Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
×
Linux Software

nVidia Unified Drivers Including Linux/FreeBSD 306

Screaming Lunatic writes "nVidia has decided to include Linux and FreeBSD in their Unified Driver Architecture and offer more tech support. Sounds like great news for Linux developers and users if Linux drivers are released at the same time as Windows drivers. (The NV30 emulation driver for Linux was made available about 3 months later than for Windows) The big push is probably from big studios that use Linux tools such as Film Gimp. More info here ." Added by Heunique: You might want to look here if you are using the latest development kernel.
This discussion has been archived. No new comments can be posted.

nVidia Unified Drivers Including Linux/FreeBSD

Comments Filter:
  • BeOS, anyone ? (Score:2, Interesting)

    by z80 ( 103328 )
    It would be fun if someone would have a look at the drivers and port them to BeOS. Geforce4 is yet to be supported.
    • Re:BeOS, anyone ? (Score:2, Insightful)

      by dinivin ( 444905 )
      It would be fun if someone would have a look at the drivers and port them to BeOS.

      It would be fun if nVidia open sourced the drivers so that you could look at them and port them to BeOS.

      Dinivin
      • Re:BeOS, anyone ? (Score:2, Insightful)

        by killmenow ( 184444 )
        It would be fun if nVidia open sourced the drivers so that you could look at them and port them to BeOS.
        No, it would be fun if nVidia shared the details on their cards so others could write drivers and port them to whatever freaking OS they wanted.

        I have an nVidia card and it is nice, but I think my next will be an ATI card because they have at least tried to be more cooperative with open source developers.

        Check this quote from an October 20, 1999 ATI press release:
        Recognizing the phenomenal growth and increasing popularity of Linux, ATI is committed to ensuring that the open source development community has access to technical development information on all its key components.
        Hello, nVidia? This is Open Source calling. We want your specs.
    • Re: (Score:3, Insightful)

      Comment removed based on user account deletion
      • Re:BeOS? No. (Score:3, Interesting)

        by UberLame ( 249268 )
        > I decided long ago that I did not want to become like those sad people that cling to long-dead platforms, constantly predicting their return and vocally defending their virtues. Let BeOS rest in peace.

        I like those people. Like the ones who do insane amounts of hacking to add upgrades and get more speed out of their Amigas. Or Acorns. Or Ataris. I think they do more than most groups to make the computing world interesting, and I applaud their efforts. And I wish I had the time and money to join them. I think the efforts of Apple II and C64 people are cool too, but have less desire to join them.
        • Re: (Score:3, Insightful)

          Comment removed based on user account deletion

          • When I think of all of the effort that they are expending and how it could benefit a modern, viable platform, it seems a terrible waste.

            It's not!

            Where do you think people cut their teeth learning how to program? That's right, those same dinosaur platforms!

      • Comment removed based on user account deletion
  • by Anonymous Coward on Tuesday December 17, 2002 @10:00AM (#4906675)
    That still means you get shafted if you use something different than i386/i64/amd64whatever. There's still no PPC support, and that sucks. Badly.
  • I think this will help them a lot. Linux runs quake much better, so if they officially support it they can start advertising the FPS of quake in linux. Users will compare it to quake in windows and think the card is even better.
    This will also help get developers(other than the gods at idsoft) to take linux a little more seriously.
  • by JSkills ( 69686 ) <jskills@goCOMMAofball.com minus punct> on Tuesday December 17, 2002 @10:05AM (#4906692) Homepage Journal
    ... to buy nVidia based graphics cards.

    They're arguably equal or (in most cases) superior to most other cards

    They've always supported Linux

    Installing a Geforce 4200i in my Mandrake box was a snap last year ...

    BTW - it was interesting to see the comment by Tim Sweeney of Epic Games (Unreal), who was applauding nVidia for their support of Linux. If we could only get all the gaming companies to pay as much attention to the Linux platform as the consoles or the PC, I could see the entire desktop shifting towards Linux next. Ok, well maybe I'm just dreaming ...

    • They're arguably equal or (in most cases) superior to most other cards

      You havent ever seen an ATi card in action, have you?
      Waaay superior 2D (picture quality) ...
      • Waaay superior 2D (picture quality) ...

        Wouldn't that be due to your monitor?

        • Wouldn't that be due to your monitor?

