Forgot your password?
typodupeerror
This discussion has been archived. No new comments can be posted.

AMD Puts Out Radeon HD 6000 Open-Source Driver

Comments Filter:
  • For all the Slashdot posters who keep begging for Linux support and talking about how big companies constantly ignore you, this is your chance.

    Buy AMD. Be vocal that the reason you're buying an AMD video card is because of their driver. Vote with your wallet.

    (On the CPU front, you can make an equal case for Intel supporting open source).

    • by CynicTheHedgehog (261139) on Friday January 07, 2011 @12:38PM (#34792286) Homepage

      For what it's worth, my AMD CPU and ASUS motherboard (quad core AMD 870 chipset) work just fine in Linux, as does my AMD/ATI video card. As did my previous AMD CPU/motherboard. I have yet to be disappointed by them.

      • by OFnow (1098151)

        With Ubuntu 9.10 , Intel Core Duo, ASUS board and an ATI card (new in late 2009) the screen had ugly little dots appearing and disappearing.
        And red thin streaks appearing and disappearing. The proprietary driver did not work at all (I forget exactly what it did).

        Switched to an nVidia card and no more weird artifacts. I sure hope ATI/AMD has fixed all that with the new open source drivers,
        but it makes no sense to me to switch back just in case ATI works now!

        • That was a vsync issue that was fixed a while ago and released as an update 10.04 (6.13.1-12 I think). I never experienced it myself. I think the drivers are on 6.13.2 now which is scheduled for Natty, but you can get it from the PPAs I think. 6.13.1 has been really solid for me.

    • Re: (Score:2, Troll)

      by drinkypoo (153816)

      For all the Slashdot posters who keep begging for Linux support and talking about how big companies constantly ignore you, this is your chance..

      What, my chance to buy an nVidia card so I can have a driver that works?

      • I have always stuck to NVidia graphics cards, because of the great performance in Linux.
        I have put up with the annoyance of having to rebuild my NVidia driver every time I have upgraded the kernel.
        I have put up with the annoyance of having to go to lengths to disable the nouveau driver.

        Now, however, I have a new system; Intel i7 8 cores, with a (bit of a wimpy) nVidia GF 8400GS. (Hey, it's a dev server, not a games machine.)
        After installing Fedora 14 and doing the usual hassle of removing nouveau and insta

        • by drinkypoo (153816) <martin.espinoza@gmail.com> on Friday January 07, 2011 @01:04PM (#34792784) Homepage Journal

          Maybe you should stop running the Pre-Alpha version of RHEL and move up to a distribution intended for end users if you don't want to solve problems and do Alpha testing for Redhat.

          • by MSG (12810)

            Is the problem Fedora or nvidia here? The Linux developers prefer drivers to be maintained in the kernel tree. Nvidia chooses not to do that. Neither the Linux nor Fedora developers can fix that decision.

            By your logic, the user should have chosen Windows, as it's "intended for end users" and he won't have to solve problems or do testing for some Linux distribution. Get real.

            • by X0563511 (793323)

              I'm having the same problems on my desktop. Quad 3.0ghz, 4gb RAM, and an NV 460 GTX. I don't think it's the hardware, as it worked fine on slackware before I got stupid and stuck Fedora on it :)

            • Kernel devs (seem to) desire open-source drivers in the kernel and no others. As the nVidia driver is a binary blob, it cannot be included in the kernel. This has nothing to do with ggp running what amounts to an experimental operating system and then posting performance issues as if it affected all versions of that OS.

              There are in fact Linux distributions intended for end users. There are also pre-alpha versions of Windows (good luck getting them), and I would bet large sums that similar performan

          • What should they use? Ubuntu is just as bleeding edge, and they would have likely run into similar issues on Debian or openSUSE with having to manually reinstall the driver with each new kernel release.

            Some of the issues could be caused by Fedora running a bleeding edge version of X, but in general Fedora isn't worlds different from running Ubuntu or openSUSE as far as stability goes.

            • Debian now has dkms, so you can just install nvidia-kernel-dkms and nvidia-glx and let the system handle rebuilding the module for you.

              • by tyrione (134248)

                Debian now has dkms, so you can just install nvidia-kernel-dkms and nvidia-glx and let the system handle rebuilding the module for you.

