Slashdot Mirror
Torvalds Slams NVIDIA's Linux Support
New submitter jppiiroinen writes "Linus Torvalds received the Millennium prize last week for his work on Linux operating system. He was already in Finland, so Aalto University arranged a talk session with him (video). During the Q&A, a person asks why NVIDIA does not play well with Linux. Torvalds explained shortly that NVIDIA has been one of the worst companies to work with Linux project — which makes it even worse that NVIDIA ships a high number of chips for Android devices (which use Linux inside). Torvalds even summarized that ('Nvidia, f*** you!') in a playful manner. What has been your experience on NVIDIA drivers with Linux?"
So why we don't moan and groan some more !!
I haven't had problems with NVIDIA cards since Redhat 5.2. ATI on the other hand...every time I try to install Linux on a laptop with an ATI video card, I end up having to futz with it for hours to get it to work.
"My immediate reaction is "WTF? What kind of moron doesn't make things 64-bit safe to begin with?" Linus
In Soviet Finland, individuals tell corporations to fuck off.
Nvidia cards are the only way to reliable game on Linux, either natively or through wine. Look at the winehq.org appdb for any game, then notice how most reported problems are on Ati video cards.
Case closed, unfortunately.
I have no experience with arm nvidia graphics drivers though.
I apologize for the lack of a signature.
In terms of GFX space, who provides better than Nvidia? I'm not saying they are angelic, but it seems to me that they produce working, viable, well sorted binary blob drivers for your OS Mr Torvalds.
Maybe I don't understand the problem, because it seems there are biger problems to complain about..
We`re all equal
Definately agree, nvidia single handedly has to be the worst company to support Linux. That, and the little Asian gizmo maker that only makes dodgy win32 drivers, I mean for f's sake if you're gonna make a gizmo make sure you make a driver for every single os on the planet, otherwise don't make one at all.
I have had to make a compromise in using the Nvidia driver. It's a "black box", so you don't know what is in it or how others might be able to improve on it, but on the other hand, it does the 3D work for stuff like KDE, Google Earth or 3D games like Brutal Chess or BZflag.
In Mageia, there is the Nouveau free driver, it works very well for 2D stuff, but does not work for 3D stuff.
So it depends on your requirements, and how wedded you are to the "Free" concept. Having said that, if there was a free driver that does 3D on Nvidia cards, I'd take it.
Take Nobody's Word For It.
Not to mention that ATI lets drivers fall out of support a lot faster...
Agreed. I don't know about the NVIDIA/Android problems but on the desktop NVIDIA has always been waaaaaay better than ATI. That may have changed in recent years, I don't know because I stopped buying ATI stuff a long time ago.
My karma is not a Chameleon.
I recently decided that I've put up with my Radeon for too long and dumped it for an NVIDIA GPU on my desktop. I like how AMD is more willing to work with the open-source community, but the driver they are working on isn't coming along fast enough and development seems more focused on adding support for newer hardware (not a bad goal, of course) rather then squeezing out more performance, and unfortunately with desktop environments like Gnome 3 demanding a decent level of 3D acceleration I found that the performance of the open-driver isn't up to the task of handling a multi-display desktop with decent performance with more than a couple windows open (it gets worse if I also want to play a video on my 2nd screen) on my Radeon 5850/4850. My 5-6 year old PC using just 2D acceleration felt snapper than Gnome 3 did with my radeon.. which makes me sad. NVIDIA's propietary driver does a lot better in that regard. Catalyst appears to be dropping support for all but the most recent GPUs, I'm not sure how well it does with Gnome 3, since they are so far behind with kernel and xorg support that I can never actually run it with downgrading a whole bunch of stuff.
I have had far less issues with any nVidia card than ATi/AMD. i have had both for more than ten years.
http://www.myspace.com/pmg-goa
Linus, f*** you!
nVidia worked well in the past on Linux. But since I got tired of all that Stallman and Torvalds crap, then I moved to Mac and now I'm asking myself why I didn't made it before.
when shopping for a PC over the last 10 years it seems that if there's a Nvida card then you know you're going to be OK installing Linux. Other cards not so much. I'm not a gamer so I don't hang on the bleeding edge and obsess over frame rates but when I work with 3D Nvdia seems to get the job done, even if it's closed source at least it's there.
From a developers point of view there are likely glitches inherent in working with closed software that need to be worked around to keep things stable, only speculating but I imagine this is what may be frustrating Linus.
"The Most Fun Possible on 4 wheels" is at SunBuggy in Las Vegas
I haven't watched the video, but probably Torvalds is pissed about the lack of contribution to open source projects (kernel, drivers). While their proprietary drivers are top notch. What will happen if one day Nvidia decide that making a Linux driver is too cost effective. Answer: a lot of unsupported video cars and SoC.
I understand the (bla bla bla) Intellectual Property they paid millions in research (bla bla bla) they don't want to give that information for free. But if the current architecture is so different than previous one, then why don't make public development information for previous generations of their chips as soon a new (different) generation is out.
While Intel graphics is still lacking, they are the only intelligent ones that have the brilliant idea of working the support for their next generation chips before those chips become available for sell.
on my office desktop.
While I still use the closed driver on my myth box, I find as long as you don't try and push it too far, and avoid flash all-together, life is peachy.
Those aren't drivers. It's just some binary files filled with random bits.
hey, why is everyone making the assumption that this is an ati vs nvidia debate?
it's not. there are so many more brands you should be considering when torvalds made this statement.
My NVIDIA 8800GT card, I get the only resolution of 320x200 and the other computer that has an ATI (not sure which one) has problems also, mainly graphics screw ups, not sure if it Linux or the card. Both cards under Windows seem to work better but the ATI does sometime flake out under Windows but not Linux, about once a month or so. Both systems duel boot the OSs.
lol nvidia suxx buttz!!!11
Nvidia is the only company where you can buy a GTX 675M and you can't use it properly with Linux.
Bumblebee is not effective at what it's supposed to do and the permanent open-source solution is still months away.
In a world of blind men, the one eyed man is king....
If I want to have decently supported video offload and remotely respectable 3D performance, nVidia drivers are about the only choice.
AMD drivers to this day cause my system to panic on shutdown attempt. MythTV's OpenGL painter and video renderer don't work correctly with AMD drivers, leaving me with video playback with XV and no recourse to sync to vblank. They do have XvBA out there, but I have to go into a more 'bleeding edge' xbmc and then be greeted by very bad artefacts with videos that are profile 5.1. AMD's open source interaction seems better, but none of the open source drivers come close to the 3D performance and notably no video decode offload is available.
Intel I heard great things about, but at least with Fedora 17 I can't seem to find the best way to get vaapi driver on there. All I see are requests to get it in being met with 'too messy'. It's also not in rpm fusion. I dug up a module from an old rpm and got vainfo running, only to find out rpmfusion xbmc build disabled vaapi support anyway, and only went with vdpau. Now I could recompile, but the point being that the larger community seems to not be bothering with trying to test Intel's solution as much.
Meanwhile, my nVidia system does vdpau beautifully, has pretty much no-brainer 3d support, and tear-free XV playback (even though I never use it anymore in favor of opengl rendering). Everything about the experience shows that both nVidia as a company and the userbase at large are developing and testing with nVidia primarily.
I could see as a developer being frustrated at supporting a kernel where a large portion is running kernel-mode code that you can't see, but from a user perspective, nVidia is about the only viable solution for Linux graphics.
XML is like violence. If it doesn't solve the problem, use more.
Linus isn't talking about gaming, performance or anything else like that. The point is : nVidia ships a binary blob and an obfuscated source portion that needs to be built outside of the vanilla kernel. That is what Linus is talking about, nVidia's lack of cooperation with the kernel people at integrating their drivers into the main line kernel in a way that respects the project's goals and visions.
Why you people are discussing the performance when that is not at issue, I have no idea. It was all pretty clear to me what Linus meant.
"Not to mention all the idiots who use words like boxen."
Anonymous Coward on Monday August 04, @06:49PM
I have nothing but problems with Intel's drivers not working right under Linux... I only wish that they were as stable as NVidia's drivers.
the worst. they either never compiled correctly or broke things horrendously when they did (my experience.) by comparison linux has open source drivers that competently handle 3d graphics for ATI/AMD, and they have for a while. That active development on such open source drivers was being considered at all, and not for NVidia, is a testament to Torvalds opinion on the matter. id like to think Intel screwed up once with their GMA500 chipset, which may or may not have been intentional to block linux from the hybrid tablet laptop market at the behest of microsoft, but even Intel works closely with the linux community in most cases to ensure their graphics chipsets work with linux, and work well.
Good people go to bed earlier.
Both systems duel boot the OSs.
Well there's your problem, having your OSes walk 10 paces, turn around, and shoot each other is bound to lead to problems. I would suggest trying to dual boot instead.
XML is like violence. If it doesn't solve the problem, use more.
I doubt he bases this comment solely on graphic card support, but for the whole way NVIDIA as a company handles linux support for their different kind of chips. NVIDIA makes a whole lot of chips that might not be related to graphics. He mentions android and nvidia chips in them causing huge problems lately.
Looks like you haven't try to use a laptop with NVIDIA Optimus technology.
Subject says it all. Yeah, I can't go look at their code, but to be honest, I don't need to because it actually works, unlike ATI's (well AMD).
Why should Nvidia subscribe to the projects "goals and visions"? Thats the projects concern, not theirs.
Jesus, I know LT is a good among geeks, but he is starting to get that syndrome where everything out of his mouth makes you want to roll your eyes and ignore him. If you want to bash someone for linux support, go trooping over to AMD/ATI's offices and loudmoth them. Last thing we need is you getting NVIDIA in a mood to provide EVEN LESS support than they already do genius!
Is he mad that they're producing drivers for Linux? Would he rather they didn't? Maybe he's losing his mind. Pity. Linux has finally gotten to the point where it is happily and easily usable ~ MintLinux, baby!!!
But if he starts attacking people who help the Linux community, but maybe don't do it the way he'd like, and they stop, or if he starts screwing Linux up, either on purpose or because he's losing his mind... I (and countless others) may end up having to switch to something else, like perhaps FreeBSD, which I've always regarded as a pain in the ass.
So if he retires, etc., is there any provision for someone taking over for him, whatever functions he is still performing? If he goes nuts, is there any way to force him out or fork Linux? If there isn't, I'd say Linux has a problem.
If Linux is Linus' own personal fiefdom, is it, and thus are we who use it, really and truly free? Free as in speech, I mean?
Hence why it's Linus (the project lead) that's talking about it ? He's voicing his concern... I don't get what you're even trying to get at here...
"Not to mention all the idiots who use words like boxen."
Anonymous Coward on Monday August 04, @06:49PM
Linus did give Nvidia the bird and a "fuck you" but he never slammed the quality of the Nvidia's hardware. His gripe with Nvidia is how hard it is for Linux to work with the company since they only provide a binary blob driver which makes bug fixing for it dependent on Nvidia's whims. Plus they refuse to even provide specs and API's for their hardware which make writing open drivers much more difficult and time-consuming because of having to reverse-engineer everthing to get a workable driver. In this case, Linus is absolutely correct.
I've had some problems with all three (I run OpenSuse) ... but NOT since I started downloading their own driver packages and building them on the target machine. I most recently did it with the AMD/ATI Radeon graphics on a new HP Probook. Both ATI and nVidia include configuration/tweaking software that will let me fine-adjust, and I just don't have problems with either.
BUT ... disclaimer: I haven't bought a *video* card in years. I always buy integrated graphics, because it's cheaper and does what I want. Your mileage will vary. But my Big Test(tm) of any graphics system under Linux is to see how well it runs Celestia. If it will run it at reasonable speed without flickering or other annoyances, I consider it Good.
Actually, I've had considerably more trouble with Intel's 9xx series of built in graphics than with ATI and nVidia combined. My previous laptop had the 945 graphics kit (I think I remember the number correctly) and I suffered through everything from flickers to outright hangs.
And the worst experience of all was with an old S3 card many years ago. Wow, what a piece of joy.
Cogito, igitur comedam pizza.
Why should Nvidia give a crap about what Linus thinks?
I concur with how terrible AMD support is with their Linux graphics card drivers. On my main office desktop, I'm running an ATI 4890 and the 2D acceleration is so bad I had to just turn compiz off and go to a non composited desktop. I've never had that problem with nVidia.
The soylentnews experiment has been a dismal failure.
Something isnt right, that card is pretty much works with no effort at all. Does you distro package manager provide an easy installer for the Nvidia driver?
I think his point is that Linus should be grateful that they even support his project at all, rather than bitching at them. If it costs them money to support Linux and they're damned if they do, damned if they don't, then what incentive is there to continue even supporting them at all?
I think a large part the problem is that Linus has a way of sounding like a dick. There are much more diplomatic ways of saying nVidia could certainly do better on Linux than the way Linus goes about these things. No one is suggesting nvidia couldn't do better, but when he attacks a company over this sort of thing it just makes him sound like an ungrateful prick.
I never had a problem with NVidea's driver ... ;) ).
(Other then the fact the are close source
Ok: maybe one or 2, but a quick update fixed that.
...NVIDIA blob drivers have always been rock solid in my experience, for both their GPUs and chipsets. Not only that, they're just as feature complete as their Windows counterparts. ATI used to do the same thing and their drivers weren't nearly as reliable.
From the kernel developer point of view binary blob drivers are hell. Yet, NVIDIA managed to provide some of the best hardware support available on Linux with it.
I'm "getting at" *your* opinion about what Linus is supposedly talking about - why should Nvidia give a toss what the projects goals and visions are, that has nothing to do with them providing support for their products (or not) on that platform. Some people seem to think that Nvidia should not only support the project (Linux) but the philosophies as well - why should they? It's not their vision or goal.
NVIDIA's driver is not open source, but the support is great. The drivers work, the give few issues, 3D, CUDA and OpenCL are fully supported.
Try that with AMD drivers.
There is a quote (somewhere, I cannot remember who said it), that the NVIDIA drivers have more LOCs than the Windows OS. So it is a huge piece of software... not so easy to "just code it again".
Thanks to the Nvidia drivers, I can play many games in Wine, have my graphics card "just work" without pain, and use fast accelarated OpenGL for several programs and my own experiments. These drivers are fantastic. WTH does Linus find wrong about them?
My personal experience with nVidia drivers on Linux has always been pretty smooth actually.
weren't the last two major issues with the proprietary (NVIDIA) drivers for linux remote security exploits? They were fixed but just sayin, i'd rather use nouveau than get pwned for pretty graphics.
If even the 'leader' is bashing you and your efforts to support, and telling you to f-off, why bother ?
If i was Nvidia id just repeat a loud "f-you" back to him, and stick with supporting actual paying customers.
---- Booth was a patriot ----
Linus sounds a lot like Theo here... old age, fame and financial independence getting the best of him?
Doesn't sound like the same guy that i "talked" to on usenet so many eons ago.
---- Booth was a patriot ----
I've never had a problem with nVidia's drivers, both windows and Linux. They've always worked consistently and perfectly for me. Furthermore nVidia seems to make real effort to support all features they have in windows drivers, while ATI's Linux drivers have usually been a bare minimum version of what they have under windows.
Even then, I have always had issues with ATI's Linux drivers. One of my laptops that had an ATI chipset wasn't even ever supported under Linux by ATI at all. I know linux isn't the most popular OS, but its use is pretty significant these days, and ATI ignoring that is ridiculous.
Consequently I always make a point of ensuring any motherboard or laptop I get has nVidia embedded graphics rather than ATI, even though the majority of laptops seem to use ATI so its not always been easy.
Would it be better if nVidia was altruistic enough to not use a blob and go fully opensource? Sure.. but I understand that in the real world companies sometimes have to protect their intellectual property.
What really matters to me is my day-to-day use of the drivers, and on that count its been consistently nVidia 10/10 ATi 2/10 so am I really effected by the fact that I can't see the source? no.. its not like I'd ever want to hack on it.
Disclaimer: I am in no way connected to nVidia other than as a happy customer.
Linus wants to have his cake and eat it too. If he really was worried about this issue the "right thing to do" would have been to work toward migrating the kernel to GPL3.
Repeal the 17th Amendment TODAY! Also Please Read http://www.gnu.org/philosophy/right-to-read.html
Ditto. No problems downloading, installing, or using NVidia drivers.
And I use Slackware.
I've had problems with a NVidia card that I have, and the last time it gave me problems was with the latest upgrade to Kubuntu 12.04. With this upgrade, NVidia's very own proprietary driver either rendered 3D scenes excruciatingly slow or never rendered them at all, instead presenting only a black window. Strangely enough, 2D rendering worked without any noticeable hitch.
The fix came only about a month ago, with an upgrade to NVidia's proprietary driver.
Slashdot, fix your code or at least hire someone who is competent at it to do it for you.
And if you either judge this out of context, or actually watched the thing and still describe it as you did, you're dumb fucking sissy.
...every time I try to install Linux on a laptop with an ATI video card, I end up having to futz with it for hours to get it to work.
That's a normal part of Linux deployment.
Because the nVidia driver forces you to a trade off. If you want KMS, you can't use it, and KMS is a much better experience for the end-user, which nVidia should care about if they want me to buy their graphics cards (interestingly, I've got an ATI card on the way right now because they support my triple monitor setup under Linux in a much, much better way than nVidia do).
-- Lattyware (www.lattyware.co.uk)
It reminds me of a brief political exchange a few years back: 1. Your president lied to get us into an illegal war that is going to end us costing us a trillion or more dollars. 2. So? Your president lied about a bj!
What I've noticed is that NVidia usually works fairly fine. Until that day you upgrade the kernel, and don't recompile the shim, and it craps out on you and dumps you back into a VGA console (if you're lucky). Or when you upgrade the driver and it clobbers everything suddenly.
The second problem exists on all platforms; I use Win 7 at work, and all of a sudden, the driver tends to crap out for no apparent reason. Usually, after an upgrade, and you either have to downgrade, or upgrade again within a few days. I think they keep releasing minor fixes and patches for particular users - games and the like, which tend to create a LOT of regressions. And don't get me started on the multiple, backwards-incompatible changes made to Cuda and OpenCL. People talk about ABI compatibility in the kernel, but compared to the (lack of) API compatibility in Cuda, that's chickenfeed. At least the kernel has a stable API, and the tools you need to work with it are in the same place over releases!
Of course, ATI drivers are even more flaky! And Intel drivers tend to claim OpenGL features they don't actually implement. But that doesn't excuse NVidia's behaviour either. It just means that all three are horrible companies, and that the GPU market really needs a shake-up.
Which one answers questions most clearly?
