Radeon Driver Enables Full 2D Acceleration For HD 7000

[garegin]

Well, Intel graphics drivers on GNU/Linux are Free Software and has given by far the most trouble free user experience for me.

So. Free Software is beating proprietary software hands down here.

cool your jets hippie. intel drivers in windows are more capable AND faster than the linux ones.

[Rexilion]

It doesn't work with it, it works around it. NVIDIA simply has enough resources (and writes most of their code platform-independently) that they can afford reimplementing everything on their own. But the reimplementation is by no means always better - security issues, debugging problems, stability and maintenance are harmed by the fact that it's all proprietary. If NVIDIA suddenly decided not to release anything else for Linux ever, after a while Nouveau would become better than the proprietary driver in more or less all aspects.

Proprietary solutions are not the devil, but they are simply pointless obstacles. They create a lot of problems that could be avoided if at least a part of the driver was open-source. And the only reason why the drivers are proprietary is that their code either has some secrets they don't want to give to others, or it has licensed code from somewhere else. Both could easily be made into proprietary modules for OSS drivers, and nobody would complain about that, because the main part of the driver would still be in a position where it could be maintained and changed as the community, and not the company, decides. And that's very important in terms of new feature adoption and security issues.

That 'part of the driver should be opensource' is kind of moot: http://www.phoronix.com/scan.php?px=...page=news_item

It's partially opensource -> Ppl are still not happy.

Furthermore, yes opensource allows ppl like me to help out. For example, a hibernate bug that affected a lot of hardware: https://bugs.freedesktop.org/show_bug.cgi?id=50121 (which I initially reported). So glad they have a patch right now. But I needed to:

- compile my own kernel
- boot and handle kernels (GRUB)
- handle GIT -> bisecting, remote branches, unbootable kernels
- handle patching
- report bugs upstream where many other ppl just 'dump' their own problems not even trying to figure out if it's the same issue!
- become acquinted with kernel peqularities like dmesg, (undocumented!) boot parameters

I could just have gone with:

<packagemanager> install nvidia-drivers

And be done with it. Furthermore, bugs can be reported to nvidia, the driver includes a dedicated script to aid in just doing that!

And as for nVidia's local root exploit, they are also in the linux kernel. Which is opensource. One of which, initially, went unnoticed... I tested that myself. As a normal user, executing the testcase gave me an instant root shell (wtf).

[GreatEmerald]

It's partially opensource -> Ppl are still not happy.

That's a separate case. The main uproar about it was that they promised to make it fully open-source in the announcement, and it appeared that it wasn't true. I'm pretty sure that there is nobody complaining that what got open-sourced is now open-source.

But I needed to:

- compile my own kernel
- boot and handle kernels (GRUB)
- handle GIT -> bisecting, remote branches, unbootable kernels
- handle patching
- report bugs upstream where many other ppl just 'dump' their own problems not even trying to figure out if it's the same issue!
- become acquinted with kernel peqularities like dmesg, (undocumented!) boot parameters

I could just have gone with:

<packagemanager> install nvidia-drivers

And be done with it. Furthermore, bugs can be reported to nvidia, the driver includes a dedicated script to aid in just doing that!

Of course - but do realise that NVIDIA has to do the exact same thing when something is reported to them. So they have a lot less motivation to do so, and thus not all bugs get solved as fast.

And as for nVidia's local root exploit, they are also in the linux kernel. Which is opensource. One of which, initially, went unnoticed... I tested that myself. As a normal user, executing the testcase gave me an instant root shell (wtf).

And again, of course, but the fact that it's open-source means that you can patch it yourself, and not wait for someone else to do so. That means it gets resolved much faster.

[crazycheese]

The general idea in this community is that opensource will solve that all on it's own (it's free!!). I'm suprised that ppl are dissapointed by the lagging support for this hardware. And are asking to donate money/manpower.