                And it's not slow. My 8600GTS has steadily run KDE 4.5.x smoothly on Debian Sid. Having 3.3 OpenGL and testing OpenCL 1.1 while building Bullet and Blender Trunk shows very smooth and responsive results.

                • by X0563511 (793323)

                  It's either something with Fedora itself, or GNOME. I have the same Fedora version with my 460 GTX, and have the same problem. I did not have any issues with KDE 4 on Slackware 13.1.

            • It is a DEV server, so perhaps the entire point is to code against an alpha-version of what will eventually become RHEL.

        • I have put up with the annoyance of having to rebuild my NVidia driver every time I have upgraded the kernel.

          Why, when you can just use the rpmfusion version, do a "sudo yum install akmod-nvidia", and never have to do that again.

          I have put up with the annoyance of having to go to lengths to disable the nouveau driver.

          Editing grub.conf and adding "rdblacklist=nouveau" to the end of the default boot line is great lengths?

          After installing Fedora 14 and doing the usual hassle of removing nouveau and installing nVidia, Gnome was REALLY slow. Rolling over menus could take up to .5 s.

          I've seen that reported, though not seen it myself, most likely it's your configuration settings. Try running nvidia settings and having it generate a xorg.conf for you.

          • Where were you when I was installing? I could'a used your help!
            Yeah, should used the rpmfusion for something...

            > Editing grub.conf and adding "rdblacklist=nouveau" to the end of the default boot line is great lengths?

            You forgot:
            - Ripping hair out when adding nouveav to modprobe.d/blacklist.conf (as suggested by nv installer) doesn't work
            - Searching internet for an hour to figure out that you have to modify the grub configuration

            I did try the nv config program, no luck :-(
            But hey, I'm not doing any 3D, a

            • Where were you when I was installing? I could'a used your help!

              I'm actually a newbie to Fedora on x86, 12 was my first, before that I ran YDL on my PS3, and before that Sony's wacky Kondara-ized Red Hat 6 on the PS2! So before I first installed Fedora 12 on a cheap Fry refurb I bought to copy over my /home direcotries before I updated my PS3's firmware into non-OtherOS ability, I googled for guides that would tell you what you REALLY needed to do to make your Fedora system usable for end-user type stuff. You know, Flash, 3D, MP3 support...I found this guy's site:

              http [mjmwired.net]

          • by X0563511 (793323)

            I've seen that reported, though not seen it myself, most likely it's your configuration settings. Try running nvidia settings and having it generate a xorg.conf for you.

            Same problem here, whether I use no config (autodetect), use/modify an Xorg generated config, or use/modify an nvidia-settings generated config. I've climbed that mountain and gave up trying.

        • by Anonymous Coward

          They don't make an 8 core i7. Hyperthreading does not magically make new cores.

          • by Anonymous Coward

            But it does make it look like additional cores to a newb.

        • by erroneus (253617)

          In fedora 14, whatever was slowing down Gnome with NVidia graphics has been resolved over several updates and I am not even sure which updates corrected the problem.

          However, you still have much to complain about with Linux and nVidia. Their hybrid graphics implementation ("optimus") is currently not workable/usable under Linux and nVidia will simply not share the information needed to make it work. There have also been an increasing number of cases where the proprietary Linux driver will not work with nVi

          • by S.O.B. (136083)

            In fedora 14, whatever was slowing down Gnome with NVidia graphics has been resolved over several updates and I am not even sure which updates corrected the problem.

            I believe they backed out a change in the Cairo 2D libraries that was causing the problem.

      • For all the Slashdot posters who keep begging for Linux support and talking about how big companies constantly ignore you, this is your chance..

        What, my chance to buy an nVidia card so I can have a driver that works?

        Nice troll. I regularly develop on both Nvidia and Radeon 3D hardware of various kinds, using both Open Source and binary drivers. I hit more bugs with NVidia. For example, black screen on text console, mipmap loading failure if not done in a specific order, random garbage on screen with window resize, etc.

        One thing I can say is, OpenGL performance is really good with proprietary drivers from both companies, thanks for that. But on balance I prefer to work with Radeon, it gives me less trouble, and I li

          • 1) OpenGL works well.
          • 2) Both ATI & Nvidia have been pretty good with Linux support. Perhaps not all options are supported, but its been generally better than I expected.
          • 3) I'd be curious to know what the support will be like (if it will be any different) when Ubuntu completes its switch to Unity/Wayland.
            • 3) I'd be curious to know what the support will be like (if it will be any different) when Ubuntu completes its switch to Unity/Wayland.