And while I understand Linus's point, nVidia has been really good at supporting their drivers. Every bug I found and reported had a real person working on it, and soon enough it was fixed. This is a lot better support than what I get with some open source projects that are not as well funded as nVidia devs. So, binary blob = bad, but nVidia does make up for it by actually supporting their stuff.
Hours? Try days. My last several computers all had nvidia, and "it just works", right up to the last (GeForce 9600). My latest has an ATI HD and I still haven't been able to get the audio portion of the drivers working right. At least I finally got the right magic combination of drivers for 3D acceleration to work, but it was a gigantic pain. Maybe my experience would be different if I had bleeding-edge hardware, but I always save money by buying a year or two behind the curve so nvidia is all right by me.
NVidia isn't required to, as you put it, "subscribe to the projects goals and visions".
But Linus Torvalds is also not required to enjoy or approve of NVidia's policies, particularly when they generate bad publicity for Linus Torvalds' project and also cause a number of people to complain to Linus Torvalds about a problem which he didn't caused nor can he do anything at all about it. Hence, a very appropriate and sorely required "Fuck you, NVidia".
Slashdot, fix your code or at least hire someone who is competent at it to do it for you.
This is my experience. I have a new AMD A4-3400 and tried to run Xorg and after several hours of bug search I found that there is the old blackscreen bug which is well-known since 2009 and even now unsolved. This makes many Radeon HD cards und IGPs unusable.
After inserting my good old nVidia, everything worked fine.
I agree partially... both manufacturers deserve a "Fuck you!" for not supporting Linux and other systems(!) properly.
Did you watch the video ? At all ? Linus is voicing is concerns, not telling nVidia how they should operate. He's describing the problem basically... What nVidia thinks and does is out of bounds and plainly obvious, you're restating what Linus is stating...
Hence why I don't get your point.
"Not to mention all the idiots who use words like boxen."
Anonymous Coward on Monday August 04, @06:49PM
If it costs them money to support Linux and they're damned if they do, damned if they don't, then what incentive is there to continue even supporting them at all?
Because they feel the Linux market is worth supporting. They don't do it to make Torvalds happy.
The "right thing to do" is create an open-source computer. This system can be constructed by anyone in their backyard without having to worry about a proprietary toolchain.
That won't happen, since it requires everyone to have a personal manufacturing facility, complete with microchip clean room.
I see lotsa people claiming that Nvidia Optimus is not supported. For what is worth, last month I helped a friend install Ubuntu and https://wiki.ubuntu.com/Bumblebee on a new Asus laptop, and it worked perfectly well , including CUDA support (that they use for numerics in math finance)
I don't know about the NVIDIA/Android problems but on the desktop NVIDIA has always been waaaaaay better than ATI. That may have changed in recent years, I don't know because I stopped buying ATI stuff a long time ago.
So, you've no idea what you're talking about. Good to know.
What I know is when nvidia came out, I was seeing thousands of posts from people desperately seeking answers on how to get them to work, and thousands more on how to make their X Window survive upgrades. Nvidia is the reason they came up with the mantra, "Don't use a GUI package manager!", because if it upgrades X, it'll kill the pkg. mgr. doing it.
Meanwhile, I've been using ATI for more than a decade with zero problems past their initial teething stage. However, I don't care about 3D, gaming, Wine, or MS Windows. I have no trouble playing DVDs, youtube, or flash. IF I think there's anything wonky going on with video support, I run "X -configure" as root and the wonkiness disappears.
I will never buy a box with nvidia in it if they continue this way.
"Tongue tied and twisted, just an Earth bound misfit
I suspect what NVIDIA is trying to do is impede the development of compatible chipsets. They don't want to encourage another competitor to get into the video chip market. They already have all the competition they can handle. So what limited information they'll give you is wrapped in NDAs and Linux developers are in no position (and of no mind) to give anybody an NDA. So this makes them really difficult for FOSS developers to work with.
But for Microsoft, who has the bulk of their market and CAN control the distribution of proprietary information, they'd be willing to share their specs and details of how their hardware works.
He made the kernel. What did you make? Your shithole?
And here is where the /. crowd divides.
/. readers want to live in, is one we're not ready for yet. So unless the guy finds a way to keep companies and people happy, I don't see an alternative. Either they use one or the other. Now, nVidia may have patents, but not much well known litigation history, because they do have protected their secrets/algorithms, etc, within their code. Wouldn't be easy for ATI/Intel or any other GPU player to come and pick up their algorithms and use them? They paid their engineers to optimize their solutions to work with their products, and they support their products, why should they support other products ideals and goals is beyond me. But I guess Linus is fine with his rant.
What's your take on a company that want to keep their lead in the market? Should they patent everything or keep trade secrets?
If it's none of the two, then what do you think companies could do to keep that leading advantage?
The utopia some
Linus wants to have his cake and eat it too. If he really was worried about this issue the "right thing to do" would have been to work toward migrating the kernel to GPL3.
Does this have anything direct to do with the story, or are you just using it as an excuse to give a bee in your bonnet some fresh air and exercise? (*)
Besides which, as far as I know, it's unlikely that the kernel *could* move to GPL3 even *if* Linus wanted it to, as it would require the permission of every single one of the countless contributors and/or the replacement of their code.
(*) This sounds like the typical "why are people doing [anything that isn't a cure for cancer] when there's still no cure for cancer", with "cure for cancer" replaced by something that *not* everyone necessarily agrees is a good thing anyway.
"Slashdot - News and Chat Sites Deviant". (Click "homepage" link above for details).
nvidia graphics cards under linux are a dream. anyone who ever had the misfortune of being forced to use an ati card under linux will agree. i don't agree with their binary-only policy, but my experience with their linux drivers over the last 12 years has been 100% good.
Go back and read your own post - I'm responding to your points, not the videos.
The new CUDA libraries are only for specific versions of Linux and unless you have a whistle clean install you can never get the *&#^$ Drivers to work. I have had a really tough time getting the GPU interface drivers working for data intensive processing. I do think they are really amazing chips - but the support is very narrow. my 2 cents
He made the kernel.
And I care because?
I made no points. I explained to people what was being discussed in the video. Why are you responding to me ? I framed the context. You don't agree with the context I framed ?
"Not to mention all the idiots who use words like boxen."
Anonymous Coward on Monday August 04, @06:49PM
Same here. I gave up on Debian ~ 2.x when I could not get it to work with my ATI 128 thingy. It was frustrating, and I did not have other options since it was a home PC and I was n00b.
I am still a n00b, but since that time I stopped buying any Laptop or Desktop with ATI in it...
I still cannot understand why the Linux community does not lobby for a good video card series that plays well with Linux and is on the high end for gaming. Most PCs can run office applications, but v few can run games. Most tech savvy customers would want a PC that has good graphics for games. Linux will not be able to dive into the Home PC market without that.
Overall my experience has been ok, except on laptops. I am still not quite sure if it's hibernate/suspend itself in the kernel,or the interaction with Nvidia drivers. Under both Ubuntu 10 and Fedora 11/13 I had issues. Both Dell and Lenovo T60. On Ubuntu 10 it was so bad, that some menus got sticky from time to time, they would stick to the desktop. You would have to kill the app. Kind of annoying. So far this hasn't happened on Ubuntu 12, and - knock on wood - the suspend has been muchj more reliable. Oddly enough Ubuntu has disabled hibernate, but that's another story.
So,overall, my experience has been ok, but could be better.
Your argument would make sense if nVidia didn't use Linux during the development of their products.
I really don't want to have to go back to ATI hardware on my Linux machines! Nvidia may suck, in that department, but everyone else sucks worse!
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Why you people are discussing the performance when that is not at issue, I have no idea.
NVIDIA's graphics drivers for Linux are full-featured and rock solid.
In this less than perfect world, users will have needs and values and a willingness to compromise which will come into conflict with the loftier "goals and vision" of your project.
Ok so you don't know how to setup graphics on Linux because I have had an 8800GT working for last four years under multiple flavors of Linux with absolutely excellent results at 1920x1200 resolution.
AntiFA: An abbreviation for Anti First Amendment.
There was also one person from the NVIDIA in the audience :)
Linus actually opens up the reason, why he made a comment with the word:
http://www.youtube.com/watch?v=MShbP3OpASA&feature=youtu.be&hd=1&t=1h00m25s
The problem isn't that nVidia doesn't put out functioning binary drivers. They do. Their binary ones are much better than ATI ones IMHO. The problem is that the binary drivers don't always survive kernel updates So people who rely on nVidia's proprietary binary drivers can't always update their kernel or they lose their graphics until nVidia puts out an update. Now, there are open source drivers for nVidia that get updated with every kernel update; however, nVidia does not put out enough specifications so that these drivers are able to use full card functionality like full 3D acceleration. It wasn't until recently that the open source drivers have been able to reverse-engineer some of the functionality.
Well, there's spam egg sausage and spam, that's not got much spam in it.
The argument you're having in your head is impressively nonsensical.
Microsoft has all the code for Apache. Why don't they use this to create a competing product and take over the web server market? Or more on point, Intel open-sources their graphics drivers. Why hasn't ATI eaten their lunch yet?
As a general point, IP protections have never been necessary to commerce and never will be. In the context of this discussion, you are wrong in every particular.
> If it costs them money to support Linux and they're damned if they do, damned if they don't, then what incentive is there to continue even supporting them at all?
It doesn't.
I think people nowadays just don't understand what Linux is anymore.
This discussion has nothing to do with desktop, performance or even servers.
Nowadays, Linux is shipped on 900000 phones every days and that's a very lucrative business.
Nvidia has a share of this with its Tegra chips but their way of supporting it makes it a pain in the ass for others inside the kernel development community.
From that point, Linux complaining is totally fair. PR is basically his only weapon.
Should Linus Torvalds feel grateful every time someone goes to him to complain about how linux sucks because NVidia's proprietary drivers aren't working properly, or don't work at all because they were developed by NVidia for version X of the kernel instead of version X+1?
The only total solution to this is a completely open source GPU. Ah, but the Open Graphics Project tried that and failed! But no. They actually produced and sold real hardware. It just wasn't sustainable, because the FOSS community didn't step up, creating a self-fulfilling prophecy of failure. In any case, the OGP isn't dead. They've recently started moving again, this time with much more realistic goals, which is to develop a complete working GPU module that can be licensed. If they make the BEST GPU, then companies will license it. If not, they'll have the best GPU model for academic research into GPUs, which is also, as it turns out, their PRIMARY goal.
i dunno what could be better. am i missing something?
i can compare nvidia on winblows and linux and
if i need to make a linux computer I just buy nvidia. ONLY!
GO nVidia!!!!! GOGOGO!
p.s. maybe soemone didn't consult their RDP (http://en.wikipedia.org/wiki/Recreational_Dive_Planner)?
Wow, I have never seen a reply that so completely missed the point. Linus wasn't talking about nvidia's driver quality, he was talking about their lack of participation in open source, specifically in relation to the Tegra chipset that is widely used in Android phones, when it is Linux/Android that enables their profitability in the current market. In other words, nvidia is an ungrateful selfish dick about the whole thing, hence his outburst.
Here you are talking about desktop video cards and driver quality, for a relevance score of 0. AMD wasn't a part of this discussion.
I'll give you the benefit of the doubt and assume you are just too ignorant to watch any significant portion of the talk before writing a response to it. Otherwise you're an idiot.
The source of the problem: drivers for integrated stuff that Nvidia makes for phones, tablets and laptops.
177 bullshit comments about graphic drivers for discrete video cards and only one that mentions Tegra. I'm sorry to say, but it is amazing how clueless slashdot has become...
Serously, intel is ok for office work. For anything video or gaming related: nvidia binary blob.
No Freedom though.
So Linus cannot voice his opinion? Sure nVidia doesn't have to support Linux at all; however, someone asked Linus for his opinion and he gave it to them. Linus did not go out of his way and hold a press conference to rail over how bad he thought nVidia was. This was during a Q/A session. This is not a black and white thing. nVidia does go out of their way to support Linux with binary drivers, but at the same time they don't work very well with open source developers like Linus.
Well, there's spam egg sausage and spam, that's not got much spam in it.
nVidia's drivers do not support kernel mode setting.
You get open source drivers.
It's just fine for most business use (outside of 3d modeling).
It's probably OK for light gaming, too.
Works out of the box, no driver download necessary.
I'm not a lawyer, but I play one on the Internet. Blog
> What I know is when nvidia came out, I was seeing
> thousands of posts from people desperately seeking
> answers on how to get them to work, and thousands
> more on how to make their X Window survive upgrades.
Yeah, and they solved that problem. An entire module rebuild facility for kernel upgrades was probably developed just for Nvidia.
That benefits ATI blob drivers too. This is a good thing since you are unlikely to get suitable performance (if you care about that sort of thing) without the blob driver.
A Pirate and a Puritan look the same on a balance sheet.
nvidia doesn't just make graphics cards. They make ethernet and other chipsets for mobos, which can be very problematic with linux.
http://soylentnews.org/~tibman
So, you're an asshole. Good to know.
> My NVIDIA 8800GT card, I get the only resolution of 320x200
That's funny because my 8400 had no problem driving my monitor to it's native resolution. Even the libre drivers were able to do better than 320x200. Their main problem was performance.
A Pirate and a Puritan look the same on a balance sheet.
Nvidia drivers were actually the driving force behind the concept of kernel tainting where an oops dump reports that non-free modules were in use. Nvidia drivers were creating a huge number of oops reports that couldn't be debugged due to being closed source.
This is only true if your definition of "full-featured" does not include KMS or complete XRandR support.
It's not even that. The driver will sort itself out automatically.
A Pirate and a Puritan look the same on a balance sheet.
I like Linus. He speaks his mind, and that is good. He does not strive to be a politically correct suck-up like most of the soulless corporate speaking heads you see all over the place.
He has every right to say "Nvidia, fuck you". How should his message be sugar-coated? Should he write a 500-page NY Times bestselling book about the matter? Hold a seminar? Issue a press release? Have a meeting with Nvidia CEO, CTO and CIO, presenting empty Powerpoint fluff and wanking around the issue in such abstract terms that it can be interpreted in any which way, after which everyone thinks they've done their part but nothing happens as a result? No. It suffices with three small words. Why waste more time and effort? To not hurt someone's feelings? Don't be such a baby.
I think part of how Linus comes through as he does is a cultural thing. Although he has lived in the USA a lot, he's still a Finn. If you need to deliver a message to someone who is not behaving, you deliver a message, wrapping it up in a pink box with a greeting card full of hearts is pointless. And let's face it, Nvidia hasn't been a model citizen - if you're a dick, don't be surprised that others are dicks towards you.
And you ask: what incentives does Nvidia have to support Linux? Well, how about not making life hard for the people who pay actual real money for Nvidia's products? And not making life hard for the people who try to support the Nvidia products on a great OS on their free time?
NVIDIA's drivers suck. And AMD/ATI's drivers aren't much better. They both need to get their act together, and soon, unless they want to be replaced by open-source drivers, or (heaven forbid!) another company's graphics.
The actual Nvidia driver works fine, the nueveou one not so much and it quite often has to be blacklisted (at least in ubuntu) to install the proprietary driver (that just works).
ATI video drivers suck for both windows and Linux although windows drivers are very slightly better for windows. In fact they're so awful that I never plan to look at another ATI gpu unless I see tangible signs of change. Right my primary system is a notebook with an ATI 4850 but this will be replaced shortly by a hex core i7 and high end nvidia gpu.
ATI is so bad that this is the only time that I can remember looking forward to new driver releases and just going that years old bugs were fixed let alone having fee bugs than last months release.
Then Linus should make sure the driver interface remains stable or shut the fuck up.
Which, if taken to heart, would kill what little remains of Linux 3D gaming. To the server racks with ye, Linux!
For me the nvidia-supplied drivers "just work". The ATI ones on the other hand have just been a cause of constant chaos. I've always had better luck with the drivers supplied by the hardware manufacturer than those written by someone who whilst very clever often doesn't actually know how the hardware in fact works and is therefore trying to work backwards (or in the rare cases of having a decent data sheet, reinvent the wheel). In work I've gone back to the intel supplied e1000e drivers as after a few days of being an MPI node the system would lock solid. I couldn't even tell if it was a panic as the screen wouldn't even display anything.
This is only true if your definition of "full-featured" does not include KMS or complete XRandR support.
They added XRandR 1.2/1.3 support in 302-series.
See e.g. http://www.phoronix.com/scan.php?page=news_item&px=MTEyMDk
You are running a binary blob. I've run ATI cards since '93 (starting with linux v0.9/ Xfree86 v??) without binary blobs (years before RH even existed). ATI/AMD has been helping in the effort for free drivers for a few years now. This should be supported, if you really give a damn about the platform you are using.
Nvidia closed binary blob drivers are beside the point. Most folks (who matter; e.g., those developing the platform you are using) won't touch them. Nvidia provides zero info/help towards open drivers, so yes, what Linus says is all true.
It is _your_ defense of nvidia's current practices (and that of people like you) that encourages an otherwise good company that uses Linux + some Solaris exclusively in their hardware design process from supporting the folks that make their bread and butter possible. The only windows in Nvidia's data center is the word printed on a large poster to make M$ happy.
The lack of cooperation goes both ways. Integrating their driver with the kernel requires that they relinquish their source code, and they would prefer not to do that, and that is their choice.
ATI often has better spec'ed cards than Nvidia, but the later has vastly better performance. The drivers are their competitive advantage, and they simply will not, nor should they be expected to give up that competitive advantage over less than 1% of the market.
Lack of cooperation also includes purposely keeping the kernel API/ABI in flux as a means to strong-arm vendors into relinquishing their source code, read Greh Hartman-Kroah's "stable abi and other nonsense" or whatever it's called, README if you don't believe that it's anything but intentional. (hint: retaining ABI/API compatibility is not impossible as is suggested, every other OS does it, and FreeBSD is able to not only retain ABI compatibility with itself, but also with Linux). Maybe if the kernel team was less ready and willing to break things they know vendors hook into and rely on, vendors would be more eager to cooperate, just saying.
I haven't had problems with NVIDIA cards since [..]
Wait, what? You actually managed to find a Video Card for which Linux had drivers and they didn't crash nor fuck up your window manager? I swear, I am serious, I have never in my life seen a Linux desktop distribution run smoothly on a computer with video drivers and games and effects and all that crap I keep seeing on YouTube. I often drool for it, but I never see it myself. I installed Linux at work recently. I couldn't find drivers for the video card, because I don't know what kind of video card it is. I had never seen Linux detect and use any kind of video drivers in my life. I've installed it on a couple dozen computers in the past decade, new computers, old computers, new distributions, old distributions, commonly known ones, obscure ones, drivers, third party drivers, etc.
