ATI Committed To Fixing Its OSS Problems

Sits writes "Chris Blizzard blogged from the Red Hat summit that an ATI marketing spokesman said, from the stage, that ATI knows it has a problem with open source and is committed to fixing it. Does this mean ATI will finally resolve alleged agpgart misappropriation, and fast track the release of open source 2D drivers on its latest cards while releasing specifications for its mid-range cards? Or is ATI only concerned with fixes to its binary driver to maintain feature parity with competitors?"

Likely binary drivers only.

I'd wager a guess they're going to fix the binary drivers only.

Why would they open a spec when they can compete with the binary drivers?

Re:Likely binary drivers only.

The video card industry is so secretive with thier software. ATI even locks the BIOS so after POST you cannot access the card to download it. They are so afraid that the competiition will find out how they work or that someone else will build a better driver. This is the only part of the PC buisiness that is this large, yet this secretive. I thing that they are just overly paranoid.

Re:Likely binary drivers only.

Or maybe they have some sloppy hacks to try to improve frame rates for certain games so that they score better in comparisons. Anyone remember the Quake 3 [slashdot.org] fiasco that ATI was involved with?

IP

They have already sold the card, so it doesn't matter as far as revenue who writes the best driver. Good open drivers might help sell cards. I would sure choose a good card with a good open driver.

I think it's an IP issue. They've bought into some fundamental patented IP, the license forbids releasing driver source (or it's something they have patented and it is counted as an asset on their Balance Sheet), and the patent covers something so integral to their design that it isn't worth the R&D it would take to get around it.

Re:Likely binary drivers only.

"I'm sure that they could even get a performance boost if they let millions of hackers with tons of free time optimize things for them."
I am not.
1. Millions of hackers? There isn't a single FOSS project that millions of hackers have contributed too.
2. There are very few people with the experience to write a good much less great 3d driver.
3. Even with the specs I am guessing that the majority of contributions will be security or code clean up and not performance optimizations.

Re:Likely binary drivers only.

1. Millions of hackers? There isn't a single FOSS project that millions of hackers have contributed too.

Pedantic behavior rarely convinces anyone.

2. There are very few people with the experience to write a good much less great 3d driver.

You see, that's funny, you get a whole set of guys who are busy writing what many consider complex programs. Here's one you often see doing fairly well. [kernel.org] Yet obviously, these same people are totally unable to write a working graphics driver. Even though they have written just about every other kind of driver between them, and had them overwhelmingly beat the crap out of the closed-source sector.

3. Even with the specs I am guessing that the majority of contributions will be security or code clean up and not performance optimizations.

So? Security in binary graphical drivers has been a real problem in the past. (note the 6th or so post down, with the link). [kerneltrap.org]

Re:Likely binary drivers only.

Actually, it's more likely that the reason all their software is locked away and kept secret is because it's probably infringing on numerous software patents. When Joe Sixpack can go down to the patent office and register a doubly linked list as his own invention, lots of possibilities for lawsuits open up.

I did some research into this for a course, but I don't have sources to cite off the top of my head. Definitely something worth looking into.

- shazow

Does this mean .......

Why dont you ask ATI what it means. How is Slashdot supposed to be privy to ATI's roadmap?

This is not news

If you read the original article that all of this questioning is derived from you realize the article summary has more content than the linked story. This means aproximately nothing. ATI pays lip service to open source software news at 11.

Open Source supporters within ATI

I know from talking to them at the Ottawa Linux Symposium a couple of years ago that the technical people within ATI were keen to support Linux the best that the could, but said they were mainly limited by management / legal to aim for competing with whatever nVidia offered the Linux community. If nVidia offered a complete open source driver, they would be pressured to do the same.

Re:Open Source supporters within ATI

I have owned or made purchasing decisions for 6 3D graphics cards.
* 2 were Matrox G400s, based on their being the first mainstream card to get 3D hardware support under Linux. I even ran Utah-GLX on one.
* 1 was an ATI Radeon 8500LE, based on price/performance and the existence of the open source R200 drivers.
* 3 are nVidia cards, since there's no competitive contemporary open source 3D any more, and the quality of nVidia's binary seems to be better. There are reverse-engineering efforts on both, but it's unclear who will be the clear winner on this.