          That's like making the assumption your photographs are bad because the paper used.

          The developer, film, lens, camera, and photographer are all variables in that equation, also.

          While a good monitor will certainly enhance a computer image, it's not going to be able to magically fix a crappy signal sent to it. 2D quality on video cards is not standard. And, yes, nVidia does take some liberty with high-fidelty for the sake of performance.

          If you want an outstanding 2D image from a video card buy Matrox. Their 3D implementations aren't that hot, but their 10bit color cards are used almost exclusively for high quality imaging devices (think medical applications) because of the clarity and quality.
          • Actually, your both wrong...

            While ATI and Matrox are usually touted off as superior 2D quality, that is just a baseless assumption based on the most popular implementations of video cards you have experience with.

            there is nothing inherantly wrong with nVidia's 2D picture quality, or anything inherantly good about ATI's or Matrox's. The difference is that ATI and Matrox produce their own video boards, while nVidia only produces chips.

            The difference here is in the digital to analog converters used on the board, which are not part of the GPU supplied by nVidia. nVidia sells just the GPU, and the card manufacturer is responsible for buying the other components from other people. Most companies that manufacture nVidia video cards have strong competition, so skimp out on some parts. It shows when you try to run an nVidia card made by a cheap manufacturer at 1600x resolution.

            The same holds true for the "Powered by ATI" video card line, which is not produced by ATI, but 3rd party board manufacturers.

            of all 3 GPU companies, there is essentially no difference in the 2D quality in either chipset. It's just that ATI and Matrox have the ability to dictate exactly what is used external to the GPU, and so can guarantee that their cards will have the quality of 2d that they want it to have.

            If you buy quality nVidia boards from ASUS, or Visiontek (now defunct) you will be sure to have quality parts in them.
            • by Quarters ( 18322 ) on Tuesday December 17, 2002 @03:25PM (#4909329)
              of all 3 GPU companies, there is essentially no difference in the 2D quality in either chipset.

              Bzzzt, wrong.

              Both nVidia and ATi are using 8bit color modes for ~16million available colors (plus some alpha, some stencil, etc..). While the new ATi 9500 has a psuedo 10bit mode it is just doing internal calculations in 10bit on an 8bit input source.

              Matrox, on the other-hand, offers true 10bit color modes on their cards. This is something that sets their cards apart from the consumer level commodity devices that both nVidia and ATi make.

              While DirectX9/OpenGL 2.x call for 10bit color modes (e.g. colors specified as 0...1 instead of 0...256) There isn't an ATi or nVidia card out yet that truly does 10bit (GeForceFX, maybe).

              There are differences between 2D output of cards, don't fool yourself. Even beyond the 10bit/8bit issues there are color quantization choices, pixel blending/dithering choices, anti-aliasing implementations....that each manufacturer does differently. These different choices do equal different output of the same source material on different cards.

              And yes, both companies (ATi and nVidia)make cards, not just chipsets. 99.9999% of the nVidia cards you buy are identical. They're all based off of a reference engineering design that nVidia makes for each chipset release.
    • by the gnat ( 153162 ) on Tuesday December 17, 2002 @10:21AM (#4906771)
      Add to this:
      • Their GLX implementation offers features that are non-existent in Mesa.

      One piece of software I use almost daily essentially requires the NVidia driver if you use it on Linux, because of display lists. The difference in speed is simply ridiculous.
    • by NerdSlayer ( 300907 ) on Tuesday December 17, 2002 @11:26AM (#4907243) Homepage
      I dunno if I can support your login. It is Tuesday, and I think that's large corporation hating day.

      Down to Intel! And Microsoft! And NVidia!
  • Nice, but... (Score:2, Interesting)

    by Kryptoff ( 611007 )
    How about releasing the drivers as open source? Then you'll have a lot of support from the Linux community. (see previous thread about the linux kernel)
    • But the drivers are open source. They just aren't GPL. Which is good enough for me, since I can compile it against any customized kernel I'm using.

      If they were completely binary, you'd have to grab the drivers for a specific kernel for a specific distro (Hed Rat being the most popular). That's the argument against binary-only kernel modules.
      • Re:Nice, but... (Score:4, Insightful)

        by vrt3 ( 62368 ) on Tuesday December 17, 2002 @10:28AM (#4906812) Homepage
        They are not open source. They have a small kernel modules that comes as a source tarball, which you have to compile against your kernel. The driver itself is binary only and communicates to the kernel using that module.
        • Even that is good enough to completely avoid all problems associated with kernel modules that are completely binary.

