The State of ATI Drivers on GNU/Linux 173
An anonymous reader writes "After 50 days of the Phoronix editor-in-chief exclusively using ATI Radeon hardware in his system, he has issued his final blog post dubbed The State of ATI Linux. Topics covered include the very low frame-rate performance, image quality, overclocking X.Org 7.1 support, Big Desktop/Dual Head, Linux CrossFire, and other relevant items to gamers and Linux enthusiasts. From the article 'While discussing this trial with a colleague that was not familiar with the quality of ATI's Linux drivers he immediately classified ATI Technologies as attempting to fine-tune a hull on a ship while there is a giant hole in the side. However, is this truly the case?'."
Not good (Score:3, Informative)
Re:Not good (Score:3, Informative)
Re:Not good (Score:5, Insightful)
It's amazing to me the author of the article can put out as much verbiage as he did about this issue without ever once mentioning the real problem here - ATIs refusal to document the card interface so that the hardware can be properly supported.
Until they do, their customers that use Linux, *BSD, etc. remain broadly unsupported. Only a small subset of free systems even have the option of using the mushware they want to substitute for documentation, and at a cost many will not pay. They're making themselves irrelevant in what is probably the fastest growing segment of the computer market. Why would a free software user shell out big bucks for the latest ATI *or* Nvidious card only to face the choice of running it without accelleration for the same performance as a much cheaper card, or with buggy opaque mushware that that doesn't perform that much better and taints your system, assuming it will even run on it, which it often won't?
Here is why they can't (Score:2)
This isn't so much a threat to business models in the software business because there is WAY more competition and charging for support is a valid way to make cash.
What do you want them to do? Give away the hardware and charge for support? Yeah. Good luck with that.
Re:Here is why they can't (Score:5, Insightful)
The hidden premise here is that somehow documenting the interface will make it easier for competitors to 'steal' some advantage. That's so obviously wrong in so many ways it's shocking someone would assert it in good faith.
What are they going to do? Copy the interface so their card will be compatible with the other cards drivers? Well, yes, I suppose someone could do that. Wouldn't necessarily even be a bad thing (standard interfaces are generally considered a good thing, even ad hoc standards.) But this is a far cry from somehow "stealing" the actual video card technology. That technology is, in many cases, patented, rather than protected as a trade secret, so the competition can (and you can bet, has) gone and read the patents right off anyway. They just can't legally imitate it too closely. And to the extent there are things in hardware that *are* trade-secrets, a disection of one of the cards would be a much better way to get at them. Looking at the external interface is the last method one would use to try to disect the inner workings of a device. Note that refusing to disclose the interface doesn't do jack to stop the competitors from disecting the hardware.
No, I'm sorry, that whole line of argument is utter nonsense.
Don't leave out the open-source process (Score:2)
Re:Here is why they can't (Score:2, Insightful)
Re:Here is why they can't (Score:2)
Re:Here is why they can't (Score:2)
Re:Here is why they can't (Score:3, Insightful)
How is not documenting the registers going to prevent NVidia from putting an ATI chip under an electron microscope to analyze their circuits?
Face it. If you're one chip fab competing against another one, documenting the externally-exposed registers for programmers is NOT going to deter your competitors in the slightest, nor is releasing binary-only drivers. Remember, decompiling code for reverse engineering IS l
Re:Here is why they can't (Score:2)
Re:Not good (Score:2)
If they'd document their interfaces, then we could get drivers that work, that's kind of the point...
Re:Not good (Score:2)
Re:Not good (Score:4, Informative)
Re:Not good (Score:3, Informative)
I also compile it with gcc 4, not 4.1 because that's what I built my kernel with and patched it for suspend. The nvidia driver even behaves with hibernating.
Re:Not good (Score:2)
Re:Not good (Score:2)
Re:Not good (Score:2)
Re:Not good (Score:2)
The process is simple, log out from and stop X, run the newer NVidia installer, start and log back in to X. Most real distros will have a method to do this fo
Short answer: appalling (Score:5, Informative)
And I really really truly regret it.