So I *have* put my money where my mouth is, and will continue to do so.

I also recommend hardware for friends and co-workers, and this is a factor. Even for a friend who is only going to use Windows, if all else is equal I would advise that he "reward" the company for its Linux support. Notice that in this case I said, "all else is equal," and let the friend know why I gave the advice I did.

Re:Open Source supporters within ATI

So much for showing a little leadership? Basically, we will just follow nVidia. I've got 1U Supermicro rackmount servers that have ATI 'Rage MX' chipsets on them and there is not even a solution for these chips other than a very slow unaccelerated driver. They don't even have a working binary driver.

ATI's lack of driver quality and commitment has always been a problem for me. I went from 3dfx to Nvidia and have never personally purchased an ATI product specifically because of their poor Linux support.

Marketing?

an ATI marketing spokesman said, from the stage, that ATI knows it has a problem with open source and is committed to fixing it.

There goes the good old problem solving by marketing. Wait until their developers hear about this :)

I'll believe it when....

I'll believe that AMD/ATI is fixing their problems when I can have a driver that:

1. Supports XvMC
2. Supports the tuner on my All-in-Wonder, either via XVideo or Video For Linux
3. Has reasonable 3D performance without locking up (GoogleEarth will kill my card dead in seconds, requiring a hard power off to fix it.)
4. Has reasonable 2D acceleration.
5. Runs on the current release of the kernel - on i386 and i386-64 AT LEAST.
6. Supports PCIe cards

This is *the* limiting factor which has prevented me from buying a new computer - any new machine would be an i386-64 with PCIe video, and right now the only real choice there would be Intel graphics.

People tend to say OSS support ...

... when they only mean 'Linux support'. And personally, I don't consider closed source binaries OSS support at all. AMD has been good about making the information available for open-source programmers so their chips can be supported. Perhaps their purchase of ATI will force a shift in the corporate culture there too. Well, we can hope.

Re:People tend to say OSS support ...

Mod that post right up. Intel are doing it the right way, by releasing DRI drivers. Any operating system that wants to support DRI can then use the Intel DRI drivers. Once the DRI code is running on a kernel, it's just a matter of porting the small glue layer for each card (or card family) and then the bulk of the driver can be compiled and used whatever the underlying architecture is.

Even the nVidia binary drivers have wider support than ATi, since they work on OpenSolaris and FreeBSD as well as Linux.

Current State

I am currently running the *newst* ati binary drivers and although they have added the Catalyst Control center (improvement ofer the old fglrx control center) mine (and a few other people i know using the same driver) cant seem to get dual monitor to work. And with the Opensource ati driver atleast AIGLX works but still no dual head display.

          ATI needs to step up the quality of their coding and there is no *good* reason why ati does not support AIGLX and why their 8.35.5 is having problems with dual monitors. Because my laptop uses ati and i was so displeased with its state of drivers forced me to go with nvidia when i built my desktop a year ago. Im sure many people using Linux stay clear of ati when possible for the same reason. When and if they get their stuff together it will receive a warm welcome...if they do it right that is.

Also why is it people need programs like envy [albertomilone.com] to install their drivers. Hopefully ATI and nvidia will pick up the slack hear and make it easer to install the drivers.

Dell ....

suddenly Dell is shipping boxes with Linux .... a big customer to ATI .... and Dell is talking to Ubuntu .... "How do we know which of our boxes work well for Linux, will cause us the least amount of tech support grief' ... Ubuntu guy says "well these drivers don't work so well .... they're not well supported by their manufacturers" ..... Dell guy starts crossing boxes with ATI cards off the list .... and tells ATI marketting who start worrying that Dell will start to not buy ATI at all .....

Re:Dell ....

That was my thought. But it gets worse. Dell is also a big and relatively new AMD customer. Intel's integrated graphics solution works very well under Linux. So for the low end Linux solution Intel maybe the system of choice. The Dell guys might start crossing ATI and AMD off the list. Intel offers easier on stop shopping and a more politically correct FOSS system.

Fast track?

I doubt ATI will fast track anything for OSS...

they may eventually solve SOME problems but I sincerely doubt they'll be throwing a team on resolving all of the issues resulting from using one of their cards with Linux.

Give ATi some credit

In some areas, the closed linux binary driver maintains feature parity with the Windows [ati.com] counterpart [ati.com].