Seriously, until Linux gets all its shit together, it's laughable as a Desktop operating system. With Windows, you buy a computer, install Windows and it works. And I still prefer Linux over Windows, but that's only because I'm a technical person.
Just a little nitpick:
They feel the 3D modelling, animation and effects market on Linux is worth it. They don't give two shits about spinning cubes or tux racer.
It actually works as good as on windows, what else should we complain?
Please stop citing Intel or VIA etc releasing full spec & source - their video chips are TOYS and not worth a penny!
Kernel panic!
from 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
to 45 2F 6E 40 3C DF 10 71 4E 41 DF AA 25 7D 31 3F
I understand his frustration, but the OS community hasn't produced a decent 3D driver for any card that I'm aware of.
The 3D might *work* on a few chipsets, but it's nowhere near the speed of vendor supplied drivers.
If it wasn't for Nvidia, Linux honestly wouldn't even be an option in the workstation or GPGPU market.
I can't say the same for AMD/ATI...... My latest computer has my first AMD/ATI graphics card, and much of the time I have had no problems.
The problem is that the AMD/ATI drivers are that they seem to be very software specific.... so each time either kernel/X.Org changes a little I can't compile the driver :( I never had that problem with the Nvidia cards/drivers.
I'm not idealogical enough to care whether the drivers are open-source.... I just want the drivers to work ALL the time.
It is too late to ignore linux 3D markets now. In addition to AMD and NVidia, now Intel and PowerVR make good enough 3D acclerators.
Try to play with Intel's HD 4000, you might be surpised. What about PowerVR? Oh, just about any ARM-based mobile phone or tablet uses PowerVR to display 3D graphics.
Microsoft has all the code for Apache. Why don't they use this to create a competing product and take over the web server market?
Because IIS does a lot of things Apache doesn't and integrates into the rest of their stack while apache doesn't. While as an application server with web-server capabilities rather than a straight-up webserver doesn't even target the same market as Apache. Never mind the slew of third party stuff built on ISAPI. It's not worth, neither in terms of development cost, support cost nor time, and there's no practical gain, Apache doesn't actually bring much to the table for the market Microsoft is targeting with their products.
Or more on point, Intel open-sources their graphics drivers. Why hasn't ATI eaten their lunch yet?
Um, because the source code for drivers for Intel's hardware is practically useless for ATI's hardware? What are you expecting, that ATI build better drivers for Intel hardware than Intel, and therefore destroy them on the graphics card front?
As a general point, IP protections have never been necessary to commerce and never will be.
Except that you've done nothing to support this conclusion. Look at all the cheap open source knockoffs of proprietary software. Without IP protection, do you think they would still be cheap knockoffs? Do you think competitors (in the same market of course, and in software markets, obviously) wouldn't just treat their competing products the same as Autodesk treats everything it acquires (gut it and put the good stuff in 3DS Max)?
How well do you think companies like Adobe, Corel, Native instruments or Autodesk would fare with no IP protection? Just because you can frame a pair examples where it makes zero sense to "borrow" IP doesn't mean there are not hundreds of examples where it actually makes sense competitively to do so, and IP protection is the only thing stopping them?
Oracle is eating Red Hat's lunch specifically because of there being flimsy IP protection. The whole purpose of OUL was to rebrand RHEL and "bleed them dry" (Larry's words, not mine) on support, and it's worked quite well over the years.
...who thinks Torvalds regularly acts like a arrogant total dick?
Please send me a video where he acts in a way that gives me a feeling of sympathy. Because while I obviously respect him for Linux, with every interview/speech I can stand him less...
There's a reason for keeping powerful tools out of the hands of the common people.. Just like with weapons, you can let them have their snub nose 38, but not a 60mm gatling gun
Linus doesn't trash nVidia's Linux support, he trashes nVidia's cooperation with the kernel team (or more precisely the lack thereof)
Right now the nVidia binary blob offers the best performance and stability of all drivers on Linux, by far.
That being said it is (if you believe the guys who do the actual work in the kernel team) a PITA to support because nVidia doesn't cooperate.
Bottom line, Linus is pissed because nVidia doesn't make it easy for the Kernel team to implement their binary drivers, so they represent a lot of work and yet they contribute very little to Linux. They are a hassle. And I think he is in his right to tell them to f*** off because they have the ressources to make it a much nicer situation.
There's a couple of points here.
1. Should the driver API be stable? Usually not: a long term stable API means that it's hard to make changes when necessary. For example, the Linux USB subsystem has been substantially re-written 4 times. This is never a problem, because when someone changes the kernel API, they also are responsible for updating all drivers in the tree that are affected by it. It means that Linux's USB support isn't hamstrung by lots of legacy cruft.
2. Should the ABI be stable? No. This breaks pretty much every time the kernel is compiled. The solution: recompile the drivers.
The only problem is that Nvidia don't want to play by Linux's rules: they want to ship a binary, rather than source driver. The result is that Nvidia (and nobody else) gets snarled up on the API/ABI policy which works exceptionally well for all the open-source kernel drivers.
The argument for a stable ABI is rather like the (claimed) USP of Java: write once, run anywhere. The thing is, if you can recompile on your target platform, that advantage disappears. Why would anyone care about binary compatibility if they can have source-compatibility?
Aside: there is a very different logic with the Posix (system-call) API between userspace and kernel space. This means that, for example, fopen() works exactly the same way now that it did 20 years ago.
Yea- this is exactly why those who give a rats ass about improving support for GNU/Linux and free software don't go with nVidia or ATI graphics chipsets. Intel graphics are nearly as on the low end as nVidia now and there more than sufficient for the majority of people. They work better on GNU/Linux than nVidia because you don't need that binary blob. The Intel graphics drivers work across distributions much better and in a number of distributions nVidia cards can't be supported at all due to restrictive licensing.
There is unfortunately only one company who is actively trying to eliminate the non-free bits targeting consumers. Check out ThinkPenguin.com and one thing you will notice is they support EVERY single product on Trisquel. Trisquel is a completely free distribution. There are a few others who project themselves as being freedom friendly and will ship with Trisquel. However they are actually shipping systems which will never work well with Trisquel due to the fact the hardware is dependent on non-free software.
For the most part everything works pretty well out of the box as the free software community is able to support the chipsets being used and GNU/Linux users aren't dependent on the manufacturers for support. Manufacturers generally don't provide very good support for any operating system. If you look at Microsoft Windows you see the same problems. When people move from one version of Microsoft Windows to the next they end up having to replace old hardware. Not everything of course. There are some devices which are standards compliant (some printers, most digital cameras, most DVD drives, most USB flash drives, etc). On the other hand most people don't have only standards compliant or freedom friendly hardware today.
People need to make conscious decisions on what they purchase if the situation is going to improve. Many people are doing this already. Many more need to start doing it though.
At least the binary GPU drivers work. I had to move off of nouveau because every time I resumed from suspend the GPU fan sounded like a UFO taking off. Granted that may arguably be nvidia's fault by not giving the nouveau dev's what they need to write the drivers properly.
But, I haven't had issues getting nvidia cards working under EFI either; particularly those in Mac's. On the other hand the only way I have gotten ATI cards to work properly is by using fbdev. I have seen supposed ways to get this working by dumping the bios when booting in legacy mode and then using loadbios in grub 2, but have never personally gotten it to work.
So, they may not be great to work with, but at least they're throwing us a bone, which is more than I can say for ATI in most cases.
Nvidia was a decent company years ago. However, they got distracted with more ways to make money by screwing over the customer. When nvidia had just started using SLI and Amd started using crossfire, They started making it a pain to use the video cards. The problem then was that supposedly "you can't run sli on motherboards that had support for crossfire" It was all b.s and a scam. You could run sli on crossfire boards and vice versa. People found ways around it but nvidia rather sell it's motherboards that was marketed for only being able to run sli. In short they sent a bunch of cease and desist letters and it's no surprise that they have lack of linux support. I'm surprised they even have drivers for linux, With the whole motherboard tactics they used I'm surprised they don't just drop support and issue everyone cease and desist notices. I refuse to support Nvidia and their nazi like regime. They aren't as bad as sony with requesting youtube and twitter accounts but pretty close.
Have to agree with this...
NO PROBLEMS!
And nVidia is the ONLY VIDEO CARD I choose, period.
I will NOT choose ati even though AMD owns them, and I ONLY USE AMD processors.
Linus can whine all he wants, the fact that nVidia has told the kernel to take KMS and to stick it! nVidia asked for a change to help support that abomination that is optimus etc. and the Kernel Illumanti got their hoity tootiy tudes out and and said no... no you can keep the the party line about this or that on GPL, open source ...
I do NOT CARE.. I use ONLY the OEM drivers, and thats all I will ever use. Open source is not the be all end all reason as to why I use Linux... and if Linus continues with his idiotic drivel I just may be a BSD user!
I have a response for Linus... F*** YOU!
1311393600 - Back to Black
As has been noted here the nouveau driver is pretty good as far as my various flavours of Nvidia cards have used.
Sound through HDMI does work under nouveau as I understand it forcing you to use Nvidias proprietary drivers.
Since I refuse to use proprietary code wherever possible that restricts me to non HDMI devices.
Anyone else have other experience?
At least some nVidia drivers do not play well with legacy Motif/X apps (using lesstif or OpenMotif); RedHat does not play well with OpenMotif, for that matter.
"My opinions are my own, and I've got *lots* of them!"
Wouldn't his problem be solved by going GPL v3?
I love that Linus is still as forthright as ever.
The fact that Nvidia don't support current fundamental features of their chipsets on Linux (like Optimus) is reprehensible. Thanks you to Linus for providing the "fuck you" to Nvidia that they deserve.
Ok so you don't know how to setup graphics on Linux
We are WAY past that being a valid statement. It isn't 1990 anymore.
This story starts more than a decade ago. There was a hugely popular software vendor concerned that maybe one day people might choose to not use their software. They had vast sums of money and controlled access to the immediate future for software and hardware vendors alike.
Foreseeing a potential difficult future they chose to defend themselves in a particular way. They formed subsidiaries they controlled and gave them patents, and filled them with developers skilled in the finer (and secret) nuances of how to interact with their software, and they kept them informed with advance knowledge of how it would work in the future.
These subsidiaries approached hardware designers with a simple message: they would accept the patented technologies and use them; they would let the subsidiaries write the drivers that had special hooks into the software; they would do this under non-disclosure and never tell - or they wouldn't. If they accepted they would not be able to publish open specifications about how their own hardware worked because that would be exclusively cross-licensed with the subsidiaries in exchange for access to the patents. The hardware makers who wouldn't play along wouldn't get as good compatibility with the big company's software, nor inclusion in their distribution CD and OEM images. The refusers would be plagued with difficult installation, buggy drivers and unhappy customers and fail in the market. The software would change in ways the refusers could not predict, but the accepters could. Some accepted, and some refused. Those who accepted survived, those who refused mostly died.
This has continued to the present day and as the hardware has evolved the agreements persist in ways that are now not removable.
Nobody involved in Linux wants hardware manufacturers to write the device drivers for them. They only want open and clear specifications for how the hardware works so they can make their own drivers. They aren't going to get that from NVidia, nor ATI, nor any others whose technology is intertwined with this compromise from yesteryear. This boon is now beyond their ability to grant without starting again from the beginning.
Help stamp out iliturcy.
It isn't damned if they do, damned if they don't. That is like your wife saying "You bitch at me if I bang your brother, and you bitch at me if I bang your best friend. I'm damned if I do and damned if I don't.".
nVidia has a third option. Support drivers on the OS properly.
nVidia is the one being a dick. The GNU/Linux ecosystem relies on freedom to function. Linus manages a *mostly* free project and has stupidly let these companies dictate the rules. We don't need nVidia. nVidia needs us. We are the customer. While I can sympathise with why one would support/include non-free code it is harming the community, making it difficult for Linus, and the free software projects shouldn't be putting up with it.
Not all free software projects accept non-free code and the mainline kernel shouldn't either. But what we have is a mainline kernel where non-free blobs are allowed in and a fork for other distributions without said blobs: libre-linux is used by most truly free distributions and then there is also debian's kernel which is also used by many.
We would be better off if the community stopped accepting non-free code altogether. We should make chipset manufacturers release the source under a compatible free software license for everything that will be distributed in any distribution or free software project (mainline kernel). We already do this to a large degree in many projects with great success. ALSA for sound, Debian for servers (desktops too), Trisquel (for desktops), amongst other. There is only one company with chipset(s) releasing proprietary drivers for GNU/Linux today. It's Creative.
I'm very much a fan of freedom and Trisquel is doing well particularly given the cooperation between that truly free project and ThinkPenguin.com (actually libre.thinkpenguin.com for a sanitised version). ThinkPenguin doesn't ship hardware dependent on non-free software so you can largely assume it's going to work without too much hassle across distributions. When users are told they need to replace certain components in order to get stuff working it's generally a really good and easy solution to the non-free software problem. It might cost a bit of money although it's nothing compared to what you would pay for Microsoft Windows or Mac and unless you are doing GNU/Linux to be cheap... and even then most likely you are still getting a better deal. From better chipsets to less expensive software. I say less expensive because if you are buying from companies like ThinkPenguin who are funding the development of GNU/Linux then you are indirectly paying for free software. And there is nothing wrong with paying for free software. Most people just don't do it if they can get away with it.
If you use GNU/Linux or want to use GNU/Linux and can't I highly recommend becoming a contributing member or sponsor of the free software community. There are a whole lot of projects which make this really easy:
http://trisquel.info/en/member
http://www.fsf.org/associate/
Some less free distributions focused on free software development:
http://www.linuxmint.com/sponsors.php (warning: not a free distribution, some non-free software is included)
www.spi-inc.org/donations/ (non-profit setup to handle donations for the Debian distribution, Debian is not a completely free project, although does exclude binary blobs in the base distribution- other repositories DO contain non-free software)
My nVidia drivers survive Windows kernel updates perfectly fine.
You want 3D graphics and decent sound don't use Linux.
You want a good server don't use Windows.
The linux fanboys can point fingers at Nvidia for all they care, but from the user point of view, Nvidia's Linux drivers generally work better than ATI's Linux stuff. And when they break because of kernel updates, the user is just going to blame Linux rather than switch to ATI and have problems all the time rather than just some kernel updates. Nor are they going to switch to Intel and be stuck with an ugly slideshow instead of smooth 3D graphics.
Other hand nvidia not too long ago was releasing graphics hardware with a higher failure rate than reasonable. I don't know about the current batch since I'm on ATI hardware now...
People like to claim ATI/AMD drivers are complete crap. I'll try to remember the progression correctly, but it's close enough to make a point. I've had ATI>ATI>NVIDIA>NVIDIA>ATI/NVIDIA>ATI>AMD, over the past 15 years.
Windows: I had one motherboard incompatibility with my first ATI card, aside from that no problems from any card at all. Nvidia and ATI both worked quite well over the years. I don't think that makes either of them perfect, but people like to blow things out of proportion and/or blame them for their own stupidity.
As for support in Linux, I'm not as familiar. I've usually ended up rage quitting with both Nvidia and ATI because I'm a gamer and couldn't play whatever title it was at the time. I know I had issues over the years when playing with Linux, with both manufacturers.
No, nVidia does it to sell more graphics cards. They make money off supporting GNU/Linux. If they do a shitty job as they have we shouldn't be supporting them. Torvald's is LETTING them do bad stuff to the community by being indifferent to the problems nVidia is creating. Torvald's has sadly let them take advantage of this community. The open source movement was intended to foster development. However it has never been in our best interests (open source) to let in non-free software to free software projects (mainline kernel, while it isn't a FSF project it is free except where we've allowed companies to
"contribute" binary blobs- this was a mistake).
If your need to play games is sooo great then go get a copy of Microsoft Windows and play them. Torvalds shouldn't have to put up with this crap. He's suffering for your benefit. And while RMS's stance may piss him off there is a reason for RMS's stance. I believe it is overlooked and the costs of going "open source" were too high.
For me it is a clear violation to ship a phone with the Linux kernel + non-GPL driver as the combined work is derived from the kernel and must be GPL.
It is ok that you download a non-GPL driver to your system and install it. But then you must uninstall it before you sell/give it to somebody else.
Absolutely, you're dead right.
My point is if the Jesus Christ of Linux, Torvalds, is attacking them for their support, then that risks harming their marketshare. It may actually lead to the Linux not being worth supporting, that's kind of the point.
He can voice his opinion but there's a massive difference between encouraging improved support for Linux, and basically trolling companies who do support Linux, one is productive, the other isn't. Torvald's is carrying out the unproductive one.
Yeah, and they solved that problem. An entire module rebuild facility for kernel upgrades was probably developed just for Nvidia.
Are you talking about DKMS? If that's the case, then you probably heard about UEFI secure boot? You know, that thing where everything, including the kernel AND the kernel drivers will have to be signed? I don't know if you got me, so I'll make it more clear: if you want to use UEFI secure boot, then you can't use DKMS, so we'll be fucked and the "upgrade destroyed my X setup" crap will happen again.
Tao Presentations is a 3D presentation software that runs on Linux as well as on Windows and MacOSX. We had mixed experience with nVidia support. Sometimes, it just works. In other cases, major functionality is missing (e.g. a laptop where the HDMI port is simply not detected). That being said, except when we hit a major bug, OpenGL performance is overall rather close to what we get on the same machine on Windows, which indicates that nVidia takes Linux somewhat seriously.
Actually, I don't think nVidia has any kind of anti-Linux strategy in place. It's rather that their portfolio is big and their overall strategy is sometimes confusing as a result. Consider stereoscopy for example, which matters a lot for Tao Presentations. Both nVidia and ATI have this puzzling idea that only "pro" customers can have stereoscopic support for OpenGL. This leads to situations where a machine with an nVidia chipset will connect to a 3D projector perfectly... as long as it runs MacOSX. The exact same hardware won't offer quad-buffer OpenGL support when running Windows or Linux. Why not?
So it's not about Linux, it's about lining up so many little pieces that sometimes, one of them is missing and the whole thing collapses. Linux is probably at the bottom of the list of things to fix, so that's where you see more problems.
-- Did you try Tao3D? http://tao3d.sourceforge.net
They should cooperate so that when their drivers crash and corrupt data on my encrypted home partition, they can get help to make sure it does not do that any longer.
Ubuntu Precise Pangolin, straight out of the box, with the standard proprietary Nvidia driver. GeForce GTX 2**M card. Crashes sporadically on heavy load, causing a hard kernel panic.
xkcd is not in the sudoers file. This incident will be reported.
