OSNews

True, but I think the the problem as far as I see it, you'll only start to see it when their new CPU/GPU thing gets going - being it a completely new idea, probably unencumbered by patents, which means that one will be back to square one unfortunately with Ati support.

Unrelated to this post, for me, I'd love to see Ati/AMD succeed and give Nvidia and Intel some much needed competition. At the same time, however, I am reluctant to believe Ati - they've had a good several months to get their 'ducks in line' and they've done nothing to improve the situation.

For me, it isn't so much needing to have open source drivers - that would be a perk, but not a necessity, the issued I have is their lack of updating their binary drivers and maintain them.

Having open source drivers won't fix that; if there is a massive delay between the product launch, specifications being made available, and updating of the source, you're still back at square one in regards to support - whether the drivers are open or closed source, you're still at the mercy of AMD/Ati.

ps. Remember, this isn't the first time they promised improved drivers.

Edited 2007-05-13 21:53

"however, I am reluctant to believe Ati - they've had a good several months to get their 'ducks in line' and they've done nothing to improve the situation. "

Integrating 2 separate companies is not an easy task, there are various duplicate components (ie: accounting, human resources) that need to be integrated, networks to integrate, business practices, corporate culture, it's not an easy task. To expect AMD to just buy ATI and then change the entire company on a dime is unrealistic. I for one am pretty happy about this, and think the timing is perfect.

Especially since, if you really read what the AMD guy said versus what the blogger wrote, it isn't entirely clear that AMD actually promised open source anything. He promised that AMD would work more closely with the Linux community, and he promised to make their drivers work better with Vista as well.

Basically, he admitted that their drivers are awful no matter which modern OS you're running, and that they're going to try their best not to suck so much in the future. Still no DirectX 10 part (not that I personally care), no viable competition to nVidia on the high end, and no reasonable comparison to Intel in their commitment to free software. I'm been an AMD fan in the not-so-distant past, but I've lost most of my confidence in their ability to execute.

Edited 2007-05-14 02:32

I've had no difficulty in producing patches to Intel drivers without extra documentation. The existence of documentation would certainly make it easier, but it's hardly impossible without it.

This ain't no good news. This is the all the same bullshit, AMD is becoming the same opensource fraud Intel already is.

Thanks for nothing, AMD.

We don't want your filthy code, full of magic numbers.

What about providing full and free documentation for your chips, so we can code our own maintainable and quality opensource drivers?

"AMD is becoming the same opensource fraud Intel already is."

Intel is an open-source fraud, now? How so?

They are working with the upstream X/Mesa people to get quality drivers into the standard X.org distribution in a timely fashion. Sure, their binary-only regulatory daemon was evil; but they're fixing that with the iwlwifi driver (which has that regulatory software as part of the firmware). They want people to be able to buy their hardware and have it "Just Fricking Work" with Linux and BSD. No screwing with configurations or kernel modules or anything. It should automagically work and work to its fullest.

Sure, their drivers are still not feature-complete, but they run well enough to use Compiz/Beryl quite smoothly, have nice EXA support, and support Xrandr 1.2; and they've told us that things like hardware-offloading of motion compensation (XvMC), HD video decoding, etc. are in development, slated for publishing within the next several months.

Quite frankly, a company which has such devotion to making their hardware F/OSS-friendly rightly DESERVES our business.

My new workstation is almost entirely Intel-based, for what it's worth - An E6600 processor, an onboard G965 video chipset ("GMA X3000") and HD audio, etc. I'm quite pleased with its functionality. The only thing I wish I had was an Intel wireless ethernet card, as my current Prism2-based card is starting to show its age...

So I ask you again: How is this fraudulent on the part of Intel?

It just a matter of time! I believe, one day *everyone* must embrace FOSS or else they *will* be out of business.

Well, provided I get some support at all for my 9600xt, I will be happy. And here was me thinking my next system would either have to be underpowered Intel or overpriced nVidia grahpics... This is good news, regardless of what they give us, it is better than the nothing we currently have!!

And what made you think you would get nVidia? They are as closed as Ati's right now.