            My guess is, no effect on OpenGL per se. The migration to Gallium drivers is the one to watch.

      • by Carewolf (581105) on Friday January 07, 2011 @02:00PM (#34793730) Homepage

        NVidia's linux drivers have been incredibly buggy for years now. I recently changed the open source nouveau drivers which easily solved all performance problems I had. NVidia is both poorly implemented, has poor support for moderen X11 extensions (XRandr), the recent 26x series also has VERY serious memory-leaks which slowly brings your computer to a crawl.

        Btw. I am not stating this as a user, I am stating this as a develop who has been working on getting KDE to work better on NVidia hardware.

    • by Hatta (162192)

      I tried this. I bought an AMD card (4550 iirc) for an HTPC I was building. I'd be happy to have 2d acceleration working on it. Haven't been able to get it to work with any X driver besides Vesa. Drivers install, but X won't run. Running the util that generates an Xorg.conf just segfaults. This is with any of the drivers from Mepis or direct from AMD, open source or commercial.

      I gave up. Maybe I'll try again this weekend, but I'm leaning towards just buying an nVidia card. The only time I've had nVid

      • Re: (Score:3, Informative)

        by Anonymous Coward

        Only reason you'd get a segfault is because fglrx is installed. Remove it or ati-drivers (it's called one or the other) and you should be able to configure with the regular xf86-video-ati driver, which should at the very least give you smooth 2d if drm isn't available, and 3d/kms if it is (and the kernel is later than 2.6.36 or so.)

      • by Kjella (173770)

        To be honest, then I'd try something other than Mepis because then they're doing it wrong. Replace "vesa" with "radeon" and it should boot right up, if it doesn't they've horribly mangled kernel/drm/xserver/driver version somehow. That'll be the open source drivers as the time of release, if you want the latest I don't know on mepis. On Ubuntu you'd just add the xorg-edgers ppa and dist-upgrade. Of course, then you're really on the bleeding edge which is not for everyone.

      • by Menkhaf (627996)

        With more recent versions of Xorg, its probing works great and in many cases eliminates the need for an Xorg.conf.
        Granted, I haven't messed a lot around with Xorg lately, but when I last did a few months ago, it worked great without an Xorg.conf. If you haven't tried yet, give it a go!

    • by vadim_t (324782)

      Will do. All my cards have been nvidia so far. Will be ATI from now on.

    • Nvidia has had excellent Linux drivers for years, open source or not they work. I avoid ATI video as much as possible and it'll take more than token of support on AMDs part to change that now. They may be geniune and may win me over, but I'm not jumping on this bandwagon just yet. I'll keep voting with my wallet in favor of Nvidia for now.
      • by PitaBred (632671)

        Token? They've been working on open source drivers for the GPUs ever since AMD bought ATI. What more do you want? They can't just dump their current drivers into GPL due to patent and copyright issues, so it takes time to redevelop them and do it the Linux way.

        ATI is the ONLY way to go if you want decent performance and open-source drivers.

    • by Terrasque (796014)

      Already done. My current ATI card was a direct response to AMD starting to open source the drivers.

      And this from a customer that have bought Nvidia exclusively since the first GeForce card.

    • by jiteo (964572)
      On the CPU front, I'm buying AMD as well because Intel caved and gave Hollywood studios their DRM. I'd post the link to the earlier /. post, but /.'s textbox is broken in Chromium on Linux.
    • Sorry, but ATI driver is still PISS POOR relative to nVidia. It seems that everytime I install ANY distribution (... well, Fedora, Ubuntu or SUSE,) nVidia Just Works (tm)(r) but with ATI I have to download some weird thing from ATI, and the stuff (almost) never work right.

      You can say what you want about having an ideologically "open driver" but frankly most end users want the stuff to Just Works (tm)(r). My own experience is that ATI is still VERY FAR from Just Works (tm)(r).

      • Sorry, but ATI driver is still PISS POOR relative to nVidia. It seems that everytime I install ANY distribution (... well, Fedora, Ubuntu or SUSE,) nVidia Just Works (tm)(r) but with ATI I have to download some weird thing from ATI, and the stuff (almost) never work right.