I'm right there with you. Several years back I decided to quit buying laptops or video cards made by ATI because of the hell I would go through EVERY TIME in order to get the GPU working properly in linux. I'd dread upgrading my kernel because it was guaranteed that I'd be screwing around for a good while in order to get graphics working properly. I went exclusively with NVidia and ever since I have never had problems. I upgrade my kernel or distro and then simply install the nvidia driver and it just plain works. Every time.
It's been established for a while that nVidia is not going to open source their driver. It's not like this is a surprise. In the past, Linus has gone on record to say that binary blobs are better than no driver at all.
The turnaround on a kernel release would need to be a month or so for a company like nVidia to make the required changes to their driver and have it ready when a new kernel is launched. Hell, even the Evil Microsoft provides developers access to pre-RTM builds with enough time to make sure that their drivers work at launch.
What kind of OS changes their driver API so much that a driver compiled for one version doesn't "just work" with an incremental update? An immature one. Frequent kernel updates are basically patches. If your driver API is changing that much, you need to re-examine how you're doing your development.
We have every right irritated with Linus here. I've written software for Linux, Windows, and embedded RTOSes. Linux was by far the most painful. Boost was the only thing that made it tolerable. Berating a company that supports a user base which represents 1% or less of their market without any modern-day language tools at their disposal (STL, Boost, a decent IDE) is just counter-productive.
If Linus wanted to do something productive, he'd help to stabilize the various APIs in the kernel so that this stops being an issue.
Bottom line: this goes both ways. Either you allow binary blobs and give the providers of those blobs a chance to get their drivers working before you put out a new kernel, or you launch a kernel whenever you feel like it and live with the fallout.
Graphics cards are, nowadays, a bit more than 3D gaming. Nowadays there are a number of markets that companies such as NVidia may cater and are of fundamental importance, such as smartphones/tablets (remember Tegra?) and HPC (remember CUDA/OpenCL?).
What these markets have in common is that linux is the only reason they exist and are relevant. Windows is,at the very best, a "also ran" in mobile devices. In the HPC world linux is essentially the only game in town. In fact, there is currently only a single entry proprietary OS entry in the Top500 list.
Do you expect NVidia to abandon any of those markets in protest of Linus Torvalds pointing out that NVidia sucks at supporting linux? Think again.
Slashdot, fix your code or at least hire someone who is competent at it to do it for you.
Errr... if a driver can't survive kernel update... isn't that really the problem of the kernel development process?
I mean, imagine you'd have to update / compile your drivers with each windows service pack (yes, I know even windows break binary compatibility sometimes, but that's from a major release to another and it's communicated well before it happens.
So who's really to blame here???
Linus Torvalds is also not required to enjoy or approve of NVidia's policies, particularly when they generate bad publicity for Linus Torvalds' project
Except that's bullshit. Linux Torvald's project is generating bad publicity because it's bad at graphics drivers. ATI is more than accommodating and helpful--and the open source community fails miserably at creating stable drivers. So just about everybody demands Nvidia or ATI *fix* Torvald's failed project. If Nvidia opened up their specs more there's nothing to suggest that the drivers would be any better than the Open Source community's crappy job with ATI drivers. At least Nvidia offers a workable driver set even if it's a binary blob that gives Linux a workable video card option.
If it weren't for Nvidia's blob drivers getting anything done in Linux that requires a video card would be a hell of a lot harder and far less interesting. I sure as hell wouldn't rely on an operating system that performs as well as Linux on ATI's open source drivers. Having a working driver option for Linux generates far more good publicity for Linux than the alternative of a crappy but ideologically pure driver.
I got a feeling the same guy who's so die-hard he runs linux on the desktop, is so die hard he reads reviews of video cards and likes to build great computers, and just might like to have the "latest and greatest". A nerd with a computer is a biker with a motorcycle, and will probably spend much more of his disposable income on computer parts compared to the general public(as do bikers with cycles).
So while he's going to bitch and complain, he's also going to go out and buy an older nvidia card, hurting sales of new cards for nvidia, or mabey even try the secondary market on ebay. How many people REALLY will plunk down $500 for a video card?
the average joe will not.
Also, it'd almost really be better for nVidia to open source, and we won't have this. The only thing better for a company than fanbois, are fanbois who program, and submit bugs. Thats why the open source development model works so well, it puts fanbois to good use. The more fanbois, the more people looking for bugs.
nVidia is probably the best option for performance discrete graphics on Linux, I already tried ATI and was deeply dissapointed, while nVidia drivers for my 4-5 year old laptop handle perfectly connecting to my external 1080p monitor my new (few months old) laptop with ATI graphics (A6 APU) totally fails to even drive the monitor over 1440x1024, even tough the ATI card has way more memory available to it.
nVidia propietary drivers work and are well tested, while ATI drivers sort of work most of the time but they are filled with glitches here and there.
Of course, not everything is pink and ponies on the nVidia side, the downside with nVidia is the little to no support to modern X stack and linux graphics standards, xrandr? fuck that use the nVidia panel or no dual monitor for you, KMS? hell no, its all or nothing, ATI is just only a tad better but still useless to me.
On the OSS side of the cards drivers the ATI drivers sort of work, KMS and basic stuff is solid but for modern cards you get mostly no acceleration for anything, which is just a waste of all that silicon, Noveau drivers are the same, basic stuff working but thats all, but Noveau developers have it harder as they have to reverse engenieer all, with no docs at all or help from manufacturer and they have been around less time than ATI devs.
C-x C-c
The customers who are using literally tons of their very expensive Tesla hardware in supercomputers might be one such incentive.
Another angle photo of Linus "pointing" ...
http://i.imgur.com/zEC7j.png
NVIDIA's graphics drivers for Linux are full-featured and rock solid.
Fuck, no, they're not. They categorically refuse to support anything with their Optimus stuff, which is just about any recent laptop you'll come over.
It's not like they just don't support the GPU switching. I could live with that.
But their driver won't even load if it detects an Optimus configuration. So fuck them.
While nVidia cards will work with binary-only drivers, if you're using fully supported X.org drivers, ATI is your only alternative. Seriously, the X.org drivers for nVidia don't even have bearable 2D-performance - you simply *can't* run X11 on them without the binary drivers (which may or may not be compatible with the other things you have...because you can't recompile them).
So they can benefit from that sweet 1% linux desktop share!!!
I have been waiting for some sort of response from Linux for a while now, with regards to Nvidia.
Linux is right, Nividia is nothing but a pile of poo, and it has caused my business a ton of problems with virtualization for example in the Linux Desktop space.
Many of my customers use qemu-kvm, with desktop linux terminals to run UNIX, Windows and other pieces of software for AT&T.
Some of these are autocad based, and the AMD open source drivers works really well. However, a few of the workstations I have had to deal with use Nvidia GPU's and the drivers wreak havoc on the kernsl stability and patch cycles for virtualization.
I now don't even bother and require customers to buy AMD gear otherwise I don't support it.
So I have to agree with LINUX, for anyone out there actually trying to create a LINUX based desktop future, the Nvidia hardware really crucifies the customer experience in dealing with the idea of a LINUX desktop.
-Hack
Got Geometrodynamics? Awe, too hard to figure out? Too bad.
No. The thing that cause you to be "fucked" is not being able to boot a new kernel.
The idea that you won't be able to compile your own kernel modules because of Microsoft's locked booter is a novel and interesting concept.
A Pirate and a Puritan look the same on a balance sheet.
You should tell the guys in Siberia with the cool 3D RTS for Linux.
Obviously they didn't get the memo.
A Pirate and a Puritan look the same on a balance sheet.
Telling someone "FUCK YOU", even as a joke, is definitely the wrong way to ask for increased support.
I managed that in 1999.
I suggest attributing your problem to "user error".
> because I don't know what kind of video card it is
There's a very easy way to get around that. Google is your friend.
A Pirate and a Puritan look the same on a balance sheet.
"I had no point, I was just trying to slam Linus for daring to criticize my favorite corporation.
FTFY.
I hope you're not actually actively buying graphics cards at all, then, since they're only useful at all for 3D and gaming. You can play DVDs, Youtube, and Flash just fine with integrated graphics. Intel has decent free/open-source drivers built into Linux, for example.
2d is easy. 3d is hard. Your anecdote throws out all the stuff that doesn't work properly with ATI, and you claim that what is left is representative of the whole and just as good as Nvidia.
You appear to have no idea what you're talking about, either. Likewise, good to know.
Nonetheless, they aren't sisu-fueled supermen and there's a reason that during the Cold War they made serious compromises with Moscow.
Ah finnlandisierung meine freunde.
Seriously, Cold War time wasn't all that bad time for us. It was hard, but not completely bad. I know it well as I'm old enough to remember it well since sixties. We did have homie-russians some time that tried hard to influece and mostly they succeeded to get people self-sensor telling truth about many matters what people really thought about eastern neighbour that is what finlandization wich is german invented term of the time describes. We lost part of Karelia true, we did lost Petsamo too up north, and we got Stalins advisors (valvontakomissio ie. supervisory comission) to look after that the war criminal prosecution went on direction they wished. Hanko peninsula, the south tip, was a short time also rented for Russian military base, but it did not last long before they left. Supervisory comission left quite soon after the war criminal prosecution was over. Surely KGB had is's men Russian embassy but so did many other western and eastern block embassies too.
We also paid heavy compensation to USSR, but luckily that also rapidly helped build our factories to produce goods that later took us from agricultural society to a modern industrialized nation. Once compensations were paid it was nothing much but just playing with sordino after that. We have saying that if you are made to sleep next to bear, better not make too much noise to wake-up. it It was open secret how screwed USSR was, and besides politicians in officials and trade negotiations jokes were kept aside otherwise russians were laughingstock of everyone. And they were not only in Finland, ask anyone from Baltic countries or anyone from Poland, it was same there. Ofcourse there were communist minded minority, being communist here is not criminal offense, but they never got large support in elections and besides only a very minor fraction of them actually wished we would merge with USSR. Finally they got the clue after USSR dissolved and disappeared from political map.
We did quite a bit bilateral trade with USSR and that provided work to many on our side and russians got some industrial products we produced, that was the good part of it to both sides. If you still remember those two MIR mini-subs Russians used couple of years ago to visti Nort pole, those were built on that time.
After Cold War it suddenly stopped for some time and that was greatly contributing the recession early nineties, then bit trade continued with Russian Federation now with currency. There have been and will be challenges next to Russia as any neighboring country can tell. Finland has managed quite well not to kick russian bear ass too much and often to wake it up wrong foot and hopefully it remains that as well. Not all USSR/Russian neighbors did not manage as well, unfortunately for them.
Cheers,
ac
Sorry, ATI is worse. At least Nvidia supports their hardware for more than four-ish years. His deciding not to purchase ATI products was the right call. I once thought that I'd give ATI another chance. Now I have a motherboard that with an Radeon HD 3200 integrated video card that now has no official drivers for the video card. So it is now useless for running XBMC at a reasonable performance. Another motherboard of mine with an integrated Nvidia video card which is the same age still gets driver updates from Nvidia.
http://www.phoronix.com/scan.php?page=news_item&px=MTExMTA
People need to make conscious decisions on what they purchase if the situation is going to improve. Many people are doing this already. Many more need to start doing it though..
If you want stable, hardware-acceleratd OpenGL 3 or 4 support, you buy nVidia cards. Since I work a lot with OpenGL 4 (and sometimes 3), I have no options. I had an ATI/AMD card once, and will never buy again. Anything beyond basic 2D is crappy with these (due to the awful drivers). Same goes for the Intel chipsets.
So, some of us do not have the luxury of picking our hardware based on ideologies. I am happy that I can use Linux at all for my projects..
This sig does not contain any SCO code.
> No one is suggesting nvidia couldn't do better, but when he attacks a company over this sort of thing it just makes him sound like an ungrateful prick.
If it was my kernel, I sure wouldn't be grateful to have to handle thousands of bug reports because of that binary blob and have to do 1970 style debugging and reverse engineering every time there is a problem because they thought running unknown code in kernel mode is so great.
If it was my kernel, I'd blacklist nvidia on the first bug and not enable them again until they did what everyone else managed to do: integrate into the kernel on a source code level. Linus was nicer: he just sets a flag so he can see from afar who it was and loads it anyway.
It's not like today's graphics adapters have much secret sauce, they are just parallel floating point units.
The main problem I have with the NVIDIA drivers (aside from the obvious closed source) is that it doesn't play nice with the X RandR extension. It's 2012, RandR has been around for years, all the utilities for managing multiple screens make use of it, but not with NVIDIA drivers because they can't be bothered to even set up an interface between RandR and their own "TwinView".
Then they are welcome to stop using the low-end crowd as testers for their high-end folks. See how well that goes over with them.
It's not just desktops which use GPUs these days...
How about the 90%+ supercomputer market share, where GPUs are starting to be used...
What about the server market, where GPUs are also being used for computing where linux is fairly widespread.
Also mobile devices, where linux in the form of android is widely used.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Actually it costs them more to develop their own binary drivers in house, than it would to open source them and allow third parties (including the kernel devs) to take up some of the slack...
Not supporting linux would lose them business, almost certainly more than it costs them to support the drivers... Linux may not be huge on gaming desktops, but GPUs are also targeted at high performance computing, and Linux is very big there.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
his work on Linux operating system
Linux is not an entire operating system, it's an operating system kernel
GNU+Linux is an operating system.
Have you tried the FOSS drivers? Because you really shouldn't complain when AMD did what the community asked them to do and handed over the specs, or did everyone forget how many times we've heard "Just open up the specs and we'll support the hardware" on this very forum and others? AMD took that one step farther by actually hiring developers to assist the FOSS driver devs in getting up to speed, and from what I've been reading they've been coming along nicely, although focus has naturally been a little heavy on the APUs since so many of them are out there.
This DOES highlight what i consider to be a major failing in Linux for quite some time, the fact that its damned near impossible to JUST get security updates, as package A needs kernel B and depends on packages E-G so to keep the thing updated you end up with an "all or nothing" so you can't just update say the XBMC software without changing graphics drivers and a bunch of other shit. People can scream bloody murder all they want but THIS IS WHY a hardware ABI is a GOOD thing, because if I want to keep my 3 year old graphics drivers while being fully patched and running the latest of everything else on Windows? not a problem, I just don't have to update the graphics driver. this is also why AMD's support phase isn't a problem on Windows, as one can simply stick with the last driver and be fine for the life of the system and after 4 years they've squeezed all the power they are gonna out of a chip. Again like it or not the ONLY REASON that that Nvidia or AMD have to keep releasing new drivers for old hardware in Linux is because you simply can't use the old drivers with new kernels or the whole thing falls down.
So I don't see how the community has any right to complain about AMD, you got exactly what you asked for, all the specs opened and handed to you on a silver platter. AMD simply has a hell of a lot more on its plate than just graphics so continuing to support 4+ year old chips on an OS with maybe 5% market tops is simply a waste of resources. if you want to complain pitch a fit at Torvalds for making driver support such a damned mess, even one of the big Red hat developers says the current way of doing things simply isn't sustainable, that a single group can't control 20,000 packages and drivers and keep it working, and recommends an ABI and a much more stripped down design that allows you to concentrate on the core while letting those that sell the hardware provide drivers. I wonder how much money Nvidia has blown keeping a team of devs around to do nothing but constantly update the Linux drivers when Torvalds constantly breaks the damned drivers with kernel fiddling? bet it isn't cheap, not cheap at all.
If handing you the full specs like you asked for STILL isn't enough? Maybe its time to look in the mirror and consider that maybe, just maybe, you're doing things the wrong way. There should be no damned reason why you can't take the last release that AMD made for that HD3200 and have it run perfectly on the latest distro and the fact that you stand here and admit that it doesn't work just shows what is wrong with linux in a nutshell. After all how do you expect the smaller hardware guys to support you if the big guys have to pay entire teams to constantly fix the damned things just to make the drivers work?
ACs don't waste your time replying, your posts are never seen by me.
Doesn't Compiz use 3D acceleration?
"Sissy"? Really? What is this, the 60s?
Supercomputers are going to use software written from scratch to access the GPUs, I don't think an over-the-counter driver is going to do much. I don't know much about GPUs used in servers but I can't really see you accessing the GPUs through the general desktop drivers Torvalds seems to be complaining about. And finally, android doesn't really follow the development model of mainstream linux, nvidia is going to work closely with the large manufacturers who actually create the android devices.
I suppose there could still be a commercial manufacturer for a completely open hardware computer.
That's a Linux problem, not an NVIDIA problem. The notion that drivers should get updated every time the kernel is updated is completely stupid. If you have that particular problem and it bothers you, blame Linus, not NVIDIA.
Torvald's is carrying out the unproductive one.
Your so-called productive approach has been entirely unproductive for more than a decade. It is way past time to say "fuck you".
Finally! A year of moderation! Ready for 2019?
ATI video drivers suck for both windows and Linux although windows drivers are very slightly better for windows.
Windows drivers are slightly better for Windows! Profound!
Man blir trött av att gå och göra ingenting.
I bought two 8800GTs the day they came out, sometime late 2007. Within a week, NVIDIA had a beta driver to support it, and the only issue with that driver was that it made the cooling fans spin at full throttle, everything else worked perfectly. Hard to imagine that nearly 5 years later you are having problems with those cards on stable drivers.
If I have a project that has problems when other people don't do something that I require from them, I *am* the cause of those problems. I don't have the right to say "fuck you" to anyone who refuses to do what is necessary for my project to succeed.
Linus doesn't have to be polite about such things. There are alternatives to nvidia, after all. Nvidia can go with linux or get lost, it won't matter that much. If you buy a computer for linux, you buy well-supported hw - of course. Linux isn't all about installing on a random old computer any more. You can buy a decent computer for it - which is why I went with ATI. I knew in advance that the open source drivers would work well. No binary blobs (other than the GPU firmware file).
Does that include driving 4 displays with working 3D accel and xv?
Because I sure as fuck couldn't get it to work properly with a pair of 560s.
Replaced with a HD6870 and 2 active DP-DVI dongles... works.
Linus has been dealing with nVidia for the last ten years and they haven't changed their tune. If I were him, I would not have been as patient or quiet.
Well, there's spam egg sausage and spam, that's not got much spam in it.
I couldn't find drivers for the video card, because I don't know what kind of video card it is.
Funny, lspci will you tell what card you have in your computer, like e.g. VGA compatible controller: Advanced Micro Devices [AMD] nee ATI Barts XT [ATI Radeon HD 6800 Series], Now, on MS Windows if the driver is not installed, the device manager will gloriously tell you Generic VGA, and you have to google for the device ID to find out which type of card you have in there.
I had never seen Linux detect and use any kind of video drivers in my life.