          • Unfortunately, it's NOT good enough to allow you to properly debug applications that use that driver, or debug the driver itself.

            It's about as open source as Cheney's list of energy policy contributors.
      • Re:Nice, but... (Score:5, Insightful)

        by dinivin ( 444905 ) on Tuesday December 17, 2002 @10:37AM (#4906857)

        Do you even know what "open source" means? It does not mean having a HUGE closed binary driver (larger than most people's linux kernel) that links to the kernel using a tiny wrapper whose source code is available. Nor does it mean having a OpenGL library and GLX extension whose source is completely unavailable.

        Dinivin
    • How about releasing the drivers as open source? Then you'll have a lot of support from the Linux community. (see previous thread about the linux kernel)

      I don't really see this happening soon. Video cards, more than other types of hardware tend to include a fair amount of nifty technology in the drivers.

    • They can't (Score:5, Insightful)

      by Sycraft-fu ( 314770 ) on Tuesday December 17, 2002 @11:20AM (#4907191)
      Look, I know that many people think ideally everything should be open source, but it doesn't work that way. There are good reasons to want to keep the source closed on many thigns. In nVidia's case, one of teh main reasons is because they liscence certian thigns (like OpenGL technology for example) they they are contractually not allowed to release.

      Really, I think people ought to quit bitching. Their drivers are fast, stable, and support all the features of their hardware. This is what one would expect from a driver. If they keep it closed for contractual reasons or otherwise, that seems like a poor reason not to use the hardware.

      We aren't talking about something like Windows, that is attempting to keep something proprietary, the drivers act purely as an interface between the hardware and the higher level software.
    • Re:Nice, but... (Score:4, Insightful)

      by Fnkmaster ( 89084 ) on Tuesday December 17, 2002 @11:28AM (#4907267)
      Your point has been raised a million times on Slashdot. The counterpoint has as well. NVidia has technology licensing agreements with third parties that, at least according to NVidia, do not allow them to Open Source portions of their drivers without violation of the agreements. Others have argued that surely NVidia could renegotiate those agreements. It has been pointed out that sometimes licensing agreements are with entities that don't even exist in the same form anymore, and renegotiating them may not be feasible. Undoubtedly, it could be an expensive proposition for a company with lots of such licensing agreements.


      In the end, like any business decision, it's a marginal cost/marginal benefit based decision. What does NVidia get from Open Sourcing their drivers? Matrox has had the best support for Open Source over the years, open specs and the works. How far has that gotten them? Ummm... I am as much a proponent of Open Source as the next rabid slashbot, but the fact is for a hardware company concerned about giving out too many details of their hardware and intellectual property, that spends more time and money developing good drivers than other hardware companies do, their stance makes some sense. In a perfect world, we would recognize that a hardware company's business is selling hardware, and the driver software ain't part of their business, thus they should Open Source it. But the fact is they MIGHT give away proprietary information they don't want competitors to have if they did that. And that's more important to them than the small market represented by the most rabid Open Source zealots.


      Furthermore, many of the problems folks have had over the years with breaking NVidia drivers are directly attributable to the fucktard kernel devs who don't seem to have a concept of a stable ABI/API for kernel drivers. This is one area that Windows technically seems to shine over Linux. Kernel modules should work seemlessly across minor kernel versions. Not to encourage binary only modules, but to encourage ease of use and upgrading of Linux systems. If I upgrade Windows 2000 to Service Pack 18 or whatever, I don't have to go download new drivers. This is just silly. The contract between driverland and kernel land should be well-specified and stable, not "the driver can muck around with any kernel structures it fucking pleases".