        Simply not true for current offerings. Both Radeon and NVidia proprietary drivers + install scripts are a pain, and both are better than they used to be. Both offerings are fast as hell, but if you want hassle free kernel upgrades, nothing comes close to the open source Radeon drivers.

    • I am going to put my money where my mouth is. I've done it before.

      Once upon a time I had nVidia and ATI hardware on my computers. I could compare how well each company provided support for Linux. In neither case was their support stellar but nVidia did more than ATI. So when came time to buy a new laptop, no machine with ATI hardware made it onto the short list. They were disqualified from the start.

      In recent years, it seems that AMD is supporting Linux better than nVidia does. So it is likely that when

    • by asdf7890 (1518587)
      I base my decision for not buying Intel chips partly on things like http://hardware.slashdot.org/story/08/01/09/1352214/Negroponte-vs-Intel [slashdot.org] and http://news.slashdot.org/story/08/01/12/1424209/Intel-Employee-Caught-Running-OLPC-News-Site [slashdot.org] - the only Intel chip I personally run is the Atom in my netbook (as there was no CPU competition in that arena at the time).
    • Except what I need to use in my upcoming machine is a Cayman (or Antilles - cards coming soon) GPU and, while Cayman cards have been out for a while now, ATI apparently haven't provided support for Cayman in their open source Linux driver. And with the recent departures from the ATI Linux team it may be a while before that happens.
    • I did - I picked up a 5770 this time around. It's my first ATI (AMD) card ever. The other ones have been Nvidia and I think there was a 3dfx in there somewhere. I suppose I should send them a letter letting them know why I switched as well.

    • lol, i can't do that now even if i try

      i already replaced all my NVidia cards and switched to ATI when they released documentation. and every computer i bought after had amd card.

      and there is no place where i could brag with replacing 8 NVidia cards because of their linux support

  • by rrhal (88665)
    So when does Myth get 3D Blue Ray Support?
    • by jmorris42 (1458) *

      I'd settle for DVD support, I ain't greedy.

      Seriously, every once in a while I pop in a DVD, just for giggles. Very few play all the way through without error. Not new stuff mind you, OLD crap fails. National Lampoon's Christmas Vacation crashes the player a few chapters in. An ancient print, not a fresh copy.

      So I don't expect BluRay to work before it becomes obsolete.

      I do love my MythTV though, still trying to get the g-ddamned cable company to tell me how many channels I would actually get when the HD

  • by 0100010001010011 (652467) on Friday January 07, 2011 @12:35PM (#34792218)

    I didn't see any mention of VAAPI [wikipedia.org] or XvBA [wikipedia.org] Acceleration for playing media? How about OpenCL [wikipedia.org] support?

    Granted the HD 6000 [amd.com] looks more like a gamers card than something you'd stick in a home theater pc, but I'd think that OpenCL support would interest quite a few people doing massive number crunching. Especially since there's even PyOpenCL [python.org] available.

    • by Evanisincontrol (830057) on Friday January 07, 2011 @12:42PM (#34792364)
      Phoronix released a followup article [phoronix.com] today. A driver supports X video acceleration, but no VAAPI yet. Like you said, these cards seem to be aimed at gamers, but with the prevalence of HD video content these days, I'd be surprised if VAAPI wasn't a high priority for this driver.
    • by drinkypoo (153816)

      Statistically nobody cares about OpenCL, but the lack of video acceleration (if there is indeed a lack) is a serious oversight.

      • Fast, cheap number crunching? Isn't that why the Army bought a ton of PS3s to utilize the Cell cpu? Amazon just released new EC3 process that gives you 2 Nvidia CPUs for number crunching, and they're not cheap either.

        OpenCL is the "Open" CUDA solution. I believe CUDA actually supports OpenCL at this time too.

        Given a choice between "100% Open Sourced" Non useful driver like this and NVidia's "Free as in Beer" driver that has been kicking ass for a long time in the XBMC and for people that do use CUDA, I'll t

        • by drinkypoo (153816)

          Given a choice between "100% Open Sourced" Non useful driver like this and NVidia's "Free as in Beer" driver that has been kicking ass for a long time in the XBMC and for people that do use CUDA, I'll take Nvidia.