Any recent linux distribution will detect the card if it is in its data base, it only happens once in a while that the card is too new to be already listed.
With Windows, you buy a computer, install Windows and it works.
But only after you went disk-jockey installing all the drivers from 3rd party CDs. We even had it at work that Windows (7 64 bit) didn't support the NIC and we had to put in another one to get a net connection.
With Windows it's usually: You buy a computer with Windows pre-installed and it works (and that was one of the problems Linus mentioned - that it's difficult to find vendors to preinstall Linux.)
nVidia could say "fuck you" to Torvalds for refusing to standardize the kernel device interface so that a blob binary will work for more than one version. That isn't too much to ask. Windows does it (as does OS-X). A binary driver will, at most, need to be updated for service packs and usually not even then, they generally work for the entirety of a version of Windows. Only on Linux does ever minor kernel update require a new driver.
I understand he wants OSS drivers, but I can understand why nVidia doesn't. While they are being stubborn and refusing to budge, so is he. Linux could be set up so that closed source drivers were easier, so that only major updates required a new one, but it is not.
It's hard to write software which directly access the GPUs without hardware specs, which nvidia don't like to provide.
Every single supercomputer deployment writing its own code for everything would be stupid, which is why so many run linux and then write a small amount of custom code that pertains directly to their individual needs. No point reinventing the wheel.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
There was all this chatter about how if the GPU companies would just open up, legions of extremely smart programmers would make grad-A OSS drivers, better than the Windows counterparts!
So AMD does and... Nothing. We have a broke ass, "sorta works" OSS driver. Apparently the legions of programmers are either busy playing WoW or maybe, just maybe, writing a graphics driver for a modern card is way harder than people give it credit.
I think part of the problem is you get people who've written something like a NIC driver and say "Oh this driver writing isn't bad." The problem is most hardware is peanuts compared to a GPU. They are just amazingly complex. You can see it in driver sizes. A NIC or RAID driver will generally be 100ish kbytes, sometimes less. On my system, the primary WDDM nVidia driver file is 20MB, and it won't even work completely with that alone. It's driver to support OpenGL is another 25MB, another 10MB for D3D10/11 components, 7MB for CUDA, and so on, never mind support stuff like the control panel.
Basically it is a really complex problem, and of course each new version of the graphics hardware brings in a new setup to deal with. So it isn't so easy to bash out a high quality driver for graphics cards.
Thus the OSS AMD drivers really aren't any good, despite AMD playing nice. The community has not managed to produce some amazing driver that is fast, stable, feature complete, that makes Windows people say "Man I wish I had that." Had that happened, I'd be more inclined to say "Get with it nVidia." However as it stands, nVidia is able to produce a Linux driver that is in every way as good as their Windows driver, and that is damn good. Given that, I'd say they are doing it right.
Nvidia drivers work just as well as any other graphic driver, LIKE SHIT!
I have a 9 year old custom built box at home and have been using Ubuntu on it since Ubuntu came out.
When I tried to upgrade to 12.04 I got all sorts of kernel error messages during the install and related problems later.
It is my understanding ( less than perfect ) that support for hardware fell through the cracks between changes NVIDIA made and changes Canonical made.
I've been window shopping for a new PC system and several people told me that onboard video has improved substantially in the last 9 years, that since I am not a gamer, that I do not need to get a separate video card anymore.
Goodbye NVIDA.
How many people would REALLY be able to look in to the box, if the code were available? To me the whole OS is a "black box". I'm not a programmer. Even if you are, kernel programming is another level of it. So for many people the whole thing is a black box. We can look at inputs and outputs, but we lack the ability to play with it. Works ok that way.
In fact, that seems to be how Linux is generally treated support wise too. Our Linux lead at work is a very smart guy, and is a programmer, yet when I watch him troubleshoot Linux problems he doesn't go break out the debugger and start tracing through kernel code. Rather he takes the same "black box" approach I do to troubleshooting Windows problems: See how the output differs from expected, formulate a logical reason as to why, change things, see if that fixes it, repeat the iteration until the problem is fixed.
I think the obsession with source really -is- more of a religious thing to 99.9% of the people who whine about it rather than anything practical. Many don't have the ability to go and debug it and even if they did, they don't have the time and even if they did it usually isn't the best way to solve the problem. It makes them feel good to have, but they don't actually make any real use of it.
The majority of the people at Nvidia don't feel that Linux is worth supporting. Luckily, there are a few vocal supporters at Nvidia and they're allowed to put in the work to make it happen.
That's the exact opposite of my experience.
I have ATI on my laptop and NVIDIA on my desktop. I freely upgrade my desktop kernel and use the NVIDIA installer to update the driver. However, every major X.Org release breaks the ATI drivers for a number of months until ATI gets around to it. The open source ATI drivers are pretty good at keeping up but they suck the daylights out of the battery on my laptop which suggests that the "specs" that they gave the open source community are incomplete.
My experience:
NVIDIA - Closed source but well maintained drivers.
ATI - Mostly open specs but poorly maintained drivers.
Some of what I say is fact, some is conjecture, the rest I'm just blowing out my ass...you guess.
Been saying that for years. ATI/AMD did what the community wanted and guess what the community found out? Writing good graphics drivers is really freaking hard.
I also agree that an ABI is not badly needed. Even if you are going to produce your own FOSS drivers and an ABI is a big help.
The current FOSS solution makes a good enduser experience really hit or miss for a desktop user.
1 Write a driver for your hardware as FOSS.
2 submit it to the Kernel.
3 wait for it to get put into the Kernel.
4 Hope that the big distros adopt that Kernel soon.
Then end user can not tell if the this or that device is supported half the time without knowing what freaking chipset it uses.
If you have a binary driver interface then you can stick a driver in the box and say "Supports Linux 4.xx" and it should be just fine
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
How would it lose them business? That small fraction of 1% of people who buy their video cards that use Linux? I mean, we all know Linux isn't used in the home for anything real. At least anything real requiring a video card requiring accelerated graphics....
Nvidia could love EVERY Linux user that has a system with an Nvidia video card in it, and it wouldn't warrant even a mention as to why their numbers looked different.
I abandoned Nvidia and went to ATI/AMD since I only wanted to rotate one of my screens. That only works with xinerama with Nvidia, and with xinerama there is no Compositing anymore. Nvidia owns "TwinView" doesn't support that. That wasn't the only issue, video playback was never running smooth, even if the card was modern enough (9600 GT). With AMD/ATI and fglrx drivers all runs fine for me.... but Intel graphics (in my notebook.. ) are unbeaten in Linux integration, no extra tools needed, just use Randr!
In the past, many closed source programs have turned out to have patent violations in them.
If a company has their own patents, then those patents are already public so closed source doesn't protect them. To claim it does is totally bogus.
But if a company is violating other people's patents, then of course, they'll want to keep closed source.
NVidia is worse than useless and I would not buy a machine which contains NVidia hardware. Having said that, I stay away from ATI for much the same reason. Usually this means I have to get by with just 2D but it's good for most things. I've not yet owned a computer able to decode some 720p h.264 video in realtime or even come close to running gnome shell but thanks to ogv (much less CPU stress) and the fact that Debian Stable doesn't like to include software written in the last 5 years I've not yet run into a serious problem.
If you want to use NVidia I recommend Microsoft Windows as much as I can recommend something I choose to use myself. If you must use Linux and need some basic graphics then consider Intel.
But with the blob driver you get non-native resolution on the console, slow switching between TTYs, and very rickety fast user switching.
Having just spent a whole day trying to get a GT525M card working with that Optimus shit, I can only agree. Nvidia sucks on Linux.
none
Well, they certainly do subscribe to MS goals and visions and release the driver in a way that's appropriate for Windows platform.
Having said that, nVidia is by far the best (the only actually) solution for Linux. We run bloody flash animations and videos 24/7 on cheap atoms - if it wasn't for nVidia ION we couldn't be anywhere near to being competitive. Although it would be nice to have a built-in support, it's not exactly rocket science to rebuild the modules for different kernel. We do it all the time - nVidia driver install provides the infrustructure (precompiled kernel interfaces), we build it once, then distribute via yum/rpm to hundreds of client sites via nightly yum update, reboot and everything else taken care of automatically. Never failed us.
You just made my day.
Have you heard about SoylentNews?
It's the other way around. The drivers are dependent on the kernel version. I think nVidia did this on purpose for control.
Well, there's spam egg sausage and spam, that's not got much spam in it.
.... you need to understand that attacking or insulting NVIDIA will not help. The reason that they don't release more source code or internal details about their chipsets can be summed up in one word: patents. The more details they release, the more they risk being bushwhacked by either patent trolls or well-funded competitors.
To bring about change, all you can do is write your Congressman, minister of Parliament, or other legislator and explain that patents are a historical artifact that, while always of debatable value to society, demonstrably do more harm than good in a modern, fast-paced technological environment.
It's not like linux users are gamers, nor are there a large amount of games available for linux (that would make use of a 3d card)
Seems like Linus' real complaint is that Linux has a tiny fraction of the 3D market, leaving no real justification for any hardware manufacturer to spend $$$ developing drivers for it. And releasing the HW specs is silly -- as others have pointed out that exposes you to patent trolls, and when AMD did it we didn't end up with a magically great FOSS driver.
Desktop Linux is irrelevant when you look at the numbers. If you subtract all the android users, desktop linux is about 0.3% of the market. So, anyone doing any real desktop work is using Mac or Windows. Linux is great for servers, but not for desktops. So it makes perfect sense for NVIDIA, (or any graphics company) not to support Linux. I use Linux on the server and I abandoned it for desktop use years ago. Linux is brain damaged: gnome2 vs gnome3 vs kde4 vs ubuntu unity. Linux desktop is an exercise in futility, it has failed the Linux users and is a mess of warring factions. I have dual boot Win7 + Mac OSX and laugh at all these problems Linux keeps having on the desktop arena. What a joke!
Better than AMD for 3d accel.
I live in Chapel Hill, home to UNC, where NVDA was born. I have some friends who are deep into NVDA's tech development. The short answer: "NVDA loves Linux and open source, but fucking God damned stupid software patents make it criminally stupid for NVDA to let the open source community in on how they actually write their drivers."
Celebrate failure, and then learn from it - Nolan Bushnell
The idea that you won't be able to compile your own kernel modules because of Microsoft's locked booter is a novel and interesting concept.
But the bootloader isn't locked, you just turn off secureboot. It's only locked on ARM devices, like most Android tablets and all iPads are so the locked bootloader is hardly a Microsoft invention.
Naturally if you want to use UEFI SecureBoot you would have to sign the kernel, but that's the nature of secureboot.
Download their drivers, chunk back to text mode, run, restart graphics, done.
What's the problem? The only problems I've ever had involve installing a new kernel and forgetting to rebuild the graphics blob.
Every time I hear of some couple in the process of splitting up I walk up to them and ask them if they haven't tried to work it out by productive methods.
Actually it costs them more to develop their own binary drivers in house, than it would to open source them and allow third parties (including the kernel devs) to take up some of the slack...
How exactly? And please don't just suggest they get free labor with no overhead by just opening up to contributions.
There's still no KMS, though. And it took them how long to add XRandR 1.2 support? The reverse-engineered drivers had it pretty much from day 1.
I had a sig once. It was lost in the great storm of '09.
it craps out on you and dumps you back into a VGA console (if you're lucky)
If you're unlucky you get a kernel panic or no display at all or that VGA output is on a disconnected screen or many other things
ATI drivers are even more flaky
I've given up on AMD's closed Linux drivers. The OSS drivers may be slower and less power-efficient, but manually setting the GPU to minimum power mode brings the battery life to ~3/4 of what I get in Windows, even with desktop effects enabled.
It just means that all three are horrible companies, and that the GPU market really needs a shake-up.
At least AMD releases the specs for their GPUs so the OSS community can write open drivers that aren't horrible. Well, they do have (relatively) horrible performance, but at least they're stable
I had a sig once. It was lost in the great storm of '09.
I can't be the only one who still has a PC with a chipset that requires the "forcedeth" reverse-engineered drivers for ethernet, right?
I had a sig once. It was lost in the great storm of '09.
Your so-called productive approach has been entirely unproductive for more than a decade.
And what approach are you saying that is?
Some Ramblings
I have an NVidia (something or other) in my company assigned Dell Laptop.
Works a treat with the open driver. No 3D, but so be it. I don't game anyway.
The AMD/ATI situation? Same thing. Just takes time and effort to get the driver to the point where it can be even REMOTELY considered ready for the kernel.
The "driver du jour" from the hardware vendors? Can't be trusted.
Sure, if you are building your Supreme Gaming Machine, go for it. For any real work? Not so much.
Now, it gets weird. Because I am about to backpedal on that statement.
In some very limited circumstances, the vendor drivers may be deployed. Specifically, for GPU calculations. I still wouldn't trust these drivers in the role of a DISPLAY driver yet.
(I consider the GPU calculation testing to be more comprehensive and useful, although I find the use of OpenCL to be.. abhorrent).
But the set of features implemented by these drivers for 3D rendering (OpenGL) tends to be oriented towards gaming, and not the kind of visualizations needed for "real work". For example, 3D depth cued lines. A feature handily supported by SGI and SUN in the past, but missing from NVidia, ATI and Matrox the last time I looked -- 10 years ago, but I suspect still missing. Not that the feature was available in Mesa, either, but Mesa is the LOWEST level of support expected.
I would be happy if I were wrong, but, as far as I can see, Intel graphics is just as good (or bad) for my 3D visualization needs.
Now, I do have to give a tip of the hat to NVidia. They (at least) tried to support OpenGL with their implementation. But, I really don't understand how NVidia managed to create an OpenGL implementation that was arguably inferior to the SGI and SUN implementations. Possibly (and I speculate) that their attention was split by DirectDraw, and the perceived need to micro-optimize. The second reason was the need to replace a good deal of the driver stack, which NVidia tried to do without the cooperation of the kernel developers.
Which brings us to the present day, and a question: "What to do now"? Is it too late to have NVidia assist in laying out the driver stack? Most likely. The only beneficiary of the current situation is Intel. Intel has participated in laying out the ground work for display on Linux. Intel will reap the rewards of this. Both NVidia and AMD will be relegated to providing GPU processing, but will be squeezed from the bottom. After all, Intel will control the GPU sharing protocols. OpenCL will probably continue to entrench, and NVidia is trying to keep their compiler presence (they own the space right now). Intel is likely to release more general compilers and infrastructure to squeeze them.
AMD? I am afraid for them. They deserve better than to become a footnote in this saga.
Just another "Cubible(sic) Joe" 2 17 3061
I've been using Nvidia drivers on Slackware with various kernels since 2000. Never had a problem.
except NVidia considers pretty much the whole driver as 'firmware'--and it is OS-specific.
I usually claim that NVidia's abreviation NV stands for No Video. Yes, I'm that fond of their Linux support. Yes, I'm exagerating but unfortunately not much. My experience is either the mess with the binary drivers ir slow unreliable graphics with the open source driver. Unfortunately it's almost impossible to get a high-end laptop with an Intel CPU and AMD/ATI graphics these days. Fortunately Intel's integrated graphics is good enough for my needs so for my latest laptop which features Intel integrated graphics along NV chipset I didn't even bother with the NV chipsets at all. Yes, this time it truely is No Video for the NV chipset :-)
Meanwhile, I've been using ATI for more than a decade with zero problems past their initial teething stage. However, I don't care about 3D, gaming, Wine, or MS Windows. I have no trouble playing DVDs, youtube, or flash
I hope you're not actually actively buying graphics cards at all, then, since they're only useful at all for 3D and gaming. You can play DVDs, Youtube, and Flash just fine with integrated graphics. Intel has decent free/open-source drivers built into Linux, for example.
I am actively avoiding nvidia graphics based laptops (and Intel CPUs; but that's another story). I've had no trouble with Intel's integrated graphics, ever. I believe my machine no. 3, a Compaq Evo (PIV-2.4 GHz) running OpenBSD, has one. Runs X fine (but it needs a WiFi interface, which I'm still deciding on).
"Tongue tied and twisted, just an Earth bound misfit
What I know is when nvidia came out, I was seeing thousands of posts from people desperately seeking answers on how to get them to work,
The first time I ever used linux semi-seriously was in a computer class in like 2004. During downtime, since the comps had nvidia cards, I wanted to play Unreal Tournament, so i decided to install the drivers. The steps: .\driverinstaller
1)Go to nVidia.com, and download the driver installer.
2) run
3) play unreal tournament.
Ever since then I have periodically gone back to Linux, each time with an nVidia card, and each time the process was the exact same. Download the installer (helpfully linked here!), run the installer, reboot.
Ubuntu (and I imagine other distros) has for the longest time also had the nvidia drivers in its repos, so you can install it from there even easier ("apt-get install nvidia-drivers-binary" or whatever) and not have to worry about kernel upgrades.
Its not that hard, and it basically has never been that hard (at least for the last decade).
What I know is when nvidia came out, I was seeing thousands of posts from people desperately seeking answers on how to get them to work, and thousands more on how to make their X Window survive upgrades
Synaptic is basically a front-end for apt, and I imagine synaptic's other-distro brothers are similar in that regard. If an upgrade like that is breaking things, the package itself might be broken. Regardless, IIRC the driver is a kernel module and shouldnt be broken by just an X upgrade, though kernel upgrades do have the potential to cause issues if you DONT use a package manager.
Your anecdote throws out all the stuff that doesn't work properly with ATI, and you claim that what is left is representative of the whole and just as good as Nvidia.
Not just as good; better. I've supported people blessed with nvidia who felt the same way as me about 3D, et al. They just wanted working X Window. Nvidia makes that needlessly complex and problem prone.
If you want to do 3D and gaming, use Windows or buy a game console. Don't try to foist nvidia on people who just want a working video setup. "Don't try to teach a pig to sing. They're not good at it, and it annoys the pig."
"Tongue tied and twisted, just an Earth bound misfit
Which is why I can't stand those that treat FOSS and Linux as a religion, the ones i call FOSSies, because even when you point out what should be as plain as the nose on your face, that there is no damned reason new drivers need to be written for old hardware except that the current driver model simply doesn't work they scream like you must have taken a crap on Baby Jesus.
I mean for the love of Christ they call me an "MS Ninja" for DARING to point out you can't sell Linux to consumers in its current state when what do we see right here on this very forum? That unless the hardware OEMs constantly release new drivers, even for old hardware that isn't even fricking being sold anymore, that the software WILL NOT WORK with current kernels. i'm sorry, that is just royally fucked up. hell on Windows you can use 2K drivers on XP and Vista drivers on 7, and i can take the driver that came with my 2 year old card, slap it in a fully patched Windows with fully updated software and it just works and my customers never have to worry about it.