      • Your point has been raised a million times on Slashdot. The counterpoint has as well. NVidia has technology licensing agreements with third parties that, at least according to NVidia, do not allow them to Open Source portions of their drivers without violation of the agreements.
        As an Nforce use, I have found this to be a little white lie. The audio is Intel810, the IDE is AMD. Doubtless, their Ethernet implemenation also is somebody elses. They have different entry points to distinguish themselves from the stock branded chips. Using binary modules for their NForce boards is an unnecessary headache.
    • It has been already said a trillion times, but I will repeat it for you yet again: They can't. There are many things in the driver code that NVIDIA doesn't own. Those things are patented and therefore cannot be released in open form to the general public. And puh-leese, dont start preaching that they should get rid of that "patented crap" of whatever you would like to call it.
  • by OmniVector ( 569062 ) <see my homepage> on Tuesday December 17, 2002 @10:08AM (#4906704) Homepage
    Id software can practically drive the industry sometimes, and without decent driver support for the platform Id would have a hard time putting out games like doom III or Quake III for the linux platform.
  • Is this news? (Score:2, Informative)

    by Aardpig ( 622459 )

    NVidia have been "supporting" Linux for a few years so far. Unfortunately, their drivers have been closed-source, binary-only -- a fact which has caused quite a bit of grief for kernel developers, since it makes it impossible to trace the cause for a kernel oops when using the NVidia drivers.

    I did a search through the article for the word "open". I found "OpenGL", but no "open source". So, IMHO, this news release is just PR bullshit (apart from the BSD bit, which may be new) -- there appears to be no move whatsoever for NVidia to open up their source.

    I wonder what implications the continuing close-source approach of NVidia will have, what with the upcoming abolition of binary-only modules in Linux kernel 2.6?

    • Closed source may be what the Linux kernel needs to keep things stable.

      Maybe if it weren't so easy to diagnose what errors mean for more products, the kernel folks would take a more conservative stance toward keeping the api's and constructs more consisten, particularly in "stable" kernels.
      • YOU ARE WRONG. Actually not really, I just like saying that.

        The "kernel folks" don't give a rats tail end about your video driver or you when your video driver fails, so closed source isn't going to change things that much.

        What I do wonder is why device drivers can't live in sandboxes or even user space. I know it would be difficult to design an efficient API, but it isn't impossible.

        Joe
        • What I do wonder is why device drivers can't live in sandboxes or even user space.
          Why, sure they could. In fact, they did - got any old copies of DOS lying around?

          But without the OS to regulate access to devices, there was no pratical way to arbitrate among multiple processes and users, especially if you care about security at all.

          • No, not my user space, but not in full control of kernel space either. (In Unix, there is only kernel space and user space.) I don't generally want drivers to have the ability to crash the OS. Some of them of course will be able to (if the IDE driver doesn't load the kernel correctly or fails to swap in part of the kernel correctly, then it will crash).

            User space would require a context switch that we would want to avoid, so I mentioned the sandbox.

            DOS had no protection, nor even multithreading/multiprocessing, so I'm a bit confused by your point.

            I want an OS where I can confidently install a video driver and know that it doesn't have access to trash my filesystem. I've heard NT3.5-- had video drivers like that.

            I can imagine that an easily readable description of the driver would include what resources the driver needs and the OS would only allow access to those resources. It couldn't be fullproof, but it could give a bit of confidence in binary drivers or even open source drivers that we don't want to audit first.

            Joe
      • Closed source may be what the Linux kernel needs to keep things stable.

        Troll, or just ignorant?

        First off, while I own two Nvidia cards, I won't be getting or recommending another. Lack of stability is the reason. The 3DFX card and open source drivers that I used to have were rock stable when rendering 3D. Every release of the Nvidia drivers have been unstable and have caused crashes that do not occur with the open nv drivers. These problems occur with others and are similar to those that occur under Windows. (Yes, I checked the AGP settings and other troublesome issues.)

        Second, it is not the resonsibility of one programmer to fix the code of another. In the case of the kernel developers, they can't fully debug each binary release of Nvidia's code on each system. That said, when code has a defect -- but is openly available -- it is often fixed.

        Bottom line: Most of the reasons for using open source are practical and have nothing to do with agendas, politics, philosophy, or views on business.

        • Why didn't you read my post?

          3dfx has been out of business for years now. Of course open code that old is so stable.

          The kernel crew has no responsibility to fix the code of others. They do have a responsibility to make their own code usable and consistant.