The main purpose of this computer was TV-Out, a feature only supported by the proprietary firegl drivers. The version I first got (8.16.20) didn't feature any overscan controls, so it sat in the middle of our television. After a couple of releases of this, we got 8.21.something which broke it even more - in fact, now you could only see the top third of any video you were watching with XV. At the same time of course, there was no 3d support at all.
Since then, I got a VGA->RGB Scart cable, and I've been able to switch back to the free drivers. The quality is significantly better - working 3d, a full screen picture and snappier menus. I plan to be very very careful when buying ATI again.
Re:Short answer: appalling (Score:2)
I suggest an alternate course of action. (Score:4, Insightful)
After learning exactly how "awesome" ATI's driver support was when I tried to setup 3D with my Radeon 8500 (and also Xinerama, etc) to play WoW under Cedega, in both 32-bit and 64-bit modes, I switched to nVidia and haven't looked back (yes, nVidia's drivers ran with Cedega and WoW in both 32-bit and 64-bit Linux installs perfectly well).
So, I suggest to you, to never buy ATI again. Saying you'll be careful when buying ATI again, is like saying you'll be careful when shoving a live scorpion into your pants again. ATI is shit. Regardless of what their hardware might do, if you don't have drivers to make it do it, it's the same as not having the card!
Re:I suggest an alternate course of action. (Score:3, Insightful)
But yeah, Nvidia drivers if I want performance.
What was the question again? (Score:4, Insightful)
Yes.
The state of ATI drivers on Windows is pure crap. It's even worse on Linux.
Re:What was the question again? (Score:3, Insightful)
However, I have tried with a hell of a lot different Linux/bsd distributions [freebsd, gentoo, fedora, Mandriva (full powerpack DD illegaly bittorrented version), Ubuntu 6.06, DSL are the ones I rememb
Re:What was the question again? (Score:5, Funny)
The Open Graphics Project [opengraphics.org] is your friend.
Re:What was the question again? (Score:3, Informative)
From the linked link:
It is a project to produce a PCI graphics card with fully specified programming interfaces. This card will be optimised to be fast for current and next generation GUI environments. This means it is mostly designed for 3D operations, specifically those that are used to render GUIs (Graphical User Interfaces). It will accelerate games to varying degrees, but that is not its primary purpose. It is intended to be a well-documented card that can be easily _
Re:What was the question again? (Score:2)
Intel actively supports the DRI drivers. It's really too bad that the performance is so amazingly bad.
As far as I'm concerned, your best bet is Nvidia, followed by Intel Extreme Graphics for non-gaming.
Re:What was the question again? (Score:2)
Re:What was the question again? (Score:2)
Re:What was the question again? (Score:2)
Suppose they got the HDL for a super-duper 16 pipelined vertex shading GPU today. Could they still roll them out to end users for less than $500 a pop? Probably not.
Tom
Re:What was the question again? (Score:3, Informative)
Not true at all. There is no *high performance* graphics hardware for free software at the moment, but 'decent' doesn't imply 'top of the line' but rather 'good enough for general usage' and there are definitely some choices there. The on-board Intel video has great support, the Via is nearly as good I'm told, the Matrox G550 (±$30 retail, and awesome performance for the price) is fully supported with D
Re:What was the question again? (Score:2)
Of course, there is always the possibility that some other company reconsiders its stance on documentation for their chips.
S3 would be a candidate, if they are not blocked from releasing their specs due to third party IP inside. Right now their the marketshare is quite small, and capturing the
Re:What was the question again? (Score:2, Troll)
As far as that line from ATI is concerned (the "third party IP" claim) it is BULLSHIT. There is nothing proprietary about "to do foo, send value (n) to register fc70" that can reasonably violate any NDA, violate any copyright or patent, or give an edge to any competitor. ATI is just a bunch of megalomaniac pricks, a trait they inherited when they purchased the rotting carcass of Diamond Multimedia, which had a similar elitist a
Re:What was the question again? (Score:2)
Perhaps you have a reversible configuration error somewhere?