Oh Oh!

A friend of mine recently had his dog "fixed". What, exactly, does ATI intend?

Consumer point of view

I own an ATI 9800 PRO graphics card. It's a great piece of hardware. But "I need a good driver", which is translated to: If they don't release an outstanding driver in the next few weeks, my next card will be nVidia. Or better... If they don't release an outstanding full open-source driver in the next few weeks, my next card will be nVidia. Yes. I know that nVidia drivers aren't outstanding, and aren't open source. But I've been stuck in the bad side so long, that I won't be satisfied with "just the same as the competition".

R300 opensource drivers

There are open-source R300 [sf.net] drivers that cover the 9500 up to x850 range of cards. I've used it for on old 9600XT AGP card (AGP chip) and HIS-overclocked X800 AGP card (PCIe chip with PCIe-to-GP bridge). The performance seem to be acceptable for my needs - which is surprising, knowing that R300 driver was completely developed from reverse engineering.

Recently the driver has been included in the official DRI tree. Most distro use it to provide open-source 3D acceleration. It is the default drivers for near every GPL-compatible Beryl/Compiz LiveCD (like Kooraa, for exemple) and function well enough with them (the same can't be said for official binary drivers).

As usual you should stop focusing on the hardware maker - who doesn't { have the possibility to / want to } throw resources at an OS that represents only a smaller fraction of their market share.
You should instead seek what has been produced by the OSS community - through large-scale collaboration they often manage to put out some marvels.

There no way one could except ATI to open-source drivers. They may have problems with code in their drivers that wasn't produced in house and that can't be opened cheaply.
BUT what AMD/ATI realy need to do is to help the DRI/FreeDesktop guys develop their own driver, and for that they need to document a little bit their chips. The best thing could do to the OSS community isn't trying to make their BLOB drivers less borked. The best thing would be to provide list of registers and samples so the community could write a R500 driver.

In other news

Announcing free software drivers for the new Intel 965GM Express Chipset [marc.info]

ATI, NVIDIA: fuck you. Open source graphic drivers are possible, period.

Just do it! Show us the results!

I will believe it when I see the results. I am not trusting ATI/AMD.

Drivers need a lot of work

Unfortunately at work I am stuck with an ATI X1300 card with Linux. I have to put up with text in my editor getting constantly corrupted, my mouse cursor corrupted and lots of other weird quirks. I tried to fire up Google Earth and that just hung. All of these things work perfectly on my nVidia cards at home, and over the years I've used nVidia (since the Gforce 2) I've only rarely had problems.

Even ATI's installer sucks badly. It took a week before I could finally get the ATI driver to install on the computer, in part because it had integrated graphics (which did not work at all with X). The Vesa drivers for the ATI card are far too slow to be usable.

An operating system MUST be license neutral

How do we feel about Microsoft's decision to exclude open source drivers by requiring signatures on everything in XP/Vista? Would we want them to rule out GPLed software based on MFC and .Net? It's no better for Linux to enforce a particular license for drivers or impose license restrictions on KDE/Qt apps. An operating system should be license-neutral for any applications and plugins it supports. A user should not be limited in what kind of hardware he can buy for his Linux computer.

Ok I am stupid ...

But would it not be possible to use windows drivers for *nix?

I naturally do not mean that they would be plug and play. But the Windows driver API must be pretty well known, and they run on the same hardware. So it should be simpler to reverse engineer.

A thin wrapper around the windows drivers could perhaps make it work and hold us over for the short term? Something like Wine for graphic drivers.

I had decided to buy a 14.1" notebook purely for Linux, but i decided against it as it is pretty much impossible to guess if it will run. Trial and error is a bit expensive ...

Re:Ok I am stupid ...

Short answer: no.

Long answer: No. X11+GLX is very different from GDI+DirectX. In almost all cases, it would be easier to reverse-engineer the hardware, rather than wrap the driver api. Also, it would probably be impossible to use windows graphics drivers in a secure manner. And the extra translation layer would kill performance. If you are going to reverse-engineer the drivers, you might as well look at the hardware info, and not the software api.

Note that in some cases, it is possible to use Windows drivers on a *nix operating system. The NDIS network card driver api is well documented, and is supported by projects for Linux and FreeBSD.