          If the kernel developers planned ahead and didn't arbitrarily make major changes to "stable" code at a whim, things like device drivers would be stabler and easier to write.
    • Re:Is this news? (Score:4, Informative)

      by OmniVector ( 569062 ) <see my homepage> on Tuesday December 17, 2002 @10:22AM (#4906780) Homepage
      Not that i'm trying to be an ass,

      but this is posted every time the topic comes up. NVidia can't release the drivers because of legal reasons. There are things in the code that they do not own, thus cannot release.

      Not to mention right now, the Nvidia cards win hands down on driver quality, which is a good advantage over ATI cards.
      • Not to mention right now, the Nvidia cards win hands down on driver quality, which is a good advantage over ATI cards.

        Hardly... Take my home computer, for example. A brand new i845 motherboard, with a single P4 processor. Using a GeForce3 and the drivers from nVidia, there is literally a two minute wait between the time I type 'startx' and the time X actually comes up. In the mean time, the screen flashes about three or four times, but the machine is completely unusable, even through ssh because the entire computer just stops functioning till X comes up.

        Compare that to the FireGL drivers from ATI with a Radeon 8500 on the same machine. X starts up in less than 3 seconds.

        Unfortunately, no one at nVidia can seem to tell me why this is happening. Until this is sorted out, and nVidia actually learns the meaning of "support", they've lost my business.

        Dinivin
        • I have a Geforce3 in a P2 machine, and it only takes me about 45 seconds to startx.

          I bought the Geforce3 because I previously had a TNT2 and they both use the same driver. I would have rather had a FireGL card, but most of all I needed something that would work in under an hour because I needed faster GL performance and had no spare time for extensive tinkering.

          I guess it is time to upgrade my nvidia drivers again. Unfortunately, it will be sometime before I can afford a FireGL card now.

    • Re:Is this news? (Score:2, Informative)

      by jdkincad ( 576359 )
      I wonder what implications the continuing close-source approach of NVidia will have, what with the upcoming abolition of binary-only modules in Linux kernel 2.6?

      The hernel developers aren't abolishing binary-only modules, they're just changing the way they interact with the kernel.
  • by Quarters ( 18322 ) on Tuesday December 17, 2002 @10:11AM (#4906718)
    The big push is probably from big studios that use Linux tools such as Film Gimp.

    Right, because workstations that use a 2D, time based, painting program need programmable pixel shaders, programmable vertex shaders, hardware transform and lighting, massive fill rate, AGP 8X transfer speeds, and astronomical triangle throughput.
    • Give this man a cookie.

      Maybe it's because of the use of acutal studio tools that were released for Linux in the last two years. Maya, Softimage|XSI, Shake (a compositor, but it uses GL heavily). Throw in renderman, which has been in Linux for years, but doesn't require video (it's just a renderer), and your whole CGI pipeline is running under linux. And with the NVidia drivers, very very well.

      Film Gimp? Give me a break. I like where it's heading, but it's not what pushes studios to linux.

      Rich
    • Programable vertex and pixel shaders and massive fill rates are potentially very usefull for paint programs.

      For instance, instead of the CPU rendering each tile of the image, the paint program could send all the tiles of all the layers to the graphics card and let the graphics card render the screen. The tiles would reside on the graphics card until they are changed, or they need to be removed for new tiles from scrolling or zooming the image. Well, so far this just would require AGP8x and massive file rates. But there are other tasks that could be implemented as shaders here. For instance, pixel shaders could handle the alpha blending rather than redrawing the scene for each layer. There are many tasks that could be pushed to the video card, but some of them would be hampered by the fact that while AGP 8x is fast to write to, reading back from it is about the same speed as PCI, which could be detrimental for many tasks.

      None of this is supported by the Gimp, or FilmGimp, or any other paint program that I know of for linux. I'm led to believe that some Irix programs do things like this though.

      BTW, I hear that the new NVidia cards did away with hardware T&L, and instead they do it using vertex shaders, with the GL driver making it work seemlessly. I can't really confirm this from anyplace officially, but it would make sense to reduce redundancy this way.
      • by Quarters ( 18322 ) on Tuesday December 17, 2002 @12:14PM (#4907680)
        BTW, I hear that the new NVidia cards did away with hardware T&L, and instead they do it using vertex shaders, with the GL driver making it work seemlessly. I can't really confirm this from anyplace officially, but it would make sense to reduce redundancy this way.

        Hardware T&L and programmable pixel/vertex shaders are not mutually exclusive.