          It doesn't matter if you use CUDA or not, if you want your video to work right on Linux you use the last generation's nVidia card. Often nVidia will lag on hardware support, they lagged on my GT 240 for literally months. (I bought it for low power consumption... And it's more than enough card for me these days since I don't play the latest, greatest, masturbate-est games.) But AMD will lag even longer and then when they claim hardware is supported it still fails. I've been through this too many times to go

      • by Kjella (173770) on Friday January 07, 2011 @01:20PM (#34793070) Homepage

        It's not an oversight, blame DRM. Everything about their video decoding block (called UVD) is heavily coupled with the protected video path they have to provide for BluRays and other protected content. That means you have to have lawyers and tech people going over everything with a fine tooth comb and how much easier it would make it to reverse engineer the missing bits. It doesn't matter if AACS is broken and BluRay rips are everywhere because the contracts are still valid and the terms and penalties are as nasty as they get. AMD has said they will try to get changes to make it more open source friendly in the pipeline but new designs are started 3-4 years before release and it's probably not on their top 10 must do changes.

        That said, multi-threaded H.264 decoding has improved very much in software and I have no problems decoding 1080p video with that on my desktop CPU, it probably hurts a bit in power usage but at least you *can* do without. Hardware acceleration is more important for laptops and battery life, AMD is working on it but this is a very hard problem and they need most of the resources getting support for new architectures like the HD6000. This is not like much other software, hardware moves fast and in close sync with the closed source drivers. If the open source developers don't keep up, there won't be any support at all. P.S. DRM is also one of the reasons you can't share more of the "fundamentals" with the closed source driver. That would make it too easy to decompile the Windows/Mac driver to track and grab the protected content in transit. It really is a big hindrance to open source drivers.

    • by MBGMorden (803437)

      That's a major dissappointment. It's been my experience that in Linux, you simply WILL NOT get smooth video that is free of tearing without employing one of the GPU-assisted rendering methods. Me personally I bought a recent (but still desktop oriented) Nvidia card for my Linux system just for VDPAU support. That was one of the main things that made Linux finally usable to me as a "daily driver".

      • by bfree (113420)
        I've a 785g based motherboard with an energy efficient X2 5050e, running the stock Debian sid free radeon xserver (though I do have the non-free firmware) and have had no problems (or artifacts) playing back anything so far (up to and including 1080p h264). Maybe it's taking up more cpu load then it might if I used fglrx (or the non-free nvidia driver with a suitable card) but on the other hand I don;t have to worry about the non-free kernel-module conflicting/failing with updates to the kernel or x.org.
        • non-free firmware, or non-free driver... if it's the non-free driver, that is kind of the point... AMD is releasing partially usable open-source drivers, but if they lack full support for accelerated video, or 3d gaming, then it's kind of a half-hearted step. I've found that nvidia parts in linux tend to work with less issues (though closed source drivers), I bought AMD in last year's upgrades... I got an ATI HD 5770, and my son has a 5870 (from last january)... I still don't have reliable OpenCL encoding,
          • by Kjella (173770)

            AMD is releasing partially usable open-source drivers, but if they lack full support for accelerated video, or 3d gaming, then it's kind of a half-hearted step.

            AMD has never claimed they would release open source drivers of any given quality at all really. What they did promise was to give enough specifications so the community could make an open source driver, generally play nice with them and support them and contribute a little manpower themselves towards that. Gradually that has shifted towards a wish to make the open source drivers their "official" legacy drivers for cards they've dropped frglx support on, so that drivers for old hardware keeps up with kernel

            • If you want a driver that's all AMDs work it's called frglx
              Last time I worked with FGLRX it had the glaring fault of crashing the system to the point that no further interaction was possible if the kernel module was missing. Nvidia's driver OTOH just gave an error message

              The impression I got last time I dealt with ATI graphics on linux was that the OSS drivers were unfinished and the propietry ones were more of a PITA to deal with than the nvidia propietry ones. Has that changed?

              • by Kjella (173770)

                The proprietary drivers certainly become more even. ATI had horrible Linux support when AMD bought them and things didn't turn on a dime but they've improved very much. I haven't really used them that much though. I've been trying very hard to use the open drivers since those were the main reason I got it (for performance I mostly boot into Windows and game anyway). How the open source drivers are depend a bit on what generation card you have, I'd say the older the better it's been. Personally I've been on

          • by bfree (113420)

            non-free firmware, or non-free driver

            You could just have re-read what I said ... non-free FIRMWARE, I sure as hell won't let the non-free "drivers" touch my systems.