Believe me friend I WANT Linux to get better, I really do. I want to offer nice looking Linux machines right next to the windows units and even better to get rid of XP on the older units for a nice new shiny Linux OS but I just can't, and that is because if you keep the damned thing updated the drivers fall down and go BOOM!
Please Linux community, please stop taking this half baked horseshit from Torvalds and Co, please stop accepting forum hunts and pages of CLI mess just to get the damned drivers functional after an update as par for the course, you should and do deserve MORE. After all a free shit sandwich is still a shit sandwich and right now the driver situation is a fucking disaster. How can you expect to get support from those thousands of smaller companies making hardware if fricking companies like Nvidia constantly have to release NEW drivers for OLD shit because the old drivers will no longer function thanks to all the BS going on?
so please, demand a stable ABI, with say a 5 year cycle between major changes to said ABI and a way for the previous ABI to be used beside or replace the new ABI if the person has older hardware that requires it. there should be no damned reason why my customer should have to update his GPU driver just to run the latest XBMC or any other software for that matter, and frankly if you fix this glaring problem? then there is a good chance you could grab some real share, as you have nice software, GUIs, pretty much all of the top layer stuff is quite nice but its the low level stuff like drivers that are really putting a couple of slugs into Linux before it can even get out of the gate.
ACs don't waste your time replying, your posts are never seen by me.
Care to provide a link to those specs?
I can't find any docs on Radeons that date past 2010, which means all the cards out in the last two years are back to being secret.
We have every right irritated with Linus here. I've written software for Linux, Windows, and embedded RTOSes. Linux was by far the most painful. Boost was the only thing that made it tolerable. Berating a company that supports a user base which represents 1% or less of their market without any modern-day language tools at their disposal (STL, Boost, a decent IDE) is just counter-productive.
Seriously, if you write a device driver in Boost, you should be shot in front of your kids. And I know, I used to work for a company that wrote most of its software in Boost, luckily only in user space. They had all the usual problems with Boost, all the static and dynamic cast bullshit, 8 different kinds of smart pointers, all of them thread unsafe and of course, error messages the size of a small house. They had a bug for 2 years because they couldn't follow the code because of Boost and all Design Patterns they had used. There is a reason why Boost is not STL.
So no Boost or STL in the kernel is a good thing, only to keep coders like you out. Just read what Linus has to say about it:
http://article.gmane.org/gmane.comp.version-control.git/57918
(full dislosure: I now work for a company doing almost all coding in Erlang, and we rely on one peice of code written in Boost, and we hate it.
It is almost impossible to compile, lacks a good api, and if you change any configuration while it is running it drops all its connections. We really would like to replace it and probably will at some time.)
It's because the lead-in to this article asked "What has been your experience on nVidia's drivers with Linux?" and for the vast majority of people here, that experience is as an end-user, not a kernel developer.
You're right of course, that they're two completely separate issues, but nVidia's refusal to "play nice" with integration into the kernel is purely a theoretical concern for a lot of us, while the performance boost is very easy to see.
This "article" was about Linux an "open" OS and it's main man, Linus, the main reason Linux exists .. right? Why do some of you defend the practices of a company that continues the "closed" philosophy they have always had? I am sorry but I don't think this article was about "oh my god I can watch 3D chesticles on Nvidia hardware running Linux better than I can do it on AMD hardware , right of the box". .. this is about the issue of openness ... which is the basis of Linux. .. and the fact that their drivers seem to work better out of the box for the casual "consumer" in the generic scenarios doesn't mean anything. .. once I got the environment set up for what I needed and stabilized . And that is only because I was able to play around with AMD's drivers .. not to mention having direct support from AMD engineers. I would have never been able to accomplish that if I wanted to use Nvidia's chips. But at the same time my needs were more specific. However for generic purposes if I really wnated to use Nvidia hardware I would use it in a windows environment where they actually put a LOT more effort and you actually get the performance you paid for whn you bought that ridiculously expensive card. .. if you are not exactly a power user and you are content with "out of the box" stuff .. Nvidia might be a better ( or should I say easier) choice .. but for those who want/need to pull all they can out of their system and know how to do it .. AMD is actually a much more viable solution. .. or fanboysm ( if those are even words .. but you get the point) to me.
No
Nvidia never believed in it
I personally had better experience with AMD
Soooo. to each it's own
Crying about Linus criticizing Nvidia sounds like homerism
Linus's comments I think are related to how difficult it is for the open source nouveau driver to properly work with nVidia gear, especially newer chip sets. nVidia's proprietary drivers I have found to be very good, although the fact that they don't support dkms is a major PITA. Without support for it, every time you update your kernel, you have to reinstall the driver! Ditto, booting into a previous kernel. Personally, I would give nVidia a major attaboy if they were to open their systems so that open source kernel drivers could be more easily written, or to (similarly) provide FOSS of their drivers and firmware - both are needed to create a fully open sources driver for these boards.
Points for nVidia:
1. Their drivers work really well with Linux.
2. They have always responded (to me at least) to dealing with bug reports. I have not had any problems in contacting engineering personnel in getting issues resolved.
3. They provide thorough documentation for software integration with their boards.
Points against nVidia:
1. There is too much proprietary (hidden) stuff going on to make developing foss/3rd party drivers very difficult.
2. They don't support dkms. making system updates problematic.
3. See number 2... (that one is worth several negative points on it's own).
I'm going to give Torvalds the benefit of the doubt and assume he might have overlooked this on the mailing list: http://www.phoronix.com/scan.php?page=news_item&px=MTA0ODE but NVIDIA proposed a very reasonable solution to the Optimus on Linux problem, and everyone just started arguing that it needed to be more Stallmanesque. Sitting there demanding that NVIDIA open up[ their source code is silly.
I never experienced any problem with NVIDIA linux drivers, whichever card or OS version (since 2001 and counting, with tens of boards). However, I tried to install Ubuntu 12.04 with a GeForce GTX550 Ti, and installation failed. I could not even finish the install process (just a purple screen). I was really disappointed, and fell-back to Ubuntu 11.04 with NVIDIA version of their board driver.
Actually it costs them more to develop their own binary drivers in house, than it would to open source them and allow third parties (including the kernel devs) to take up some of the slack...
No, it wouldn't and even AMD agrees on this. Between licensed code, patents, DRM and a host of other reasons an OSS driver could not fulfill their customers' requirements on Windows. That is to say OEMs who have obligations to Microsoft who have obligations to the MAFIAA, not you. Not to mention the FPS junkies would probably not enjoy the performance hit because the third party IP would have to be ripped out.
Effectively you can either share closed cross-OS code like the Catalyst/nVidia driver does, or you can share open source code with the rest of the Linux community. Those two options are mutually exclusive, and no the those few bits you get from the community are no replacement for the 90-95% "free" cross-OS work the closed driver gives them. What you say is only true if you could flip the switch from "closed" to "open" and have everything you had before plus some more free work.
In a technical, nerdy sense you of course could do that since code doesn't magically break or become useless by publishing it. But legally that is impossible, they'd get sued to hell and their permissions to play any kind of DRM'd media would go away and with it any rights to be in a "designed for Windows $foo" sticker machine. How bad is that? Well, I'll let Bridgman that handles the open source OSS driver answer that, this isn't an official statement or anything just a forum post but:
If the worst case was as minor as not being able to release any more specs we wouldn't be worrying so much. The kind of risks we are worrying about are much larger, ie things which would either kill or cripple our graphics business.
Worst case is that we lose the ability to sell our products into the Windows market as a result of releasing info which results in our DRM implementation no longer being considered sufficiently robust. Without the Windows market (which is >90 % of our revenues) we would, for all practical purposes, cease to exist as a GPU manufacturer, especially since we would probably lose the Mac market at the same time.
Next worst case is that we find a way to continue shipping into the Windows market but get sued under one or more of the DRM-related agreements we have signed. These all have high dollar-value penalties, again enough to significantly impact our ability to continue operating.
There are a bunch of smaller risks but we spend proportionally less time worrying about them. What makes all this complicated is that we have to consider not just the information we release but the information which is likely to be reverse-engineered and published. Each time we release information we simply raise the bar for where reverse engineering starts, and it's the combination of released plus "likely to be reverse-engineered" info that we need to consider.
If we tripped any of these risks then the impact would not only affect the GPUs we are shipping today but anything we have in the pipe. Best guess is that we would lose the next generation (ie the one after 7xx) and see significant delays in the one after that.
Since we don't want that to happen (right ?) the alternative is to trim back the information we release until it appears safe, going through the review process each time until we find an appropriate level.
That's the consequences of open sourcing the binary driver, could it get much clearer?
Live today, because you never know what tomorrow brings
I've been using Linux for quite a while (since 1992 or thereabouts) both for work and gaming and have had good results with NV hardware.
However the latest drivers (301.xx or thereabouts) run the 8800gt's fans at full speed even after the system boots, which seems to be a bug affecting quite a few other users across all major distributions.
Ever wondered whats wrong with the world? http://www.ishmael.org/
FUCK YOU :))
"It feels like I'm at the Zoo when reading this thread - I'm frightened, but it's interesting" (c)
In my recent experience, it's completely 180 now. Had huge problems with nvidia, both getting 3d to work and dual screen.. In the end, since this is work pc, switched to an amd card, it came right up, and it has worked perfectly since.
It's The Golden Rule: "He who has the gold makes the rules."
His immature and vulgar comment goes to show the kind of guy Linux Thorvalds truly is. I'll never touch anything related to Lunix because of the childish attitudes of the people who support it.
Great story.
> However, I don't care about 3D
Welp that's you being made irrelevant right there. All mainstream operating systems these days use 3d compositing to render their desktops. This makes the desktop more responsive, more intuitive and easier to use - not to mention a whole load of other plusses as well - and this trend isn't going to reverse any time soon.
Of course, I know you know this. You're just posting to brag how your own fringe use-case works for you, even though nobody else would be interested.
> I've been using ATI for more than a decade with zero problems
> I think there's anything wonky going on with video support, I run "X -configure" as root and the wonkiness disappears.
lol what
Absolutely, the nVidia drivers wipe the floor with ATI and Nouveau. Once you get your head around the x-org configuration for older cards and the hit and miss "will it compile on my system" that is!
I use my machines mostly to watch movies, video, etc. I have ATI, Intel, and nVidia cards. The only card where I could get hardware acceleration for video decoding was nVidia (via VDPAU in mplayer). That has worked for 3 years now, while I still can't get other cards to decode video in hardware.
Hey I have this Dot-Matrix printer, the driver says Windows 3.1, I'm trying to install it on Windows 7, but it just won't work. I mean Windows is Windows isn't it? You seem to really know what you are talking about, maybe you could help me out?
If you look at the versioning between Windows 2000 (NT 5.0) and Windows XP (NT 5.1) then compare that to Windows Vista (NT 6.0) and Windows 7 (NT 6.1) you might start to understand why you can use a Windows 2000 driver with a Windows XP driver and a Windows Vista driver with a Windows 7 driver, probably because if they arn't using the same kernel, they are using a kernel that is only off by a couple sub-versions, still in the same generation of NT kernel.
Your problem is that in Windows you can not visualize the kernel versioning as richly in a project like Linux.
Additionally, most big vendors that do sell Linux commercially tend to pick a kernel version and stick with it for the duration of that version of their product. Lets take RedHat for example, even in version 6.0 they are still using Linux 2.6, the next kernel update they do for 6.0 of RedHat is going to still be a 2.6 kernel.
So if you had been using a 2.4 kernel and decided to update to 3.5-rc3, don't complain about things breaking. That would be roughly equivalent to you taking Windows 7's kernel and trying to boot off of it into a Windows 2000 system.
Every so often a kernel is picked out of the generation and designated as long term support, and new stuff is back ported to it. Big vendors (and anyone else who chooses to) is free to decide that they want stability, thus choose to build there system around the long term support kernels.
As previous posters mentioned, the issue is with not only the sheer complexity of a GPU but OpenGL's slow update process too. Having a functional driver is only half the battle (which typically the older drivers DO work with newer kernels). Without OpenGL support, the driver isn't worth crap. To the point where people (in particular game developers, even AMD has its own version of OpenGL they package with their proprietary driver) tend to fork OpenGL in an attempt to build in features they want but somewhere getting one unified, up-to-date OpenGL version breaks down.
The issue of market share mainly has to do with people not being able to walk into a store other then the Apple store and being able to select something with an operating system other then Windows. How many people do you know that actually want to/know how to/even know that it is possible to install a different operating system on their computer?
It is very clear you have no clue what you are talking about, nor the slightest idea about how kernel development or driver development works on ANY platform and should just stop talking, I feel dumber for having actually read your post.
A LOT FASTER!
Sorry, but I don't see the story submission as anything other than a troll attempt by either ATI and/or an ATI fanboy. Its extremely well known that NVIDIA, by a wide measure, has the best Linux support around. That, however, doesn't mean at the kernel driver developer level, they are a joy to work with. The fact that such an obvious distinction is purposely directed to imply NVIDIA just be a shitty driver, is complete stupidity.
Long story short, NEVER get a laptop with ATI unless you plan on replacing your laptop fairly quickly; as in a couple of years. I've been bitten by two ATI cards and two laptops by ATI. NEVER been been by NVIDIA. The lesson learned - will never buy another peice of shit GPU from ATI. Their drivers suck ass. Doubly so on Linux. NVIDIA's drivers are super easy to install, regardless of the platform. They work. Period. And they dont' EOL a chipset just because they like pissing people off, like ATI does.
Sorry, but ATI has fucked me and my friends for the last time. Not to mention, ATI literally laughed at dumb Linux users. Sorry, but ATI's obvious half ass attempts to bring driver parity to Linux is for the foolish and dumb. Hell, ATI doesn't even have peer quality drivers with NVIDIA on Windows.
Yes, but not for the reasons your post implies.
At the time, NVIDIA had the largest third party, proprietary driver available for Linux. They had oppps. So did just about every other third party, proprietary driver. The problem is, the kernel drivers constantly changed the internal APIs which caused breakages of these proprietary drivers in subtle ways. NVIDIA had roughly the same % of breakage as everyone else. Its just they had far more users which resulted in a larger number of reports.
So yes, they created, "tainted" kernels BECAUSE NVIDIA showed them it was a requirement of the kernel developer's bad habbits. Not to mention, the simple fact is, its extremely difficult to track down kernel bugs (frequently these opps actually were kernel bugs, but they also wasted a lot of time on driver bugs too) when you don't have the source. So to save time, they simply created the "tainted flag."
So while your post implies "tainted" exists because NVIDIA sucks, nothing could be farther from the truth.
for the desktop you are correct.
The current Linux model is actually extremely friendly to developers. AKA if you make hardware. Lets say I am going to make a router with Linux.
I have no problem getting the drivers for the chips I am using since I can decide to pick the chips I want.
No problem with feature control with the hardware I want. I can put in a USB port and put a different price tag and sticker on it based on what the USB will support.
Yea I will put the sources up on my website but I can require a signed package for updates. And even if so one figures out how to flash it anyway big deal. I will have a new model on the market that breaks that.
Updates? As long as I sell it sure but once the new model is out who cares.
BTW You can keep Linux updated if you use Intel everywhere. Sure the Intel video chips suck but they are in the kernel.
BTW this is one reason that Linux does well in servers and the embedded space. They both offer tight control over hardware selection.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
The GP wasn't talking about UEFI. They were talking about the nvidia - easy - rebuild the driver after a kernel update/upgrade. If you use the nvidia installer, you'll have noticed that X dies after a kernel upgrade. However the reinstall of the nvidia driver is just that. You literally just reinstall the driver. The install procedure automatically recompiles those parts necessary to make the driver work with the new kernel. You don't need to download the next updated version of the driver.
This has nothing to do with UEFI and everything to do with Linux having regular/reliable kernel upgrades. Because all the other drivers are OSS, this means that they get updated automatically with the kernel. However, since the NVidia driver is a blob it doesn't auto-update with the kernel and thus a kernel update kills the driver. So NVidia built (or commissioned) an installer/wrapper that will recompile the components that interface with the kernel easily and reliably.
Now that's not to say that UEFI doesn't influence this. Because of course now you'll have to update and SIGN your kernel before it will boot. Since nvidia has already built an automated tool for rebuilding their driver to match the kernel, they probably have it easy compared to the kernel devs/end users on this front. NVidia can update their installer to include using the user's/computer's key to sign the rebuilt driver as part of the reinstall script.
Caveats:
While not an nVidia fanboy, I do have only nvidia cards. This is more a quirk of history than through specific planning.
I do wish nvidia would open up their drivers because I would prefer being able to trust my computer. My next purchase is going to be AMD for this reason.
The problem is that the binary drivers don't always survive kernel updates So people who rely on nVidia's proprietary binary drivers can't always update their kernel or they lose their graphics until nVidia puts out an update.
There's an easy button way for Nvidia's drivers to survive a kernel update on Fedora:
Install rpmfusion repositories.
yum install akmod-nvidia
Do that once, and you never have to worry about kernel updates....or the F16 to F17 update either, messing with your Nvidia driver.
I think I made very clear that I have very little concern for the opinions of Linus.
If you can't follow Boost / Design Patterns, it makes sense that you program in something as ridiculous as Erlang.
That's funny because my 8400 had no problem driving my monitor to it's native resolution. Even the libre drivers were able to do better than 320x200.
When I first installed Linux on this machine, I had trouble getting it to do anything over 1024x768 (or was it 800x600) can't really remember, but it sure wasn't 320x200. But the REAL problem is that my monitor (which is actually a small HDTV) doesn't properly respond to EDID requests over VGA, so I had to do some xorg.conf editing.
Now when I bought a GT200 and installed it, I switched to HDMI...and I got remove my xorg.conf edit, and let it handle configuration automagically.
That makes no sense whatever, and sounds like a really fucking BAD, lying excuse for Nvidea's hostility to Linux. If a thing is patented, its design and workings are open knowledge. Unless, of course, you're telling us it has to keep its shit secret because it knows it's willingly violating other companies' patents? Either way, it's evil.
Free Martian Whores!
I do NIC drivers for pretty much all popular *nix OSes. Linux is, by far, the biggest PITA to develop for. Developing for a particular version of Linux is fine, but keeping a driver compatible with all commonly used variants is murder.
Almost every other OS, even open source OSes like FreeBSD, maintain a stable binary device driver interface (DDI). That means that a module compiled for one kernel will work on any other in that major release series (and, depending on the OS, in future releases). For example, my company's NIC drivers compiled for S10 work just fine on both OpenSolaris and Solaris11.