        All pixel/Vertex shaders give us is the ability to move from a fixed-function pipline (mostly in lighting) to a programmable one.

        All nVidia cards before GF3 had fixed function lighting. You were given the lighting algorithms on the card and that's all you got. With programmable shaders, though, the lighting equations can be completely re-written by the devleopers. At around the GF4 timeframe they completely removed the old fixed-function pipeline transistors and just added pixel/vertex shader code that did the same equations. That is probably what you are refering to.
  • I suspect (Score:3, Insightful)

    by WindBourne ( 631190 ) on Tuesday December 17, 2002 @10:12AM (#4906720) Journal
    that the economy is Linux's best friend. While the recession continues(and grows deeper), companies will have to work harder to make their profits and grow their company. Like hollywood, these companies will lower their IT spending; this means Linux. This also will means that some smart companies will offer better service which will lead to improved sales which their competitors will notice. 4)Profit!!!
    • Note: This discussion is a diversion, not necesarily off-topic. :)

      While the recession continues(and grows deeper)

      A recession is a period of general economic decline; specifically, a decline in GDP for two or more consecutive quarters. We've actually had growth (in the USA anyway) for at least the last two quarters and, I believe, the last four.

      Anyway, you're correct that a weak economy is a good way to encourage companies to at least consider open-source software, especially when companies cannot afford to take the hits of security breaches and downtime that competitors are notorious for providing. That doesn't even take into account the licensing costs that are saved by using open-source software.
  • by 91degrees ( 207121 ) on Tuesday December 17, 2002 @10:20AM (#4906769) Journal
    A disproportionate number of Linux users have nVidia cards primarily because they have decent drivers which work for all kernels, and tend to install without too much hassle. I downgraded my Linux box to a TNT2 because of this.

    This has resulted in a large chunk of the market share going to nVidia, encouraging them to invest a little more in Linux. A sort of feedback loop.

    It may only be a niche, but it's another chunk of income for them. nVidia will sell chips to anyone if they can get more money back than they spend.

    I doubt filmgimp has as much of an impact. This is a smaller market than 3d enthusiasts with dual boot Linux systems.
  • Interesting News (Score:3, Interesting)

    by Alethes ( 533985 ) on Tuesday December 17, 2002 @10:25AM (#4906792)
    Considering this story about binary modules in the Linux 2.6 kernel [slashdot.org], it's either not a problem for proprietary software developers to have these restrictions placed on them, or they haven't gotten wind of the news (which is highly unlikely).
  • but.... (Score:2, Insightful)

    Installing nVidia drivers means changing open source operating system into open-closed source hybride. Keep in mind, that you have choice with other cards, like ATI or Matrox.
    I know most people answer "I am not interested with changing or even reading driver source", but they forget why Open Source is good. I will write just two reasons:
    • when something doesn't work - you can ask for help, open source projects (DRI, emu10k1, etc...) has mailing lists when you can submit bug raports, but help is not limited to drivers author - you can also find help on newsgroups, from people who has same hardware like you, open source help works better, because there is a lot of coders floating around on newsgroups, these coders can do nothing with closed source
    • spyware - yes, it's hard to believe there is some kind of spyware in device driver, but it is possible (in future!), and remember - open source Linux means no spyware at all, with closed source you can be never sure

    • and remember - open source Linux means no spyware at all

      You really are clueless, aren't you? I submit the recent trojan horses implanted in BitchX and tcpdump. It certainly is far easier to detect trojans in an open source project, but only a fool would believe that their open source nature makes them immune to trojan horses being implanted.

      Please stop spreading idiocy that only makes the problem worse.
  • by DCowern ( 182668 ) on Tuesday December 17, 2002 @10:27AM (#4906806) Homepage

    NVidia still hasn't realeased a set of drivers that work with the 2.5.x development kernel which, unfortunately, I must use day-to-day -- albeit on a non-production machine.

    I won't criticize NVidia too harshly for distributing binary-only drivers -- I understand their reasoning and I accept it. I only wish that since we can't have the source, they'd support us developers with beta drivers that work with the 2.5 series kernel. It'd be nice to have an idea of what and how things will work in kernel-next.