Re:What was the question again? (Score:2)
Re:What was the question again? (Score:2)
Re:What was the question again? (Score:2)
Re:What was the question again? (Score:2)
One more reason to love Gentoo:
Done. There may be configuration required later, but the point is, drivers which aren't part of the kernel, or which you want a newer version than the kernel supports, can be distributed as "packages" on Gentoo.
Only thing left is to make a script that rebuilds them all when you install a new kernel. That's simple enough, every time you find one of these packages, add it to a file, one line per package -- say, /etc/rebuild_on_new_kernel. Then you can
Re:What was the question again? (Score:2)
Re:What was the question again? (Score:2)
I never got distcc working properly -- that and ccache, and every other solution I tried to speed up compilation, simply didn't work well, or didn't work at all. After my latest round of hardware purchases, I figured out that nothing was going to take too long to compile anymore. Longer than I'd like, but not too long.
Any speed boost you get is nullified by the amount of time you spend compiling the system. Not to mention, both of my main Linux boxes are
Re:What was the question again? (Score:2)
I got distcc working properly not only for x86, but also for cross-compiling MIPS. (I ran another distcc on a different port for that.) My Indy died though, so I haven't messed with that in a while. ccache will only help you recompile the same exact object file with the same flags faster, it's meant for people doing maintenance, where make wi
Re:What was the question again? (Score:2)
There is no simple install script for this, but if you want to build modules in isolation without going through a kernel compile cycle, I believe you're looking for module-a
Re:What was the question again? (Score:2)
One of these [matrox.com] would fit the bill, I think.
However, is this truly the case? (Score:5, Interesting)
yes.
Use nVidia if you want performance. They use a standard code base between all OS's. 95%+ of the code that is in your Linux driver is in your windows driver. The drivers are stable and have great performance. This has been hashed out many times on various OpenGL forums...
Re:However, is this truly the case? (Score:3, Interesting)
Re:However, is this truly the case? (Score:2)
(I'm not just a gamer, I do some 3D visualization programming
horrible drivers (Score:2, Informative)
Using dualhead is just unusable because the driver fucks up X and requires a full restart of the system each time I want to change users. Opengl apps crash randomly. It's just sad...
Nvidia all the way guys! Don't fall for the cheap ATI cards!
Re:horrible drivers (Score:2)
Does anyone know of any nVidia-based PCIe 1x cards on the market?
Re:horrible drivers (Score:2)
Wait a second, you want multi-screen on Linux with PCIe 1x? Screw ATI and nVidia; you want this! [matrox.com]
Accelerated Linux Users stay away from ATI (Score:2)
Gaming? (Score:4, Informative)
The point is, without solid support for gaming, I don't care much about the drivers as long as I get a good display and reasonable 2d performance. But when I start gaming, I need the performance to just be there. There is no excuse for it not to be really freaking easy!
poor (Score:3, Funny)
Oh, the article was longer than that? Weird, there isn't much to say. I know this: If I had been able to see into the driver future two years ago, I would've bought a different notebook, one with an NVidia card.
if you've wondered why ATI & Nvidia aren't coo (Score:5, Interesting)
Re:if you've wondered why ATI & Nvidia aren't (Score:3, Interesting)
Re:if you've wondered why ATI & Nvidia aren't (Score:5, Interesting)
Re:if you've wondered why ATI & Nvidia aren't (Score:2)
They'll have to do better than that to get me off Nvidia. Yes, Nvidia is proprietary, closed, and does have some problems. They also are rarely more than one minor version behind either kernel or X, and I haven't yet run into something that I can do on Windows that I can't do on Linux with these cards.