            • I did read what you said, but wanted to clarify, as many people don't even consider the proprietary nature of firmware.
        • by MBGMorden (803437)

          I've tried the unaccelerated video playback on a myriad of configurations. Different distros - Slackware, Gentoo, Ubuntu, Arch, and others. Different CPU's - a 1.8Ghz Intel Core 2 Duo - a 2.5Ghz AMD Phenom quad core, an AMD Sempron 3400, a Celeron 2.8Ghz, and others. Multiple video cards from both AMD, Intel, and Nvidia as well. Not just swapping components either - I've tried completely different machines (such as my work machine).

          Every, single, one, exhibited tearing of the video without GPU accelerat

          • by bfree (113420)
            I saw the tearing, I haven't seen it since the upgrade which brought in the fixes which specifically said they repaired the tearing. Just for fun before I posted I went and loaded up a 1080p h264 video file and went to a very aggressive section of it just to see if I could spot any visual glitches and I didn't. I detest digital artifacts in video and regard myself as overly sensitive to them, but I haven't noticed anything not in the source footage since said tear fixing upgrade.
    • by wuzzeb (216420)

      Phoronix released a follow up [phoronix.com] article that looks a little closer. There is XVideo support but no VA-API, XvBA, or VDPAU. But this article [phoronix.com] says that there are plans for XvBA and VA-API for the ati drivers, and work is progressing on VA-API support for gallium.

      For OpenCL, the classic mesa drivers have no support and no support is planned. But gallium has some support for OpenCL (see mesa/clover), and just today the gallium support for HD 6000 series has been released. So now the HD 6000 has both classic m

    • by CynicTheHedgehog (261139) on Friday January 07, 2011 @12:48PM (#34792468) Homepage

      I think the issue with the open source radeon drivers isn't that the developers don't know how to implement stuff, or that it's a closed API or whatever. The problem is one of manpower. Reading the Phoronix forums it looks like there's maybe 2 guys that do the majority of the commits (for the really hard stuff anyways) and they just have too many things to do. I recall one post explaining that some kind of compiler was needed, and so one of the developers stubbed out a very simple one to be able to move forward on something else, with the intention of coming back and replacing the stubbed out compiler later, when the more critical issues were addressed. The compiler was a key component to all sorts of 3D stuff and was a critical factor in performance, but functionality and performance take a back seat to compatibility and stability.

      My hope is that with the release of these open source drivers a lot of that boilerplate stuff will come with it, so that the community can truly focus on implementing newer APIs and such, although I don't know enough about video driver development to know whether that's necessarily the case.

      • I think the issue with the open source radeon drivers isn't that the developers don't know how to implement stuff, or that it's a closed API or whatever. The problem is one of manpower.

        I am sure that's it. However it's wonderful that I'm able to develop complex OpenGL apps with a reasonable expectation that things will work. Things that don't work: no buffer objects yet; no anti-aliased lines; mipmap filtering seems to always be "nearest"; and that's about it so far, not bad. Then there is the fact that triangle rate is only about 1/4 of the proprietary driver. But that's still a lot.

    • HD 6000 is a whole series of cards. The fanless version [downloadatoz.com] hasn't been released yet.

    • Very good point. I'm currently running my HTPC on an nVidia Ion based Acer nettop, but as soon as someone supports all the important video decode features in a fully open-source driver I'll be building myself a new machine. Dealing with the binary drivers, particularly for HDMI audio, is a pain in the ass.

    • Gallium3D architecture can easily bring opensource OpenCL support to modern GPUs.

      It's unlike classical Mesa or closed source, where each platform gets its own full fledged drivers with its own set of supported APIs.

      Basically, Gallium3D is a whole infrastructure, where, on one side you have hardware driver which just expose the basic functionality and capabilities that modern hardware feature, in a generic way. At the other side, you have state trackers, modules which listen for peculiar high-level api (for

    • by jonwil (467024)

      AMD have said that the open source drivers are unlikely to support the dedicated video decoding silicon due to DRM issues but that work to decode video using shader programs was being carried out.