HTPC use *better* get them motivated

right now, as much as I dislike it - nvidia IS the linux owner for HTPC use.

all the howtos talk about the nvidia binary (sigh) driver and how it helps (but isn't a full solution) to mpeg motion accel. in hardware.

but with ati, there IS no solution. "don't use ATI" if you use linux and want fast video for home theater use.

I bought an ati card for the windows side of my htpc design - but I won't be buying them again until they show an xvmc driver for linux.

its just a shame they ignore unix like that; especially in the days when HTPC building is really starting to get popular.

Thanks ATI

While everyone is harping about ATI's past sins, I'd like to thank ATI for committing to fixing those problems. We should commend (and purchase from) companies that make our lives easier (I'm looking at you, Broadcom...)

If I know ATI at all

Or is ATI only concerned with fixes to its binary driver to maintain feature parity with competitors?

Yes, they do NOT care about the open source stuff.

Put on the cheerleader uniform

I attended the Red Hat summit last year. Lots of good information (there were a ton of talks about Xen, a good one about the finer points of LVM, etc), but the price wasn't worth making it a yearly thing.

That being said, I think the conference has the potential to quickly degrade to LinuxWorld-level, and this announcement doesn't surprise me. Companies will come out of the woodwork and start screaming "Yaaa, we like Linux! Hooray for open source!" for a week, but then not do anything until the next conference/expo rolls around.

(On a related note, the last notebook I bought came with Intel graphics. I specifically chose this because I didn't want to deal with the headache of ATI and Nvidia's binary drivers. Intel is no saint, but at least having full 3D drivers in Xorg is nice.)

Open? Does not play with DRM, so forget it.

Ever heard of, "Certified Output Protection Protocol (COPP), Protected Video Path Output Protection Management (PVP-OPM),
Protected Video Path User Accessible Bus (PVP-UAB) and Protected Broadcast Driver Architecture (PBDA..."

All lovely things that Microsoft and ATI (will/do) use to piss you off, and make connecting all of your expensive new PC & AV kit virtually impossible.

Better binary drivers? Maybe.

Genuinely 'open' architecture that would enable the OSS community to bypass (more easily) current and future DRM, while still being able to view the result on the lastest hardware? No way.

BS

I had an Ati Radeon Moblility 7500 which I used on Ubuntu Edgy Eft. I used the open-source drivers and it worked fine. 3D support was kinda iffy in some places, but it worked all right. That being said, I now own an nVidia Geforce Go 7300 and using the restricted drivers, it works like a champ. I don't give a rat's ass if the thing is closed or open, if nVidia is committed to releasing a high quality driver for Linux, I'm going to side with them. I can't speak for the Ati binary driver, but given that my old video card wasn't even supported by the binary driver, I'd have to say to hell with Ati...

ATI? Who F**cking cares ATI.

They don't give any f**ck to Linux drivers. More than 5 years, oss people begging them to do something for Linux drivers.

So ? DONT BUY.

Thats simple.

ATI??

Haha, ATI ceased to exist a long time ago. They changed their name to AMD.

Marketspeak = bullshit

Tons of companies say they're committed to lots of things. It doesn't mean anything.

Wait until they produce something that fixes the problem.

So I just called their presales

You know, cause complaining to their Linux department has accomplished jack-shit in the past 2 years. The peon transfers me to his manager, and his manager says "I'm really sorry, I have some very exciting news to tell you, but I can't under our NDA". I asked him if he could generalize the news, to see if it maybe fixes the problems I'm having. He says he will liaze with me to ensure I get a proper response from their Linux team, that will somehow keep me from selling my (STILL) $650 X1900. Anyways, I would kill to be able to install Beryl as easy as I do on my Dell D610. It's the slowest crappiest laptop alive, and yet with the new Intel drivers, Beryl runs awesome! Anyways, I'll keep everyone posted, but like we've been doing for the past 2 years or so, don't hold your breath.

Uh huh.

I'll believe that when my shit turns purple and smells like rainbow sherbert..

More lip service

ATI has proven itself to be unreliable and basically dishonest. Additionally, there is no money motive to do this. Linux users are such a tiny fraction of the graphic card buying market that there is no reason for them to do this work.