  • by caldaan ( 583572 ) on Tuesday December 17, 2002 @10:32AM (#4906830)
    Look, I agree in most cases it makes sense for stuff to be open source and have nice open drivers so that if there are bugs they can be addressed in a more timely fashion.

    NVIDIA's track record is already to make high quality drivers, now whenever they make a Windows driver the Linux driver will be right there with it because 95 percent of the code base is now shared.

    This is a brilliant move to hurt ATI were there problem is .....drivers... one ATI's drivers suck, they don't update them across all cards at the same time, and they surely don't update them across all supported platforms at the same time. Guess what NVIDIA now can.

    Now if NVIDIA were to release their code ATI wouldn't have to decompile it and scratch their heads trying to figure out exaclty how to compete they would have the freaking code.

    NVIDIA has a unified driver for all of their cards, for all of the platforms now so that you know that all of their cards will always work with the latst drivers, the operating systems they support and so on. Good luck finding that for ATI, and I for one don't think that NVIDIA should hand them the code on a platter just because people want everything in Linux to be open source. Sometimes vendor support is just as important.
  • by JeffVolc ( 89846 ) on Tuesday December 17, 2002 @10:34AM (#4906837)
    I have a Geforce2 Dual card which I have been trying to get seperate X screens on so I can play OpenGL games (i.e Terminus, UT2003, etc) on one screen and put XAWTV on the second to watch TV (i.e. Simpsons & Samurai Jack). Well, the new driver supports that...

    Here are the highlights of the new driver:

    Linux Display Driver
    Linux Graphics Driver Download

    Version: 1.0-4191
    Operating System: Linux IA32
    Release Date: December 11, 2002

    Release Highlights:
    * OpenGL 1.4 with CineFX architecture support
    * Support for AGP 8x and nForce2 IGP
    * Support for index overlays on Quadro4 to support legacy applications
    * Support for separate X screens on nView enabled GPUs
    * GLX 1.3 support

    Yup, seperate X screens now with the dualhead cards. Hopefully I can put this to the test in the next few days.

    BTW, Don't try this on Windows kids....

    Jeff
  • by martinde ( 137088 ) on Tuesday December 17, 2002 @10:41AM (#4906879) Homepage
    The drivers that the Weather Channel was paying for - did anything ever come of that? This announcment from Nvidia is definitely great, and I have no doubt it's the most they can do right now, but...

    It would be much better to have open source drivers available. You'd get more people looking at things like security and performance issues, and then we could have support for architectures other than x86. (Which is probably something Nvidia themselves isn't going to bother with.)
    • by pawsa ( 92107 ) on Tuesday December 17, 2002 @11:46AM (#4907385) Homepage
      The Weather Channel-sponsored drivers are doing great, I use them since August and now they are basically ready - see http://dri.sf.net/. The only disadvantage with respect to closed source drivers was inability to use some patented techniques (in particular texture compression), since the patent owner (S3?) did not grant (yet) the rights to use them in the driver. Otherwise, there are much more convienent to use than Nvidia drivers I have been forced to use on my other box. And more stable, too.
  • by Keck ( 7446 ) on Tuesday December 17, 2002 @10:45AM (#4906916) Homepage
    Companies that are bottom line driven (accept it, you won't change capitalism over night) can't usually afford to jump in head first. This goes double for Nvidia, ATI, etc. Their driver source is like a blueprint of the important parts of their hardware.

    I prefer open source, but to say there is no place in the world for closed source modules, applications, whatever, is too extreme IMHO.

    For me the dividing line has always been commodity vs non commodity. Example: Of COURSE the OS, office software, web browsers, MUA's, MTA's, etc should be open, they are commodities. Specialized programs like AutoCAD, Drivers for up-to-the-minute video cards, and various other areas do NOT lend themselves to the open source model, and I don't believe they have to.

    So right now the devil's choice is,

    a) fast nvidia drivers for linux/bsd that get released with the windows drivers, which is 2 steps ahead of where we were in July, or

    b) only a community supported driver, created by reverse engineering the chipset or windows drivers, released months (and years) after the windows versions.

    It's not a perfect world, we have to change it in small steps. Your idealism is duly noted. Give nvidia credit for moving in the right direction, maybe at some point it *will* make sense to go GPL for them..
    • For me the dividing line has always been commodity vs non commodity.