Linux does not do this. Heck, they don't even maintain a stable DDI between the same kernel version compiled with different options. Worse, they change their APIs for no sane reason, adding and removing function arguments, struct elements, etc, just because somebody looking for name recognition wants to "clean up" something.
So if Linux had some kind of stable DDI like,. well, everybody else, a lot of these problems would just go away.
Before somebody whines "Well, just get your driver into the kernel" --- it is. But our customers tend to want the latest version *without* updating to the bleeding edge 3.x kernels. Which means that we have to maintain compile shims all the way back to 2.6.9 (RHEL4). The last I checked, the compile shims alone were ~2000 lines of code, which is nearly the size of the *ENTIRE* driver on some other OSes.
If that's the case, then you probably heard about UEFI secure boot? You know, that thing where everything, including the kernel AND the kernel drivers will have to be signed?
I swear, kids should learn history, because they keep coming back with the same old bad shit over and over again until a clueless generation accepts it. Back in the late '80s there was a boycott against copy protection, and companies dropped it -- for a decade or so, when it was reborn as DRM, and you kids pretty much accepted it.
Ten years ago they had Palladium and we users squashed it like a bug. Now MS is trying to do the same goddamned thing all over, and you kids aren't complaining nearly loud enough.
Wake up, you're being oppressed. Fight back.
Free Martian Whores!
Well if honestly you HAD a dot matrix printer? Most likely it would work OOTB friend, you wouldn't even NEED a driver. in fact here at the shop i have ran into exactly TWO, count 'em TWO, items that had to be completely tossed for lack of any drivers. 1 was a no name CCC (Cheapo Chinese Crap) analog capture card which only had a 32 bit driver that frankly only ever ran stable on XP SP2 so naturally trying to find a driver for a technology no longer used like analog broadcast was pretty much a lost cause, The other was a scanner from 1997 which again the device had gotten so old and the DPI so low that while I probably COULD have found a driver as i pointed out it was cheaper simply to replace it with an all in one.
That doesn't change the fact that Linux drivers are deep fried ass on a stick because its a damned miracle to god if even 80% of the drivers are fully functional after a SINGLE update. With Windows the device will in all honestly be most likely dead before the version of Windows it was originally written for is EOLed, after all XP has two more years, Vista is good until 2017, and Windows 7 until 2020. Frankly VERY LITTLE hardware is gonna be functioning past those dates in the average home or office. Ever try surfing on one of the PCs that came with XP RTM? I have for shits and giggles, its a 733MHz Compaq with a now whopping for the time 384Mb of RAM and just loading a modern web page is quite painful on something THAT old.
So your argument? kinda pointless. I can take any desktop or laptop in the shop and slap Win 7 for it and I seriously doubt I'd even have to hunt for a driver, the included drivers already work. Can you say the same for Linux? will the EXACT SAME DRIVERS you are using now even be functional in a year if the machine is fully updated? How about a graphics driver from just 4 years ago, will it work? Most likely the answer is HELL NO! And until THAT is fixed my friend linux will stay a geeker toy, because retailers like me as well as the big boys like Best Buy and Walmart will avoid it like the black death mentioned in the earlier article. Your product? it be broke, and that ain't no joke.
ACs don't waste your time replying, your posts are never seen by me.
And since they won't produce drivers except certain versions and makes of hardware for certain limited OSes, they don't actually work. Just temporarily do some things.
ATI can reverse engineer their work no problem. They have the technology and the expertise.
There is no advantage given away to their competitors by releasing "propriatory information".
Now, if you wish to disagree, prove it. What info.
Nvidia may suck to work with, but thier drivers work well, and get the job done.
Im all for using the best tool for the job. For the same reason I dont use linux to game, I dont use windows to try and host. Everything has its pros and cons, and as far as Nvidia goes, its thier hardware, they arent required to support anything besides what they want to. I for on applaud thier dedication and time spent supporting linux when for years of thier support, it was very much a niche market.
Sure they dont play nice, but at least they play.
I agree but why can't AMD and NVIDIA define an API for the functions they need and get that into the kernel? Then they have a stable platform to base their drivers on.
thegodmovie.com - watch it
I mean for the love of Christ they call me an "MS Ninja" for DARING to point out you can't sell Linux to consumers in its current state when what do we see right here on this very forum? That unless the hardware OEMs constantly release new drivers, even for old hardware that isn't even fricking being sold anymore, that the software WILL NOT WORK with current kernels.
I see these complaints all the time, but in ten years of using various distros I've only had ONE problem with drivers, and that was over 5 years ago (maybe closer to ten). I got a new video card with an S-Video output so I could use the TV as a monitor, and although it worked with a real monitor, it was only garbage on the TV screen under both Suse and Mandriva (there wasn't any Ubuntu then), but it worked fine under Windows. That card is now happily humming right along on kubuntu 10, so someone has obviously fixed that problem.
OTOH I've had many driver problems with Windows. When I first upgraded from 98 to XP, the first thing Windows did was to replace the perfectly good network driver with one that didn't work at all. I have another box, a real old one I'm fixing for a friend that had XP with a nastily corrupted registry on it. I FDISKed and formatted the drive, reinstalled Windows, but the Dell driver disk seems to contain no drivers. Of course, no NIC drivers and Windows stupidly asks me if I want to look for them on the internet... ON A MACHINE WITHOUT A NIC DRIVER! I just hope the ones I DLed from Dell last night on a different machine work, because the latest version of Linux I can get to run on the old underpowered thing is Mandriva 2005.
You want Linux boxes in your shop? Easy enough to do, just install it alongside the already installed Windows. Of course, you'll lose money on support, because Linux doesn't get bogged down like Windows does, and people don't bother writing malware for it.
there should be no damned reason why my customer should have to update his GPU driver just to run the latest XBMC or any other software for that matter
Hmmm, I've been thinking of trying XMBC out, I'll have to get back to you on that. But I've never seen an app update break any drivers (although like I said, the newest Flash won't work without a gig of memory).
Free Martian Whores!
... even one of the big Red hat developers says the current way of doing things simply isn't sustainable, that a single group can't control 20,000 packages and drivers and keep it working, and recommends an ABI
Ingo is talking about application ABI, not driver ABI. He's objected to proprietary drivers before, and I don't recall seeing any evidence of a change of heart.
Thanks to virtual memory and so on, applications have much more of an arms-length relationship to the rest of the system than drivers. That makes maintaining a fixed driver ABI more work.
After all how do you expect the smaller hardware guys to support you if the big guys have to pay entire teams to constantly fix the damned things just to make the drivers work?
In the judgement of most kernel developers, the most efficient use of limited resources is to write an open source driver that can be included in the upstream kernel. That makes it easier for other kernel developers to collaborate with your your developers on future maintenance.
I swear, kids should learn history
Actually, kids did should learn history. History has shown that "fighting back" is temporary. Eventually history repeats itself and oppression comes back
So kids today (or rather, the kids during/after the Baby Boomers) smartened up. Why fight the inevitable? Instead, kids live for themselves, making the best out of tyranny. That usually means becoming a tyrant yourself, and that's where the whole proprietary software thing came from (which falls under the umbrella of "intellectual property", copyrights, patents, etc.)
same here...flawless.
Because when I buy an extremely expensive piece of hardware made exaclty for performance i want PERFORMANCE. And the community has proved to be UNABLE to do that!
I use only linux as operating systems and have zero doubt on when I go buy a video card, its NVIDIA or nothign, EXACLTY because of the REAL support for linux. Intel and AMD support are FAKE support, are just imaginary fairy tail support that is useful for nothing! If AMD open ups specs but the final result is crap then this is completely USELESS!
I hail and praise NVIDIA on that, if it was not for NVIDIA I would have droped linux a LONG time ago.
THIS. When I'm looking to buy a laptop, I don't even look at ones with ATI graphics. Why is Torvalds letting ATI off the hook when they're MUCH worse than nVidia?
Spin master spin is in di hisouse!
There were several possible ways to maintain a driver. Open source contribution was the least work for the contributor. Those tended not to break at all because the kernel developers could see who depended on what and either avoid breaking it or just fix it in the driver. nVidia didn't choose that one. The LK developers didn't mind looking into an oops there even if they ended up debugging the driver.
Next up was not so preferred, but was sometimes used where you have a binary blob and then glue in source form. The end user can compile the glue against whatever kernel he has and link it to the blob. Those often worked from one version to the next. Here though, only the producer of the blob can be responsible for debugging the driver but they may get really good bug reports since there are public symbols to look at and even instrument. nVidia didn't go with this one either.
They chose the 3rd way, driver is a blob. Vendor is responsible for updating the blob whenever a new kernel comes out. vendor is responsible for debugging any oops that comes about. nVidia chose this one but lagged behind on releases and wasn't so responsive on the debugging thing.
That doesn't change the fact that Linux drivers are deep fried ass on a stick because its a damned miracle to god if even 80% of the drivers are fully functional after a SINGLE update.
This is completely and unnecessarily total crap. FOSS drivers come with the kernel so 100% of them is fully functional after an update. Binary blobs are external software and are potentially and mostly 0% functional after a kernel update. Binary blobs are .1% of the kernel drivers in number. Unfortunately most non-thinking new wave ubuntu linux users own a nvidia card AND they can enjoy the binary blob right now, when it mostly works. They just did not see the crappy years of guess configuration of XF86Config to make nvidia work for the display resolution, the cryptic and non-sensical Xid dmesg messages, the hibernation madness on laptops, the installation script overwriting libraries in /usr/lib and so on.
I have had a nVidia and now an ATI Radeon. With nVidia I have a good performance but some crashes. With ATI Radeon the performance was inferior, even though the hardware was better (dual core, more memory, higher GHz); and more crashes. With the open source no crashes and no performance. At work I have an Intel GPU and and the performance seems to be way better than the ATI Radeon with the open source driver, and also 100% stable. Therefore, I won't buy anymore ATI or nVidia, just Intel for PC and laptop.
Global iconic stardom is probably not the best working condition for a software developer, IMO. Devs are notoriously out of touch with their user base, and I'd call this an example of that. I hope Nvidia doesn't take Linus seriously here... I don't know when Linus last played a video game on a PC, but those of us who work and play on the same machines, can only do so on Linux because of Nvidia's driver support. Put another way: If Nvidia's Unix Drivers team got hit by a bus tomorrow, it might actually have negative effects on the open source community. Hackers who take breaks from coding to pwn noobs aren't going to want to swap over to Windows for a 10 minute bloodbath. If Nvidia's Unix Drivers team says, "fine, then, fuck you too!" some of those coders will be writing software for Windows afterwards. Someone get Linus a cold beer, a blowjob, and a 3 month vacation before someone takes him seriously.
Every trollism an AC posts is prefixed, in my mind, with "A. Coward whined, in a weak and cowardly voice:"
He doesn't know what he's talking about because he hasn't bought ATI in a while, but you're totally on top of things despite not using NVIDIA in over a decade? Uh, okay.
Are you guys still living in 5 to 10 years ago (at least on laptops). First of all every since Nvidia came out with that damn Optimus hardware Nvidia/Linux compaitibility has gone to crap. AMD however has made great strides in getting their drivers to work in Linux. Almost every compositing desktop uses AMD cards wonderfully now. (I'm running Gnome 3 with AMD HD 6470m and Catalyst 12.2 drivers on Ubuntu 12.04, Unfortunately my wife is using Optimus and while the Bumblebee project (now on 3.0) has made great strides getting her to sort of game with her linux laptop it's still a half-assed backwards system of doing it. My wife isn't really a commandline user but now she needs me to make scripts as shortcuts so she can play).
I'm sorry but today, as far as laptops go, if Nvidia doesn't get on the ball with supporting Optimus in Linux I can't buy them anymore.
That's just how little forever alone Linux weenies like him speak. They aren't very good in social situations.
They have a binary driver so only they can work on it. They just barely released a beta driver that supports multiple rotated monitors with randr and dynamic changes of resolution/screens (think laptop docking and undocking).
I'm not impressed. The current non-nVidia open source driver (Nouveau) is better for normal day-to-day stuff.
Well, there's your problem. Device drivers should be written in C, not as shell scripts.
In the newer K/Ubuntu's (last couple of years?) it's even easier. The first boot after installation, there is a pop-up that asks if you want to use proprietary drivers. Two mouse clicks and a short wait later, they're all set up.
God is imaginary
Sorry, but you didn't understand what EFI secure boot is about.
It took us 6 months working with nVidia support to get a 3D quad-monitor display working. Our support contact was willing to help, it was just that it took a long long time before we were allowed to get direct access to an engineer and things finally began to move. It took a couple of driver updates and reverting to an earlier release of SuSE linux. The quad display worked really nicely once it worked but there were many times we thought we would never complete the project. (And it was expensive too, with the necessity of buying a certified computer system otherwise their software would refuse to run, not to mention the two large Quadro cards.)
I've had no trouble with Intel's integrated graphics, ever.
I've also found that to be true in general, but it's worth noting that there is an entire series of Intel-branded integrated graphics chips with PowerVR-based GPUs which have no Linux drivers worth mentioning. Even 2D support for these chips is quite limited at present.
"The state is that great fiction by which everyone tries to live at the expense of everyone else." - Bastiat
I personally think that the Torvalds didn't know what he was talking about. NVIDIA drivers on Linux might be closed and they do cause kernel developers problems but for the consumers the NVIDIA on GNU/Linux is the only reliable, good quality, way to 3D acceleration and their support is top class.
Other than that I'd like to point out that the girl was complaining how the absolutely latest NVIDIA Optimus chips two years ago had (and probably still have) pretty terrible Linux support. It's hard to make Intel's integrated graphics card to play nice with NVIDIA's graphics card with Optimus in GNU/Linux. The Point in this conversation wasn't the NVIDIA's general 3D Linux drivers but what kind of support NVIDIA offers for their technologies. And Torvalds answered that it is *terrible*. However I find that hard to believe. As I stated above. The Only reliable good quality 3d accelerated rendering hardware and software stack to GNU/Linux comes from the NVIDIA.
Probably helps to know what you're doing. Plenty of people have no problem with compiz and 4890s - I did it myself and it ran flawless, actually.
Maybe you should try installing drivers first as that tends to help.
Why should Nvidia subscribe to the projects "goals and visions"? Thats the projects concern, not theirs.
As long as NVidia is interested in GPGPU programming of their GPUs on Linux, it is NVidia's concern.
It's a niche market but it's with high margins. NVidia could leave but that would mean that AMD is pretty much uncontested and AMD supports Linux: http://developer.amd.com/sdks/AMDAPPSDK/downloads/Pages/default.aspx
If you really believe that you're crazy. Many core kernel developers, Linus included, are on record as hating the idea of stabilizing the Linux kernel driver interface at either the binary or source level. And I don't mean "hate" in a trivial way, they're passionate about this. They believe that to build the best kernel they can, they must be free to change internal interfaces at will, and since Linux is a monolithic kernel that includes driver interfaces.
Historically there are countless cases where a core Linux kernel developer got an idea for how to streamline a driver interface in a way which would totally break backwards compatibility, and just went ahead and did it, grepping for every in-kernel-tree driver which was affected and hacking every one to match the interface change. They just don't care if they break drivers which aren't in the kernel tree. They believe all drivers should be open sourced and in-tree so they can make such internal optimizations at will, and refuse to make any concessions to outsiders who want to distribute binary drivers. In their view, the fact that entities like NVIDIA have to do extra work to keep up is what they deserve for not playing the game the way the kernel developers want it to be played.
So no, NVIDIA didn't set out to deliberately tie drivers to particular kernel versions. That would be stupid of them -- it makes extra work for them! The problem is simply that they, as a company, do not want to open the source to their drivers, and as long as that's true there will always be friction with the Linux kernel and the community which develops it.
Nvidia is the reason they came up with the mantra, "Don't use a GUI package manager!", because if it upgrades X, it'll kill the pkg. mgr. doing it.
No.
That is because when you patch a Linux system, shared libraries and other resources are overwritten with complete disregard to the processes using them. With the exception of some things like the MySQL RPM which handily restarts your database for you [awesome] your processes are left in limbo.
Your bash process is much less likely to be affected than say a Gnome or KDE app with tons of dependencies.
Maybe you should read this article where one of the RH devs points out having the kernel devs take care of drivers is a FAILURE and why linux on the desktop is in "its death throes" because a single team can't control 20,000 packages and a couple of hundred thousand drivers and end up with anything other than what we have now, a broken mess.
Its no coincidence that the ONLY places that Linux has gained ANY traction is places where the hardware is extremely limited and controlled, servers (many servers still using Rage II for graphics, parts rarely change) embedded (parts carefully controlled and never change) and cell phones (ditto) because frankly that is the ONLY way to have a Linux system functional for any length of time and logically you should obviously be able to see it. A handful of guys simply cannot control and deliver QA and QC to the amount of code you are talking about, dozens if not hundreds of millions of LOC, which is why you have a buggy mess that doesn't upgrade worth a shit. here is another article with over 100 links of examples but I have a feeling from your tone you'll ignore it, as I have a feeling I'm talking to a FOSSie, aka one who treats Linux as a religion and not an OS since you blame everything on those that will not beg for the devs to bless their code. BTW AMD GAVE YOU THE SPECS, how's that working out?. Since the vast majority of the planet couldn't give less of a crap about the 4 freedoms as evidenced by the lines around the block camping when the latest iDevice is gonna come online, you really have to have something more than dogma, aka a fully functional modern OS that doesn't require forum hunts and bullshit just to keep running. Sadly your OS is pretty damned far from that goal and if anything getting farther away by the day.
ACs don't waste your time replying, your posts are never seen by me.
Maybe you should read this article where one of the RH devs points out having the kernel devs take care of drivers is a FAILURE and why linux on the desktop is in "its death throes" because a single team can't control 20,000 packages and a couple of hundred thousand drivers and end up with anything other than what we have now, a broken mess.
I read the article, not only that, I also actually UNDERSTAND and agree with what Ingo Molnar is saying. You even ignore the difference between a distribution managed software package and a kernel driver, but still, you feel the need to comment on that. So any further discussion turns to be futile. No offence, you are the kind of user (the one who wants to use without understand) linux does not need.
Linus isn't talking about gaming, performance or anything else like that. The point is : nVidia ships a binary blob and an obfuscated source portion that needs to be built outside of the vanilla kernel. That is what Linus is talking about, nVidia's lack of cooperation with the kernel people at integrating their drivers into the main line kernel in a way that respects the project's goals and visions.
Why you people are discussing the performance when that is not at issue, I have no idea. It was all pretty clear to me what Linus meant.
nVidia does that because that's how they protect themselves.
Drivers for Linux suck because it doesn't have a valid ABI, and not even a very stable software API.
Fix that problem, the political bullshit, and the "let's change it just because we can" mentality, and you might see things change for the better.