Sure would be nice to have open source drivers for any decent 3d graphics card under linux. But it's all about money. Corporations are beholden to their shareholders, and board members can even get sued for pursuing a non-profitable course of action. This would most certainly fall into the non-profitable category.

Please please please vote with your dollars (it's the only vote you have that counts in this country). Even if that means not buying another video card. You're only supporting crap buy purchasing ATI.

Radeon 9600

Does it mean I might be able to get better resolution on my radeon 9600 than 1024x768 on Ubuntu?

Installing any other drivers causes the system to hard lock upon the log in screen.

interesting

considering ATI doesn't even exist anymore. Do you mean AMD?

ATI is pulling your (our) leg

Read my Journal for the scoop, but it works like this: they (ATI) know their OSS/Linux support is "teh suck," but choose not to fix it. Why? The answer is simple: why should they? In the 3D realm, you have two choices ATI or nVidia. That's it. Linux isn't where the bread-n-butter is, Windows is where the revenue is. As a business, you go where the money is, not where your heart may lead you.

What's more, it may not be just one component that's truly sucky: All I know is that ATI's FGLRX + 3D + Xorg = failure. Their driver may be fine, there could be an issue with Xorg and ATI together, or some unseen combo that nobody is looking at--or it would have been fixed. So, as a result you have, really, only one good choice for Linux 3D, and that's nVidia. Nvidia knows this and loves it. ATI chooses to chase the other guy rather than fix things and gain new converts.

In a month or two when nothing has come of this, at least you'll know why. Pay no attention to the flapping heads of ATI until they actually DO something.

Commercial uses

We've been using OpenGL and Linux on ATI cards for our arcade game for over a year now. We're facing a major hurdle, though. AGP hardware is getting harder and harder to find in quantity, and the fglrx drivers don't correctly support vblank in the PCIx cards they have [cchtml.com]. We're trying to use the commercial end to get pressure on them through the buyers, but it's slow going.

When they can't be bothered to get their drivers to pay attention to vblank properly, you know it's not their top priority.

Money talks evert thing else is BS

If Dell had released it's announcement that it was going to sell laptops with Ubuntu on them a couple of days earlier my daughter would be running around college with a Dell laptop instead of a HP laptop. if you want to get good ATI drivers go to Dells website and buy a Linux computer. When Dell calls ATI and says they want a box car of hardware with a option to buy 40 more but only if they have fully functional LINUX drivers ATI will a lot more interested in Linux OSS drivers. Money talks everything else is BS.

Whichever

Competition is inherently good. Just wake me up when they succeed.

It's all in the street cred. Linux, I use nvidia. When I ran OS/2 it was Matrox. As long as ATI realizes PR is cheap but it's places like /. where results are broadcast.

Ati + Dell = headache

I can only speak from the windows side, but suffice it to say it is "difficult":

Typing this on a Precision 450, with a Fire Glx1 and use dual monitors:

Problem1) Ogl screen savers will run on only one monitor and freeze on the other.
Solution: Omega Drivers for fgl. remove/reinstall .net 1 and 2...don't use .net3, at all.

Problem2) Display properties don't reflect current state (mirror/one big display/etc):
Solution: drop the refresh rate/size and check again.

Problem3) Dell PW 410/420's have weak PSU's.
Solution: Sort of. Got a 9800pro (agp 4/8x only, despite specs' claim) for work machine and 9500 for home machine
(agp 1-4x correct, wonders never cease). Stealth upgrade, 410/420 had a 230W PSU, min rec was 300W...thing
would only boot after power up and reset shortly thereafter, and system specs listed Agp as 2x, max.
So a PW 4x0 got a 9500 and sort of worked. Luckily this box had months of uptime and a warm reboot was
fine as long as the card had power applied.
350W PSU in home system worked well with 9800. Worked *very* well (SEG).

Problem4) Dell drivers are *YEARS* out of date for even the most mid/current of cards. Laptops are locked
out as are workstations. Cats won't install on dell systems, hence the omega and other driver releases.
Sadly this is on Dell systems still under support contract, gets worse when you need the latest drivers
(on say CAD/CAM/GIS) for a system/software to work properly.
Solution: ignore the "void the warranty" in order to have a working machine. {hurmph}Nice.

