the Raspberri Pi and others SoCs support multiple OSes, but they all look like variants of Linux to me. And, I'm still not 100% sure its the exact same Tegra 3 SoC in a Windows RT tablet and an Android tablet. Note that the wikipedia page for Tegra 3 actually lists three variants of Tegra 3, and I bet more are coming. And I hope we can agree that Nvidia is #1 when it comes to confusing naming. Yes, I think that this is either a case of an Intel person talking out of turn, or the media only giving part of his answer. I certainly do not think Intel wanted to upset the Linux community.[/citation]

Many SoCs support Android, several other mobile OSes (not all of which are Linux-based or are at least compatible with Android and/or contemporary Linux), and several versions of contemporary Linux distros.

Beyond SoCs, we have many other CPUs that also support multiple platforms. Technically, it's usually just the software that has to support an SoC, so if someone really wanted to, they could port many operating systems to any SoC that they wanted too. From what I understand of this article, Linux simply won't support these Clover Trail Atoms for a while because of hardware changes with the CPU and/or GPU and/or pretty much anything else on these SoCs are not (yet) supported by Linux. Intel and/or one or more other companies that are related to these SoCs probably don't want to create Linux support nor give info for others to make Linux support for them at this time.

However, given the necessary information, pretty much any OS should be able to be ported to pretty much any hardware platform.
 

Feldii

Distinguished
Nov 23, 2011
36
0
18,540
However, given the necessary information, pretty much any OS should be able to be ported to pretty much any hardware platform.

Of course, but most drivers are written by the same company that builds the hardware, and it is rather difficult to write them on your own. I thought that was the core issue behind Linus flipping Nvidia the bird, they were refusing to help with writing various graphics drivers.

In any case, Intel has now come out and said that Clover Trail will support Linux after all. This either shows that Intel did not expect the reaction it got or did not mean to make the statement the way it did---probably both. It looks like it may be a bit of a name game though, just as I suspected Nvidia might be doing with the Tegra 3. It sounds to me like the SoC called Clover Trail in a Windows Tablet will not be the exact same SoC that is called Clover Trail in an Android Tablet. I don't know if it will be possible to buy a Windows Clover Trail tablet or a Windows Tegra 3 tablet and later install Linux on it. I guess time will tell.
 
[citation][nom]Feldii[/nom]Of course, but most drivers are written by the same company that builds the hardware, and it is rather difficult to write them on your own. I thought that was the core issue behind Linus flipping Nvidia the bird, they were refusing to help with writing various graphics drivers.In any case, Intel has now come out and said that Clover Trail will support Linux after all. This either shows that Intel did not expect the reaction it got or did not mean to make the statement the way it did---probably both. It looks like it may be a bit of a name game though, just as I suspected Nvidia might be doing with the Tegra 3. It sounds to me like the SoC called Clover Trail in a Windows Tablet will not be the exact same SoC that is called Clover Trail in an Android Tablet. I don't know if it will be possible to buy a Windows Clover Trail tablet or a Windows Tegra 3 tablet and later install Linux on it. I guess time will tell.[/citation]

Linus's anger at Nvidia is because Nvidia doesn't let Linux devs do the driver creation/maintenance work, not because Nvidia isn't helping to write graphics drivers (Nvidia writes the drivers themselves, but doesn't fix issues quickly enough when they fix them at all). If Intel gave the info of the chip to the open source community (copyrighted/patented to avoid stealing, of course), then they'd write the necessary drivers themselves.

I don't that Clover Trail will never have Linux support unless as you said, there will be multiple versions of it with differing platform optimization. If the latter is the case, then there's nothing wrong with having multiple versions. It wouldn't be because it's necessary, only because each version would be optimized for its intended platform.

Sacrificing compatibility for performance and performance per watt and having several versions of the chip to address the compatibility concern is a better way to go than designing a chip without optimizations for its intended platform. For example, having hardware acceleration and/or architectures and cache designs to give the intended platform excellent performance would be better than making a similar chip that is built instead to be compatible with multiple systems so it has to sacrifice more specific optimizations that would break compatibility.
 
Status
Not open for further replies.