  • by NtwoO (517588) on Friday January 07, 2011 @01:07PM (#34792848) Homepage
    In the past two years I've migrated from 10+ loyal Linux NVidia years to ATI. The ATI closed source drivers were reasonable whilst the NVidia drivers showed a slide in performance and stability (on my system in any case). Since September last year I've migrated my machines to the open source 3d drivers and what a beaut! My MythTV frontend with ATI onboard is impeccable. It'll require much to convince me to change away from ATI/AMD if they keep this kind of support available.
    • by Anonymous Coward on Friday January 07, 2011 @01:19PM (#34793050)

      Come back when we have hardware acceleration. Linux video playback needs CPU horse power. Even the iphone/ipod touch handles h.264 video better than linux. Whether we like it or not, that's the defacto codec these days and decoding is built into just about every device and video card. And yet, linux still cannot use them. Windows had it in '06.

      • by morgauxo (974071)
        That's a valid point. NVidia seems to have some support for that in the closed drivers now.
      • by Yfrwlf (998822)
        Valid point, but one I'm hoping may be ignored if much of the public ends up transitioning to WebM instead.
    • by PitaBred (632671)

      Does it have VA-API support for the 3000/4000 series? That's the main thing that's making me keep my Nvidia 220 in the machine instead of using the integrated GPU. I want to watch my HD videos, but it pretty much needs GPU acceleration with the relatively slow machine that's running it.

    • by morgauxo (974071)
      I had a similar but opposite experience some years ago when I switched from ATI to NVidia. Is ATI actually working better than NVidia again? My card works great for it's age but my hardware is pretty old. Back when I preferred ATI I remember shelling out money for a few cards while never getting 3D acceleration to work decently. Finally I switched to NVidia. I don't want to risk buying 3 cards to finally end up with a working one again. Should I try AMD/ATI?
      • by macshit (157376)

        I had a similar but opposite experience some years ago when I switched from ATI to NVidia. Is ATI actually working better than NVidia again? My card works great for it's age but my hardware is pretty old. Back when I preferred ATI I remember shelling out money for a few cards while never getting 3D acceleration to work decently. Finally I switched to NVidia. I don't want to risk buying 3 cards to finally end up with a working one again. Should I try AMD/ATI?

        It's not just whether the ATI closed source drivers are a little better right now -- it's that ATI (AMD) chose the right route for the future, whereas Nvidia dug in its heels, stuck its fingers in its ears, and chanted "la la la la I can't hear you." That means that ATI/AMD GPUs will be far better integrated into the FOSS infrastructure in a few years, will be far more portable to many environments, and far less dependent on the whims of the market and dodgy closed-source coding practices.

        If ATI's closed

        • by morgauxo (974071)
          I totally agree. I have this NVidia setup now where I have two xorg.conf files and a couple of shell scripts which copy them in to place and restart xdm each time I go between gaming (single monitor) and desktop (dual monitor). I'd love to see the driver better integrated and maybe just throw a switch in the KDE control panel to do this. Also, rebuilding nvidia-driver for every kernel update sucks.

          But as a consumer I don't want to spend the money until I know it is the better choice now, not will be in
  • by fph il quozientatore (971015) on Friday January 07, 2011 @01:18PM (#34793022) Homepage
    I thought the biggest stopper in building OSS 3D drivers were that they usually contain technologies covered by software patents. What happened to the patented parts? Have they been stripped out of this OSS version, with the effect that it is now slower than its closed-source counterpart? Or did they find another way out?
    • by Kjella (173770) on Friday January 07, 2011 @01:42PM (#34793458) Homepage

      The closed source drivers remain closed for many reasons, but I think licensed code is a much bigger part of that than patents. The open source drivers are built separately from scratch and are in general much slower yes. They have estimated - and don't take this as an official AMD statement but guesswork from the people working on it - that the open source driver could reach 60-70% of the closed driver on average using the simple architecture they've chosen. Simply because the closed source driver has a ton of code paths and optimizations for various situations, the OSS team is much, much smaller than the closed source team and can't possibly replicate that anyway.

      • OSS team is much, much smaller than the closed source team and can't possibly replicate that anyway.

        I would be surprised if there are more than half a dozen engineers in either AMD or NVidia working fulltime on Linux drivers. There are maybe two or xorg developers working somewhat fulltime on Radeon, so the difference in team sizes is not really gigantic. I would think that access to information and historical code base to draw from is more of an issue. However I do not agree with your "can't possibly" at all. It is just a matter of time.

        • by spitzak (4019)

          The Linux drivers reuse code from the Windows drivers (easily seen by locating identical bugs in both the Linux and Windows drivers). I suspect most of the interesting speed-related parts are shared. So there is a lot more than the "Linux team" working on the driver.