In my experience, ATI's Linux drivers are really nasty. Stuff like AIGLX is unsupported where drivers are available if you have nVidia. ATI's and nV's drivers are closed, but they're not equally bad at running on open source systems!

"Proprietary, patented optimizations are part of the value we provide to our customers and we have no plans to release these drivers to open source," the company said in a statement. "In addition, multimedia elements such as content protection must not, by their very nature, be allowed to go open source."

That was ATI last August: http://news.com.com/8301-10784_3-6104655-7.html

In line with those statements, I'd be concerned whether they're going to open up their codebase beyond basic 2D support.

I'd say that's mostly lame excuses for not wanting to cooperate. Patents are open docs anyway so they aren't revealing any secrets by releasing the source.

And the DRM stuff is a non issue as nobody wants it.

I'd say that's mostly lame excuses for not wanting to cooperate. Patents are open docs anyway so they aren't revealing any secrets by releasing the source.

And the DRM stuff is a non issue as nobody wants it.

The patents may be published but opening the source brings up the question of how they want to handle licensing the IP that the patent represents.

Just because a patent is open for others to see dosen't mean its a free ticket to use that patent.

More correctly, did they use their own code, or did they licence the code off the patent holder - that's going to be the interesting thing - if they were to opensource it, could they re-implement the technology using their own code rather than the licenced.

With that being said, if Ati came out, created drivers for Linux, Solaris, and FreeBSD - they would get a lot of Kudos from the *NIX community.

which version of opensolaris? Can you elaborate, as I am in the process of making thinkpad purchase, Thanks

My Thinkpad z61m comes with an ATI X1400 and last time I looked, the current open source X.org driver did not support this card, meaning I only got a painfully slow VESA desktop in Nexenta. Maybe I didn't try hard enough, but does anyone know of a solution to get accelerated 2D in any OpenSolaris version with this card?

Try opensolaris 11 build 63, it might be supported on it.

My nvidia 6600gt was recognized flawlessly, with the included nvidia drivers. I was even able to see GPU temperature and enable digital vibrance.

You need to create a xorg.conf file in /etc/X11 and use the radeon driver - there have been many people who appear in #opensolaris chat room and find that the reason why they've got crappy snappiness with graphics is drived from the fact they're either using the vesa driver or the ati one.

As for Nexenta, I'd personally give it the heave-ho, and install Solaris Express Developer Edition - there is a new one being either this coming week or the following, which is based off B64/65. This includes Xorg 7.2 server + updates.

You need to create a xorg.conf file in /etc/X11 and use the radeon driver - there have been many people who appear in #opensolaris chat room and find that the reason why they've got crappy snappiness with graphics is drived from the fact they're either using the vesa driver or the ati one.


Wrong. The "radeon" driver, which you tell him to use, does not support the x1400 he was talking about.

Adam

You know what this news means right?

FreeBSD drivers
Solaris drivers
Timely bug fixes
AIGLX support

But people lets be honest. All this is gonna take time. AMD has to resolve the IP issues tangled in their drivers.

Edited 2007-05-13 16:54

A couple years? Really? That's interesting since AMD only bought ATI in late 2006.

I am rather puzzled why the crowd cheered. What I would expect to hear is crickets chirping and a voice from the back row - "Haven't we heard this before?! Less talking, More releasing!"

Well considering that this particular promise is from ATI's new management (AMD), I'm not really surprised people are expecting delivery.

How about we wait a day or so before we start accusing people of things? Perhaps, given how big this story is/will be, there will be an official press statement from AMD. Until then it is just one website saying one thing and another saying something else.

> They need to do *better* than Intel to get my attention.

Well, they would do A LOT better than Intel by simply providing good, free drivers for very fast 3D accelerators, when Intel provides good, free drivers for very slow 3D decelerators.

"[..] when Intel provides good, free drivers for very slow 3D decelerators."

Uhm...The G965 ("GMA X3000") is far from slow. It has a 667 MHz GPU with a directly link to the motherboard's potential DDR2-667 or DDR2-800 memory. It supports hardware T&L, OpenGL 1.5+, hardware video decoding and motion compensation, SM 3.0, etc.