It is exciting, yes. I would buy Intel graphics if they were as good as or better than that generation of ATI or Nvidia -- especially if they were willing to work with the community,
Re:if you've wondered why ATI & Nvidia aren't (Score:2)
- Every Intel video solution I've seen uses solely 'shared memory' - e.g., your system's main memory rather than dedicated RAM
- Intel likes to offload everything to the CPU
- Performance is not up to par compared to Nvidia, ATI, or even S3.
The software support for Intel's solution is VERY good, but until the above drawbacks are addressed, I'll choose Nvidia's free/proprietary solution over Intel's Free/Free solution.
Re:if you've wondered why ATI & Nvidia aren't (Score:2)
Works pretty well for me (Score:5, Informative)
Re:Works pretty well for me (Score:2)
Well, duh -- that's because you're using an expensive-ass professional card, where ATI put actual effort into the drivers! Those of us with the normal consumer cards, on the other hand, are screwed.
ATI bad under linux, worse under windows (Score:3, Interesting)
Way back when, I had my old box set up as a dual-screen, dual-boot (linux/win98). 1 ATI card, 1 integrated video card. Both linux and Windows had no problem using both cards.
Now - upgrade to a better box, throw in a Radeon 9200, and nothing works properly except under SUSE. Ubuntu, for example, insists on using only the PCI card (doesn't matter which one you have configured as the primary in the bios, PCI or AGP).
So, throw on a copy of Windows. Ha - the situation is worse. W2003 uses the 9200, but in 4-bit "colour", 800x600 res. The other ATI card is invisible to the system. Installing the drivers - oh joy - they refuse to install. XP Pro? No real diff.
SuSE 10.0, on the other hand, saw and configured both cards. However, trying to install ATI's drivers under both Ubuntu and SUSE failed - the install program craps out.
When it comes to video cards, from now on ATI means "All Time Ignore". I didn't have these problems with the old GeForce 2 with TV-out that worked perfectly.
Re:ATI bad under linux, worse under windows (Score:2)
Look at their support (Score:2)
"Linux is a clone of the operating system UNIX"
If they don't even know what Linux is, how well do you think they can support it?
Re:Look at their support (Score:2)
OpenGL Lockups (Score:2)
It's darned frustrating. I've written a fair number of graphics drivers in my day (all for BeOS, I'm afraid), so I have plenty of sympathy for driver writer
Re:OpenGL Lockups (Score:3, Insightful)
1. Spent more time to document the damn hardware
2. Opened the interface to the public.
The problem is the hardware is always in a state of flux and just incremental improvements. Your GeForce 7800 is probably based on the same HDL source as the 6600 with appropriate changes. This means that legacy symbol names from one project creep up into the new space. You get odd names, combined with lack of comments and documentation [compliance] leads to hardware with "oops" that t
Re:OpenGL Lockups (Score:2)
Except that Nvidia is certainly the faster of the two, and tends to keep pace with development better than some open source projects.
I wouldn't be surprised to see this fixed in a week or so. Certainly not more than a month.
Re:OpenGL Lockups (Score:2)
Which is ok because I'm too busy writing my book to miss the 3D accel.
Mostly I like my cards (NV43 iirc) because they were decently priced, work over PCI-E (a must for PCI TV tuner owners), don't need fans and provide good performance. Most of the time I'm in 2D mode anyways so it doesn't matter. But accelerated bitblts are nice and handy (specially for overlays)
Weird. (Score:2)
I really don't know what's wrong with your setup. I might be able to help you over at irc.freenode.net#gentoo
Re:OpenGL Lockups (Score:2)
The only time I encountered serious problems was when I reflashed the bios of my motherboard with the wrong firmware version, and AGP disappeared.
Re:OpenGL Lockups (Score:2)
I think there may be a BIOS upgrade available, but I long since blew Windows completely off the machine. I might be able to re-flash the BIOS under BartPE, but that's a research project...
Schwab
Latest drivers are a significant improvement (Score:3, Informative)
I don't know what the support is for desktop cards, but for laptops ATI is now a viable option to consider.