        • by Kjella (173770)

          I think you misunderstand how this works. Some 97-98% of the AMD driver effort goes into frglx, with probably 80%+ shared between Windows/Mac/Linux. The remaining 2-3% work on OSS drivers. And these 98% aren't just passive consumers of information, they work tightly with the hardware guys on creating the design and interfaces and programming model and it's probably just as much their work as the hardware designers. And they got access to prerelease hardware, multi-million dollar hardware simulators and so o

          • in this world, with their actual size and other limitations I'd estimate they could probably reach matching performance with frglx for one generation of cards in 50 years.

            Ha ha, I'll make a note of that claim.

            • by Kjella (173770)

              Well if you do, please understand me correctly... The closer you want to be to the hardware limits the harder it gets and it's not linear it's exponential. Like with today's team maybe you can reach 60%. If you triple it, you might reach 80%. If you triple it again, you might make 90%. Triple it again, you might make 95%. Then there's all the little gotchas you won't know unless you made it which you might never figure out.

              To use a car analogy (yay), as long as you got a 600 hp engine it'll be a fairly fast

      • by morgauxo (974071)
        Maybe so. What happens when developers outside of AMD start hacking at it?
    • by AC-x (735297)

      Doubt it's anything to do with patents as the whole point of patents is you have to publish your algorithm in full publicly, so they're not revealing any secrets if they (as the patent owner are allowed to do) release source code. If if anything it's because they use a load of optimisation hacks in their drivers that they want to keep as trade secrets.

  • by thomst (1640045) on Friday January 07, 2011 @01:18PM (#34793036) Homepage
    I didn't realize it was on fire.
  • I can't find any info, so which version of OpenGL is supported? Is GL4 supported? Also how's the performance compared to the closed driver? I see everyone cherishing AMD for this, but being the news about a high performance 3D card, i'd expect the usefulness of their drivers and the real effort of open sourcing is on the 3D side, not so much on the 2D side and there's practically no information about this.
    • by Kjella (173770) on Friday January 07, 2011 @01:49PM (#34793558) Homepage

      No, OpenGL 2.1 is the highest supported [x.org], but this is because Mesa - the open source implementation of OpenGL - doesn't support anything higher. Somebody needs to implement OpenGL 3 and 4 before there can be drivers for it.

      • by goruka (1721094)
        That makes no sense, MESA doesn't have to support anything. Are not OpenGL extensions designed so the driver can provide higher versions of GL without the API (headers) having to care? (which is the case in Windows) If AMD releases a driver, they can easily release opengl 4 through extensions..
        • by Kjella (173770)

          That makes no sense, MESA doesn't have to support anything. Are not OpenGL extensions designed so the driver can provide higher versions of GL without the API (headers) having to care? (which is the case in Windows) If AMD releases a driver, they can easily release opengl 4 through extensions..

          OpenGL is a big Graphics Library - that's what GL stands for. While cards say they support OpenGL 4.1, it doesn't actually mean it's just to map OpenGL calls directly to hardware functions with 100 lines of code and you're done. It means it has the necessary hardware to accelerate the primitives, once you get that far. If you want to support a higher version of OpenGL than mesa, then you have to supply your own OpenGL implementation. Since you don't appear to have any clue what that means, let me tell you t

      • by Daniel Phillips (238627) on Friday January 07, 2011 @02:47PM (#34794344)

        Mesa GL 2.1 is already quite kickass, with a very capable shading language and plenty of extensions covering most of the advantages of OGL 3/4. The main improvement I'd like to see is geometry shaders, which are getting close [phoronix.com].

  • I've been pretty impressed with the closed source NVidia drivers but I'll definitely consider switching to AMD the next time I'm shopping for parts. I totally get the idea of not caring about politics so long as your video works but there are a few disadvantages to the closed source drivers.

    1) Polish - They have their own configuration utility. NVidia-settings is nice enough but it makes a distro seem kind of unpolished when there is this nice KDE or Gnome control panel for setting resolution and such b
  • I have been avoiding ATI cards and binary driver in my Debian/Linux boxes. I want stability and speed like NVIDIA's. Is it time to switch yet like with my two years old ATI Radeon 4870 video card?

Never trust a computer you can't repair yourself.

Working...