Previous generations did a lot of their tasks in software (especially T&L and clipping, which made games a no-go); but their newer generation stuff is quite good.

Granted, the G965 is in a lot of ways a glorified stream processor, but the fact remains that these are *NOT* slow GPUs.

I confess my ignorance as to the real performance of the latest Intel 3D hardware, though I'm afraid they are far from the minimum for gaming at the currently most common resolution of 1280x1024.

I did have Intel graphics in my previous PC at work, and 3D performance was horrendous, though far more than enough for 3D screensavers, wobbly windows, and cube multidesktops. Not having to fiddle with the Linux drivers on Linux was nice too. If I were to choose what video chips to use in a non-gaming setting, today I'd spec Intel, for sure.

I prefer fully working drivers(closed source) not FREE half-assed OSS drivers that have miserable parts of features implemented.

Ahh but they don't work. X crashes almost every time I log off or switch users. It was only recently that ATI fixed the problem with X crashing when resuming from standby or hibernation.

Don't start a licens flame war.

well,

for this moment I only vote on one thing:

a well performing stable 3D card

so ATI -- no go and that has bene for years
NVIDIA is the best main stream bet. and yes I would like to see open source but sometimes, it just won't happen in time.

The nVidia drivers perform nearly identical because the drivers are nearly identical (nVidia use a wrapper).

For AMD to do the same, it would mean quite a lot of work (Depending on the actual implementation of the driver). But actually, in the long run, it most likely would pay off.

>The statement quoted said nothing about open sourcing
>drivers for ATI/AMD GPUs, but it seems that this is the
>assumption many are making in this discussion.

Have you RTFA? The first sentence is: "AMD will soon deliver open graphics drivers, said Henri Richard just a few minutes ago..."

Next time, consider thinking before you talk; sometimes it helps you seem more intelligent (and intelligible).

Hey dude - you should chill just a bit. The quote said "AMD will soon deliver open graphics drivers...", not "AMD will soon deliver open SOURCE graphics drivers...". I also followed the links that some of the other posters provided addressing this - did you? What I said was that several posters here interpreted this to mean open source drivers were coming from ATI/AMD. You should read more carefully. I would like to see GOOD open source GPU drivers for nVidia and ATI/AMD, but with my feet as firmly planted on the ground as they are, I am not holding my breath - are you? Intel will do this (or so they say) but their GPU solutions are nowhere near where "A" and "n" are.

Sheesh!! Why do people take this stuff so personally.

"The open sourcing of graphics drivers will indeed be good news, but it's not a big surprise. "

From this: http://enterpriselinuxlog.blogs.techtarget.com/2007/05/09/amd-will-...


That's why it's misinformation ! Everyone saying bullshit, adding words, making it looks like nothing true.

They ain't open-sourcing anything! Even if the article says it.

Please, accept my apologies for being so rude; whether you were right or not, or whether you had read the article or not, I should not have been so impertinent.

Wouldn't you look foolish if this all turns out to be false.

So far AMD has said nothing else, and AFAIK they didn't actually come out and say they're releasing open source drivers, this is just an interpretation someone had and there seem to be conflicting ones.

Interview with AMD:

http://liquidat.wordpress.com/2007/03/13/amdati-linux-interview-wit...

"Anyway, opening the drivers is no alternative for Makedon: first of all there is too much 3rd party technology in the drivers “like for example compression algorithms”. The 3rd party companies certainly wouldn’t allow AMD to release that source code.
Second there is too much IP in the drivers. He compares the situation with McDonalds which will not release the recipe to the sauce of the BigMac also: “It is just not economically justifiable to do that.”
Also, an NDA like Intel uses is not possible because - according to Makedon - there is not “so much 3D intelligence” inside the Intel drivers.
However, he adds that the 2D part is already open source and that there are discussions to also release other parts like the install routines as open source as well. He cannot say never, but that AMD will not take this path for at least the next 6 till 12 months."

Good to see you're on the board of the directors, thanks for the heads up.

Why don't all you naysayers give the new management some time to screw up before we condemn them?