Now we just need to check back in a month or two and see if you've gotten increased support from Anonymous Coward!
Read the above again. GPUs are not just for gaming anymore, several of the top500 supercomputers now use GPUs and they are linux based.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
You're the one spewing the bullshit. The open source ATI drivers are orders of magnitude more stable than either ATI's or NVidia's binary blobs.
They're not very fast, but as far as stability goes, they're right up there with the best of them (every platform included).
What kind of OS changes their driver API so much that a driver compiled for one version doesn't "just work" with an incremental update? An immature one.
I guess that this is the 'stable API nonsense' discussion again and again read that text if you wan't to know why having stable binary in-kernel API in Linux is near to impossible.
And even when overlooking the technical difficulties you know that evolution is driven by changes, without changes you are one step from extinction...
There are actually less people reverse engineering the amd drivers than nvidia, because they have more to work with...
Same as happened with the PS3, people initially wanted to get GPU access under linux (which would have made media center use far more practical among other things), but once linux access was taken away the attention moved to getting it back which also resulted in an easier opening for piracy.
Any drm scheme will be broken sooner or later anyway, assuming there is any content that people want which is encumbered by it and isnt available in a superior form elsewhere (eg whats the point cracking poor quality streams when you can crack bluray). Any drm scheme is purely a form of obfuscation,
Patents are by definition public, and sourcecode which implements them would not be infringing if it only worked in conjunction with hardware that comes with a patent license anyway.
Of course, companies which are beholden to third parties rather than to their paying customers is a gross distortion of the free market.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
"What has been your experience on NVIDIA drivers with Linux?"
Better than my experience with ATI drivers with Linux. If I wasn't so dependent on ATI being superior to NVIDIA under Windows in terms of compatibility, I'd ditch ATI altogether and go NVIDIA all the way.
But then, the way Windows 8 is going, I may end up ditching Windows altogether. I only have it for games.
there are obvious bugs which they refuse to fix. one must install their driver and outside software to configure it. all of this just to get some stability.
AMD has comitted in trying to improve the whole process for open sourcing future generatino of graphic cards. Anything which wasn't in the pipeline until after AMD started to help open sourcing the drivers is more and more being designed with collaboration with open source in mind, so there is going to be less ping-pong between lawyers and designers before specs can be released.
Also the process of releasing specs and helping the OSS drivers is also becoming more streamlined.
All this mean that the delay between the release of new hardware and new OSS support is getting shorter. AMD has been reported hoping that HD 9000 and beyond could get same day Linux support (as is already the case with Intel hardware, for example).
Now what we also need is more workforce to similarly bring up to date the generic part of the stack (the OpenGL state trackers, etc). As AMD is having paid developpers for Linux and Intel is shifting toward Gallium (and even non-graphics companies like Google are getting interested - because of their netbook OS), that could bring also faster support for newer standards and the gap between OpenGL standards and Linux support could close over the next few years.
(Look at the situation with OpenCL: this standard was released more recently and with all the current expertise in writting drivers, OpenCL support got added rather quickly and is not lagging too much behind the published standards, although OpenCL isn't that much important for graphics - where most of the developing energy goes).
So, okay, maybe today, development is still partially hindered by bad documentation and necessary legal department clearances, but the process is getting slowly better. In a few year, AMD could probably offer acceptable opensource support out of the box.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Been saying that for years. ATI/AMD did what the community wanted and guess what the community found out? Writing good graphics drivers is really freaking hard.
And none the less, since this initial support, the opensource drivers have been steadily becoming better, with each generation a shorter time gap between hardware release and linux support, etc.
Yup, no perfect OSS drivers did insta-magically show up on the next day after the spec release. But nobody said that drivers would be written in a snap. It takes time, but on the grand scale of things, OSS drivers for AMD are improving, all thanks to the support of AMD.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
ATI is more than accommodating and helpful--and the open source community fails miserably at creating stable drivers.
Nobody promised that perfect OSS drivers will insta-magically show up on the very next after AMD releases the spec.
And OSS support for AMD is steadily becoming better (as measured by phoronix, for example), thanks to the help of AMD.
So in the long term AMD's strategy *IS* working even if it's not perfect yet.
Last but not least, thanks to these opensource drivers you can use older hardware on modern distributions. Otherwise you would be stuck at keeping the latest blob that supports the hardware, and not upgrading the kernel beyond anything that breaks the API with the driver.
If Nvidia opened up their specs more there's nothing to suggest that the drivers would be any better than the Open Source community's crappy job with ATI drivers.
If Nvidia opened up their spec, indeed, you won't see perfect binary-like level of support on the next day.
But if AMD is any indication, if Nvidia opened up the spec, that would help the Nouveau project a lot.
At least Nvidia offers a workable driver set even if it's a binary blob that gives Linux a workable video card option.
Except that the hardware is a moving target, Linux itself is getting developed too. But Nvidia isn't being helpful in trying to get everything working together.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
They took ages before supporting some basic stuff like XRandr (and the latest version of XRandr was to be specially crafted to make it easy to add support for Nvidia).
Kernel Mode Setting (KMS) is still missing.
Some other feature like Optimus are still missing or not well supported.
Multimonitor is support in their own specific proprietary way, meaning that sometimes it breaks some software not used to it. (synergy does some crazy stuff when switching monitors in nvidia-settings).
etc.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
All hail the one true god! You DO realize that is what you sound like, yes? How can you agree that the half of the system you don't agree with is broken but the half you do agree with isn't? It isn't magic you know, you only have X number of guys qualified to do that job and X+Y100 times the amount of work. I'm sorry but the math just doesn't work, it can't be done. I could post link after link, hell page after page of drivers that work in foo that don't work in foo+1, but what is the point? There were over 100 links in that last one yet you chose to ignore it.
And THIS is why I call them FOSSies, because it ALL appeals to emotion, ignoring of facts, it is like trying to have a discussion on evolution with someone who believes Adam rode a dinosaur, its simply not possible. In just the consumer desktop space you are looking at a good 200,000 drivers, now HOW MANY are working on the dev mailing list? 20 guys? 40? You simply can't do what you think can be done, there is simply not enough hours in the day for guys that ALREADY are in charge of maintaining the kernel to provide ANY QA OR QC for that amount of drivers. Anyone that can count can see what you are suggesting is impossible, it would be like claiming you could bounce hard enough on a trampoline to break orbit, it can NOT be done.
In the end you don't have the manpower to do what you are suggesting and frankly never will. As i pointed out the ONLY reason it can work in the spaces where Linux has gained traction is because hardware can be strictly controlled and changes aren't allowed until they have been run on test beds and vetted for bugs. this is the complete opposite of the consumer space which is why Linux is a massive failure there, because consumers won't put up with having to have test beds and do their own QA and QC to ensure that the system will function. For an enterprise paying 2 devs $80k+ a year to do that job is cheaper than paying for MSFT CALs for a 2000+ seat corp so it makes sense. Your design? Doesn't work except in niches that are willing to use it DESPITE the massive failings in design. Accept it and be happy with what you have, or work to change it, your choice.
ACs don't waste your time replying, your posts are never seen by me.
Actually I've found you lose money on Linux because its a fucking support nightmare from hell. How do you tell if that new device will work with Linux? you can't, any hardware lists are horribly out of date, its a total crap shoot. Hell try the Hairyfeet challenge, take ANY distro, your choice, from just 4 years ago (less than half the Windows support cycle) and update it to current using JUST the GUI as any normal user would be expected to do. Know what you'll get? A broken mess, that's what. In just the last 4 years you went from ALSO to Pukeaudio, KDE 3 and GNOME 2 to KDE 4 and GNOME Shell, and that's just the top layer stuff, the guts are even worse off, with all kinds of incompatible bullshit down in the networking guts especially. Any Windows machine I sell will continue to function for the life of the Windows install barring hardware failure, possibly even longer as I have a few customers that only recently retired their Win2K machines. Linux? Can't do it, the whole system from the kernel up is in a state of flux and shit breaks constantly.
Finally if you'd taken just 10 minutes of your time before install you'd never have had that problem with drivers as there is this place called DriverPacks where you can simply download a pack with ALL the drivers for damned near every piece of hardware, all compressed with a nice little GUI that will do the work for you. Just pick the OS you are planning on install or do as i did and download the packs for every Windows OS and you are good to go. Once done you can simply slap them on a DVD, put them on a flash, whatever, and run it once you get to the desktop and go make a sammich, it does ALL the work. Finish up with Ninite while you have your dessert and tada! From a blank drive to a fully loaded and ready to go Windows in about an hour, an hour and a half if you use WSUS Offline to install the windows updates which you have had it download previously. I keep it on a shared drive but you can use DVD, flash, whatever floats your boat. Again hassle free and an hour and a half and maybe 4 clicks total, couldn't be simpler.
ACs don't waste your time replying, your posts are never seen by me.
On one hand, Nvidia chips are some of the best supported on Linux, so I strictly buy Nvidia cards and have had excellent results especially for gaming. However, on the other hand, Nvidia has kept the drivers behind closed doors for years. While I'm thankful that they release full drivers that work, I would rather see the drivers licensed under the GPL so that they can be integrated better into the Linux kernel and released with standard Linux distributions. The Nouveau project has good intentions but it's essentially duplicating the work.
I'm a linux rookie and I could only get an NVIDIA card to work with my MythTV box, i could not for the life of me get an ATI or Intel graphics card to work.
Meanwhile, I've been using ATI for more than a decade with zero problems past their initial teething stage. However, I don't care about 3D, gaming, Wine, or MS Windows.
So, you've no idea what you're talking about. Good to know.
How do you tell if that new device will work with Linux?
Just plug it in. If there are divers needed and available, Linux will find them and install them automatically. If there are no drivers, the device won't work and you simply remove it and return it to the supplier.
Hell try the Hairyfeet challenge, take ANY distro, your choice, from just 4 years ago (less than half the Windows support cycle) and update it to current using JUST the GUI as any normal user would be expected to do. Know what you'll get?
In Mandriva or kubuntu, you'll get a fully functional computer that runs faster than before the upgrade. Of course, if a newer component has more hardware requirements than the box you're upgrading requires it will break, but I've never seen that happen yet with open source software, although I have with Flash when I upgraded kubuntu.
Oh, and the "upgrade with a mouse"? One click is all it takes. No reboot unless the kernel is updated.
the guts are even worse off, with all kinds of incompatible bullshit down in the networking guts especially.
Again, I've never seen this in ten years of Linux.
Any Windows machine I sell will continue to function for the life of the Windows install barring hardware failure
You were talking about updating the OS. I've never seen a bug fix or performance update break anything in Linux, ever. As I said, I have in Windows XP. What I spoke of in the first paragraph was the equivalent of going from XP to Vista. A popup appears saying there's a new version, one click and your kubuntu 10.04 (XP) is upgraded to kubuntu 11.01 (Win 7). And you actually believe Windows is less of a hassle??
Oh, as to the hardware failure, Linux is far more fault tolerant or hardware problems. Six or seven years ago, Windows XP started getting flaky; bluescreening and booting itself and all sorts of nonsense, while Linux just plugged along, until one evening it froze -- the power supply had been going out. Linux didn't flinch until the power supply was completely dead, Windows had been flaky for months.
A couple of years ago I had Win 7 and kubuntu on a notebook dual boot. The computer had a hardware design flaw (they've since fixed the flaw in newer machines) that if it was set to hibernate when you close the lid on battery but do something else on AC, if you shut the lid then plugged it in before the lights stopped blinking, it wouldn't boot when you tried to restart it, and you had to pull the battery to get it going again. This corrupted the registry and file system so badly that it killed Windows dead, no problem with Linux.
Finally if you'd taken just 10 minutes of your time before install you'd never have had that problem with drivers as there is this place called DriverPacks where you can simply download a pack with ALL the drivers for damned near every piece of hardware, all compressed with a nice little GUI that will do the work for you.
I prefer not to install anything on my computers from some random site. With Linux, I don't have to. The installer knows what drivers to use and where to get them. The installation is automatic. No searching for drivers, no downloads from maybe questionable websites. If I'm installing windows drivers in a Dell I'll go to Dell's website, for HP HP's. If it's a Dell running Linux, I don't have to do anything, the drivers are there, unlike Windows.
Last year I decided to get a bluetooth dongle to transfer photos from my phone to one of the computers. It had an install CD for Windows and Mac, but not Linux. I figured it wouldn't work in the Linux box, so I installed the files in Windows (requiring a reboot, of course) and it worked. Out of curiosity I plugged it into the Linux box just to see what would happen -- and it just worked. No programs to install, no drivers to hunt for, no rebooting, just plug it in and it works.
Support nightmare? That's Windows, not Linux.
Free Martian Whores!
Know what is sad? i won't do it, since you seem to be trying to answer honestly but every. single. answer. you put is actually a TM on TMRepo. that means its the same lines that have been used so damned often they actually wrote them down with a cute title so you could just paste the numbers. yet linux is STILL less than 3%, the numbers say more than i ever could, as will this link that in turn has over 100 links pointing out exactly what i was just saying complete with makes and models.
So either you are magical, able to sprinkle fairy dust onto the machine, or your memory helpfully forgets bullshit. because frankly the ONLY machine I've EVER seen pass the hairyfeet challenge was a 733MHz Intel box which i promptly threw away because it was too damned old to even get a dollar for. Everything with hardware anybody would want? Or with a decent GPU? or sound? Crapped all over itself.
In the end "Linux is a replacement for Windows!" is a damned lie, its a replacement for a Hackentosh which is a MUCH MUCH smaller niche, and like a Hackentosh you damned well better be picky as fuck with the hardware and no damned well about each brand and be ready to do some hacking, something windows users haven't needed to do in over a decade. In fact if i were to compare Linux to Windows I'd say its right now barely at Win98 but of course that isn't what you want to hear, but think about it. What was Win98? it was a CLI OS with a bolted on GUI which was a second class citizen that wasn't even required, it was buggy and flaky and CLI was often needed to really get anything done or to fix problems. what is Linux? it is a CLI OS with a bolted on GUI which is a second class citizen that isn't even required, it is buggy and flaky and CLI is often needed to really get anything done or to fix problems.
Oh and I don't know what magical place you live at friend but most B&M stores don't like to take open merch, often charge a restocking fee, and have a set number of returns per year. So using your logic frankly it wouldn't take but 1 or 2 stuck with devices to make Linux equal MORE than the cost of Windows. that is why i personally won't have a Linux machine in my shop, a single broken driver will cost more of my time that a Windows 7 HP OEM. BTW Best Buy, Asus, Walmart, what do these 3 have in common? ALL USED to carry linux and NONE carry it now, why? Because just as i said the support costs ate any savings by not paying for Windows. I'm sure you haven't used Windows since XP but frankly since Vista the only bugs I've seen have been PEBKAC which Linux doesn't give you a degree in CompSci by running it friend. Oh and if you think Linux is immune from bugs? There was an infected Quake 3 in the repos for a year and a half with NOBODY catching it, KDELook handed out infected screensavers for ages with nobody catching on, these are just two off the top of my head. look at how simple it is to write a Linux bug. as we saw with Android if anybody gave a crap or if Linux had any real numbers there would be just as many Linux bugs as there would be Windows ones, sorry.
ACs don't waste your time replying, your posts are never seen by me.
What fucking world do you live in?
Please actually watch the vid, understand what's being said, and then comment.
I know it's the fashion to not even look at the actual articles - but wow.
I am amazingly curious. If you can answer this - you are going to rake in so many farking Nobel prizes that it will eclipse the Nobel prize itself.
How can someone be more accommodating when they don't even know what the requirements for accommodation are in the first fucking place? Is it even possible to understand the language being used unless you first re-engineer or reverse engineer that language? No.
That has significant standing within IP law within any IP framework in the world! The current IP framework is actively being recognized as being not worth investigating(!). If you find that you independently recreated something after some other random patented person did --- that proves knowing infringement! Why fucking bother when you can effectively not search and claim that you didn't find anything, but then risk being wiped out financially and otherwise by massive mega-corporations that can afford to lose triple damages just to put the offending company out of action?
How fucking dense are you?
Ultimate capitcha score! "Comprehension"!
Answer to myself as I did write the above. How many times have aggressive lawyers shut me down for something of my own invention?
More than I'd like to admit. I don't have megabucks, neither I nor my family ever have.
I'm more personally proud of my restraint than of my code. Somewhat. Sometimes.
I have never just blatantly executed any of the provably lying on record asshats that sued me out of existence multiple times.
Oh really? Please tell me more about how writing software for Windows obviates the need for Boost?
Linux doesn't have STL or boost? WTF? You just said you used boost on Linux. In fact, it's easier to develop against it on Linux than on Windows as on Windows requires compiling the damn thing from source. You realize STL is a *language* library that all C++ compilers support, right? In fact, STL support has traditionally been better on Linux (visual studio only in the past couple of years started racing with gcc to implements C++0X). Regarding IDE it is true that nothing approaches quite the quality of Visual Studio, but I've found Qt Creator & KDevelop to be very good without the horrible IDE version dependency that Visual Studio injects (one company I worked at had to actually plan the migration from 2005 to 2008 in the codebase & EVERYONE had to use the same version IDE & it still broke the codebase in subtle ways for a few weeks). Visual Studio as a build system is a nightmare.
The Linux kernel is developed in a different way than NT (not just open-source vs closed source, but NT is a hybrid kernel whereas Linux is a monolithic kernel). Since Linux is OSS, there isn't as much of a need for a driver API since drivers live in the source tree; any changes across the kernel are applied to all drivers. Out-of-tree drivers have always been a problem but the philosophy in Linux has always been that it's those developers responsibility to maintain their patches. Also, the kernel interfaces don't change that often anymore. However, when they do then everyone using those NVIDIA binary blobs feels the pain.
Kernel development != userspace development.
Went through that once, and it sucked donkey nuts. So I've been buying Nvidia, because I know it'll work. It works on my windows gaming machine and it works on my workstation.
I don't care even a little bit about the closed nature of the drivers.
-B
Ash and Hickory, straight-grained and true, make excellent bludgeons, dandy for the cudgeling of vegetarians.
Have you ever wondered why so many food recipes are a trade secret?
It's because recipes can't be patented or copyrighted.
Free Martian Whores!
You DO realize that is what you sound like, yes?
You SOUND like a RANTING, unbalanced IDIOT.
"should" is a moral imperative. I think that while you're looking for a moral justification (open-source pre 2003) everyone else is looking for an effective development structure that meets the different goals of the groups involved to get a really effective result (open-source post-2003).
An alternative question you might ask is "What would be useful about Nvidia subscribing to the project's 'goals and visions'?"