A lot of the negativity in previous comments seem to be based on past experiences - try the latest driver if you have a chance, you may be pleasantly surprised...
Re:Latest drivers are a significant improvement (Score:2)
Useless bitching about no/bad open source drivers (Score:5, Interesting)
The Open Graphics Project comes up pretty high in google searches. When someone writes an article like this, it tells me that they didn't even TRY to do their homework. From reading the article (yeah, I read it!), it would appear that the author isn't seriously looking for alternatives. It's reasonable enough to evaluate ATI and nVidia drivers. What's unreasonable is to make everything totally one-sided by not mentioning the alternatives.
Re:Useless bitching about no/bad open source drive (Score:3, Insightful)
On the other hand, Intel has been providing specs and source code for their integrated graphics chipsets. This includes hardware accelerated 3D, though the chips aren't up to the nVidia and ATI top or upper-mid range. Hardware T&L is missing, for one thing.
Re:Useless bitching about no/bad open source drive (Score:2)
Besides, the long term goal of the OGP is to have open hardware. Wouldn't you like that? Don't kill them based on their short-term goals. Think about the future, lest you never end up with o
MATROX!! (Score:2)
You know, Matrox [matrox.com] doesn't have the fastest cards, but they do have Free drivers that support 3D. Since you've failed to mention them, perhaps you've failed to do your homework as well!
Re:MATROX!! (Score:2)
Re:Useless bitching about no/bad open source drive (Score:3, Insightful)
Re:Useless bitching about no/bad open source drive (Score:2)
What is this nonsense ? We bought the product like anybody else, so we are entitled to have support.
Besides, we are not asking for effort from them to support us, just some docs so that we support ourselves.
Try at least to understand what this is about.
They would make negative profit
That's not what happened to NVidia. NVidia won the movie
Re:Useless bitching about no/bad open source drive (Score:2)
Re:Useless bitching about no/bad open source drive (Score:2)
ATI on-going POOR support of the Linux community (Score:3, Informative)
On of the first attempts by ATI to provide an actual ATI supported package for Linux was the VHA Kit. This was supposed to be a library/SDK made by ATI and Loki Games to allow Linux access to the Rage chipset support for hardware assisted decompression of MPEG2 so that iDCT did not need to be done in software. When I have asked ATI about the VHA kit and if they have any commitment to providing on-going support for hardware assisted iDCT for Linux, they claimed that the kit was never distributed because of lack of interested in the community. This seems really fishy since release of the kit even made it on Slashdot and there where several comments at the time expressing interest. Later, a former developer from Loki stated that do to limitations in the Rage chipset implimentation of moving data back and forth, it was faster just to do iDCT in software.
Then the Radeon came out which should have addressed the limitations in the Rage. And nVidea released their closed source drivers with iDCT. While it is possible to do iDCT in software for the 480i resolution of DVDs, for HDTV tuners such as pcHDTV, a nVidea card is almost a requirement to view 720p and 1080i MPEG2 streams. ATI got so many requests for iDCT support that they put online a FAQ on their support site claiming that Gatos was working on the issue. In reality, the Gatos mailing list had posted multiple times that they where not working on iDCT at all. When I contacted ATI requesting to get the Radeon specs needed to support the iDCT support myself, they stated that such information is *NEVER* released outside of ATI. They went to explain that even if the developer signs a NDA, they still will never release the specs to do iDCT support.
Then the All-in-Wonder 8500 which was supported by Gatos was discontinued so I contacted ATI to offer my help to work on Gatos support of the All-in-Wonder 9700. They ask me to be patient and they would be getting back to me. A couple *YEARS* later and they still haven't gotten back to me. According to Gatos, they have gotten around to providing the specs and example hardware to one of the developers. But while Gatos is "open source" in the fact it is GPL, no one else can be much help to the project since the Gatos developers can't legally give the specs to any potental developers. All they can do is tell potental developers to contact ATI which result again with a request to be patient for *YEARS*.