Problem5) All Ati's fault: video acceleration software for DVDs can't always be installed.
Solution: nlite driver forum. Now have the 9800pro and x800pro. installing the s/w with 9800 disk works
for both cards. See, never registered the 9800 (wonder why?) but did the x800.
Ati's number check takes the 9800's #, but not the x800's #. However I can use the 9800's # for both
and both work splendidly.
Nlite forum has the steps that works, so long as your card is supported. (9500pro or better, IIRC)

In a nutshell (TLDR version) Ati, like HP, makes excellent hardware, software almost always sucks rocks.
Dell is just the opposite and sometimes the same (all things considered).
Nvidia: I lost track, what month is it?

Commitment means everything when $$ is on the line

Dell sells a buttload of Ubuntu machines. With what? Certainly not ATI cards.

ATI sees nVidia touting Linux friendly lifestyle with their cards to the newly awakening non-Windows groups (parents, teachers, *egad* politicians) and will then be more responsive. But by then, too late?

By the way, how hard is it for a company to spend the money on five or six Linux developers anyway?

Damn!

I wish I saw those complains about ATI 2 weeks ago. I spent days trying to make an ATI 1300 work on a Dell/Ubuntu, finally had to give it up.

Re:Does this mean hardware hacking is dead?

Hack the hardware? Have you any idea how complicated graphics cards and 3D acceleration is when you have no specification on the hardware at all?

Re:Does this mean hardware hacking is dead?

Uh, people have been working hard to understand how the hardware works in order to write open source drivers. See here [freedesktop.org] for example. The problem is that ATI doesn't open up the specs for their recent cards so there are very few and tedious avenues to having open source drivers (eg. reverse engineer the binary drivers, probing hardeware settings, etc). As far as I know, there's practically full opne source 3D drivers for R100-R200 based cards, somewhat full 3D drivers for R300 based cards, and no support for later models. So the OSS community is working on the driver issues, it takes time without documentation.

Two words

Omega Drivers.

More words:
Thats what I used for my ATI card. If he is still maintaining them, I highly recommend them.

Re:Does this mean hardware hacking is dead?

So far I'm hearing "commercial company hasn't written Linux drivers for their card". That's a legitimate complaint, but if the OSS community's reaction is to whine about it on cheesy blogs rather just hack the hardware...?

No, it is not, but it is strictly for pleasure only.

You see, 3d cards are complicated. On top of that the hardware itself if often finicky with lockups to the point that they should really be considered bugs. So, you can only start once your got the hardware in your hands (which means after release) and with lots of work, at best you will have something semi-working a year later. It will be at least another year before the drivers mature so everyone can use them mostly without lockups. In the meanwhile ATI will release a few more variations and, if you are aiming for comprehensive support, you are back to square one.

If ATI wants to be nice to Open Source it means releasing partial specifications (at the very least) before the card is ready so that all their cards work with 2d, Xv and multi-monitor/multi-card when they are in stores (or a couple of months later) and having full specifications no later than 6 months after release.

Anything else and we are back to scrounging for older well-supported cards - which also happen to be a good deal cheaper and have less of a margin for ATI.

The latest card I have is Radeon 1600 - and given a choice I would gladly go back to R300 (or better yet - Rage Pro) if only those cards supported the resolutions I need and PCI express.

Re:Does this mean hardware hacking is dead?

These things aren't ISA modems or parallel port scanners.

Graphics cards are immensely complex.

Re:Does this mean hardware hacking is dead?

Trying to write drivers for ATI hardware is like trying to write a HD-DVD player program without knowing the keys. Or the encryption algorithm. Or the video data format.

It's hard to do anything other than whine about it when ATI are such a bunch of tight-fisted assholes.

I'll switch back

Here's how it works: I currently buy nVidia, because they're powerful enough to handle things like Quake 4, and because they have Linux drivers that mostly work.

If I didn't care so much about performance -- like if I just wanted something that can do Beryl reasonably well -- I'd buy Intel, because they have open source drivers that rock.

I know ATI can give me competitive performance. If they can also give me an entirely open driver, missing no functionality, and as solid as, say, the Intel drivers, I'll switch to ATI. If Intel comes out with something that has performance that blows away ATI and nVidia, and keeps up the same level of quality drivers, I'll switch to Intel. Otherwise, I'll stick with nVidia.