      Ah! Commodity for whom? Photoshop/FilmGimp isn't a commodity for most people, but it is for film studios that do touch-up work. So, five studios developed FilmGimp independantly, discovered they were all doing it, and pooled their efforts.

      For resource mining companies (oil, gold, diamonds, dinosaur bones), applications like MagicEarth will probably be a commodity one day. Right now, MagicEarth appears to be the nVidia of tera visualization, so there's no pressure.

      Drivers are funny. The tend to have hardware specific portions. I think nVidia has internally found what would turn most drivers open source. Nintey-five percent of all their video products share the same pool of code. The rest are model-specific registers.

      Right now, nVidia's got the advantage by sharing all that code among their products. They won't open source as long as that is an advantage.

      Probably the most tedius and possibly the grandest hack Open/Free Source could produce would be a Grand Unified Driver Architecture for Video, Audio, and NIC. This would reproduce nVidia's internal advantage. Hardware companies could scrap 9/10 of their internal driver development, because the GUDA would have covered it. ATI would be given a large advantage if they could plug their entire video line-up into the GUDA.

      Of course, IANAP.
  • I'm not so sure this is a good thing. My multimonitor/multicard setup works fine under the latest Linux drivers, but doesn't work under the latest Win2K drivers. I have to keep using 30.82 under Win2K. I'd report the problem, but they seem to offer no support for the Detonator drivers at all. Any nvidia people out there?
  • I see a lot of whining and complaining that these aren't open source drivers, and that they don't really SUPPORT linux if its' not open source, but this is exactly the attitude that makes companies afraid of linux. Everyone feels that if they are going to put anything on linux that they have to open it up. Yes, it would be nice, but let me tell you what I think is even nicer. Having the frigging drivers in the first place. You know what's nicer than that? Having them release the drivers at the same time as their other drivers. If this doesn't seem like a nice thing to you, then just think of it as a step in the right direction and offer a little support.
  • by ChaoticCoyote ( 195677 ) on Tuesday December 17, 2002 @11:11AM (#4907120) Homepage

    Several people seem concerned about nVidia's drivers and the forthcoming 2.6 kernel. I can't say much for tomorrow, but today, I have the latest (4191) nVidia drivers working just fine with the most recent development kernel.

    To make it all work, the drivers need a minor makefile patch and updated modutils, but otherwise work just fine. You can obtain the required files from:

    Unofficial nVidia driver patch [minion.de]
    Updated modutils [kernel.org]

    Those did the tirkc for me. Your mileage may vary.

  • by GweeDo ( 127172 ) on Tuesday December 17, 2002 @11:39AM (#4907343) Homepage
    Okay. I love linux. I love having the source available to many different partsof my OS and of many of my applications. BUT I AM ALSO A REALIST. We will never get to a point where everything is Open Source. People do need to protect the their IP some times. People have a perfect right to close source their poject if that is what they decided to do. I am very happy that Nvidia is supporting Linux. I love their drivers. Quake 3 Arena and UT2K3 run really fast. All my little opengl apps I using to learn runs flawlessly. Continue the great work Nvidia...close source your drive if you want or if you are required to by other agreements. Please stop whinning everyone, be glad we are getting good solid support!
  • The page claims that the new NVidia drivers support power management, as in suspend/resume on laptops.. That doesnt seem to be the case.. Have I missed something?
  • Film Gimp? No... (Score:3, Interesting)

    by I Am The Owl ( 531076 ) on Tuesday December 17, 2002 @12:59PM (#4908080) Homepage Journal
    I think the big effects studios are moving to Linux, alright, but not for the free software. Most studios use high-end commercial compositing, rendering and modelling software (to name a few catagories) to get the job done (or they write it in house, like Digital Domain's Nuke). Most of them don't work with stuff like Film Gimp because it simply doesn't have the features that they are looking for right now.

    Anyway, a large-scale replacement of artists' desktops is taking place, moving from expensive SGI hardware to faster, cheaper x86 hardware running Linux. This is why it is so important that high-quality drivers are available for high-end graphics cards in Linux, and, unfortuantely, the best for nVidia comes from nVidia in binary format. Monkeying around wtih the kernel's binary interface with each incremental release is not going to make nVidia's job any easier.

Our OS who art in CPU, UNIX be thy name. Thy programs run, thy syscalls done, In kernel as it is in user!

Working...