The All-in-Windows 9700 is now discontinued and the new mainstream AIW card is the AIW 2006. Gatos doesn't even claim to have been provided any specs for this newer card. The ATI prioritary drivers provide no support for the tuner at all. And ATI continue to blow off requests from any potental developers except for the ones they already have an established relationship.
And for some reason that eludes me, people still claim that ATI does a better job of supporting the Linux community than nVidia! Does ATI's drivers provide iDCT support for Linux? nVidia's drivers do. If you call ATI right now asking for driver programming specs for any shipping Radeon chipset, do they actually provide it or tell you to just wait (and wait and wait and wait)? nVidia is at least honest about what programming specs they will openly release and what they won't.
Re:ATI on-going POOR support of the Linux communit (Score:2)
Bite your tongue! If you check out the Gatos project web site, you'll note that ATI actually sent the Gatos folks a couple of video cards. How dare you say ATI doesn't care about supporting a growing market segment?
What did I do wrong [right]? (Score:3, Funny)
Re:What did I do wrong [right]? (Score:2)
Its a two-way street. (Score:2, Informative)
But I can't place all the blame on ATI or NVidia for the state of drivers. Some blame lies with the Kernel Developers.
Before you start sending me hate mail, hear me out...
The kernel developers went with ideology rather than reliability when it came to the driver API. They purposely manipulate their API and hope that this will give ATI and NVidia
Not limited to just Linux (Score:2, Interesting)
Piss Poor (Score:2)
I'm immune, All AMD and Nvidia for years now. Haven't seen a thing to change my attitude to date.
Re:Forget this. We need an improved free driver. (Score:2)
close (Score:2)
Balmer is the devil.
RMS is Jesus. (looks like him, no?)
Re:Forget this. We need an improved free driver. (Score:2)
Remember: it's the GPL camp that actively impedes proprietary software. If OS popularity went the other way, problems with proprietary drivers would be way worse. You'd need to run a special kernel compiled by ATI. The BSD license offers **zero** defense against pro
Re:state of my personal machine and ATI (Score:2)
That said, buying a laptop with an nVidia graphics card is not always easy. A lot of companies reserve the nVidia cards for their super high end, extremely heavy and bulky, "gamer" laptops. I don't see why you couldn't ha
Re:state of my personal machine and ATI (Score:2)
They really are better for a select few cards. My 9800 Pro is smooth and solid now with the fglrx (this is on Ubuntu 6.06). I admit to not running very demanding games on it however, just XGL when I want to impress, and occasionally blender.
If however, you have a laptop or even a newer card, you're still pretty much out of luck.
Re:state of my personal machine and ATI (Score:2)
It is somewhat different. I remember very well the days back, when NVidia announced they would provide open source drivers, because I specifically bought a NVidia card then (something TNT2 or so). However, all they provided was an unreadable source code blob and
Re:fglrx vs Xorg (Score:3, Insightful)
Having some experienc
Re:fglrx vs Xorg (Score:2)
Rather then fund a rival manufacturer, raise bounties for deals like the Weather Channel deal which brought us the free r200 drivers. Allow people to put up money for a specific card/manufacturer, range of cards, feature set (play doom3@res*fps) or FSFs pick and have a very public page where shareholders can come and get an idea of how much money is there for the taking. I suspect most people could be convinced to pledge their cash for the FSF selection (which means the FSF has a budget to negotiate wit
Re:fglrx vs Xorg (Score:2)
I had an ATI X700 in my desktop PC. It was almost a year old. The open source xorg radeon driver didn't work, but it almost did in xorg 7.0. The fglrx driver worked fine for 2D, but I didn't run it much in 3D. I tried the Kororaa liveCD with XGL and fglrx, and it locked up a few times. I've since upgraded to an NVIDIA 7900GT. The open source xorg nv driver works fine for 2D
Re:ATI Experience (Score:2)