Proprietary software is being treated as it's the devil himself. Yet nVidia's driver is the best proof that proprietary works with opensource software.

So, you happy with your nvidia driver? Cool. I am not. Now leave the thread and take your useless posts with you please.


@GreatEmerald & Co: please do not waste your time on pointless speeches.
This guy apparently has no idea.

+ Nvidia driver works, it is best proprietary supported driver and it brought 3D to Linux.
- But its closed source and conflicts with GPL libre license, written ONLY for corporate customers with minor adaptations to fit general audiency, it lacks features compared to windows driver, it does not use advantages of Linux kernel, it is pain in butt to integrate (and Linux is designed this way for a reason). And when nvidia says its over, its over; when nvidia says - you don't get this functionality - you don't get it.

This is not case with open drivers. Open drivers are free for everybody to improve and extent. And hell yes, a lot of people are ready to pay for them. If you are not - pick other thread to troll, because your arguments are of no use.

Like I said more than year before: Good hardware, good driver or open driver - pick any two of three. Intel, AMD and Nvidia accordingly.
If AMD really understood the situation, they would punch a million dollar development into open driver. But they did not. This is what I HATE about AMD. They were dumb enough to ignore GPU opensource and they are now dumb enough to improve catalyst instead of opensource. But I don't work for them.

Yet, they do not constitute monopoly; they didn't invent HDCP and UEFI - they worked on coreboot instead; and their GPU hardware is very good, and CPUs are sweetly priced and support ECC on desktops.
If I were to pick, I would choose between AMD and Intel, because those companies actually did something to opensource, whilst having own drawbacks of course.
If you are happy with nvidia driver - go use it, no one will say a word. You can return, but when you happiness is over (if at all) any time though.

[crazycheese]

That 'part of the driver should be opensource' is kind of moot: http://www.phoronix.com/scan.php?px=...page=news_item

It's partially opensource -> Ppl are still not happy.

Sorry, but your example - the Raspberry driver, was not "partially opensource". They simply took the whole closed driver stack and stuffed it behind a blackbox layer, while punching hole interfaces and declaring them as open. That is not partially opensource. That is simply fraud.

Opensource radeon driver, for example, needs to load a small closed source firmware blob, in order to drive the GPU.
This was simply the hardware decision for microcode handling and big majority of people are ok with it, because its strictly hardware related and in no way limiting the use, support cases or OS compatibility. And it can be rev'd if there is need (but there is no currently).
Raspberry on the other side - has implemented whole logic in closed source black box. Thats not opensource.

[bridgman]

If AMD really understood the situation, they would punch a million dollar development into open driver. But they did not.

Do you mean a million on top of the >1.5 mil we've already spent ?

This is what I HATE about AMD. They were dumb enough to ignore GPU opensource and they are now dumb enough to improve catalyst instead of opensource.

We have had developers working full time on the open source stack since late 2007... perhaps your definition of "ignore" is different from mine ? What makes you think that we are improving catalyst instead of opensource ?

[blackiwid]

Its not a feature comparsion, I mean do you really believe that opensource drivers are in some way bad because their openess? They are bad because in the other driver are more man-hours involved there it does not matter if its opensource or not. at least no big one...

So that said... I dont understand even your motivation behind promoting here nvidia for his "better" driver. I can only think of 2 reasons or 3... 1. you are a nvidia fanboy and you want that anybody loves nvidia 2. you think all world should use proprietary development modell because its the better model because if you put more money in it you get in some parts better results 3. you get paid for it...


I mean its a different thing if you promote open or free drivers, because you believe in something... you cant believe that closed source is better for the world so why would you promote them, do you think that if we here flame against nvidia they stop supporting linux? They dont release this drivers because they are good neighbers or even a company that cares about their consumer linux customers. They see money in it and they will make that as long as they earn money from it... even if on each linux installation a popup would popup where nvidia would be flamed as worst cancer company ever...

if you dont care about openess of drivers its pointless to compare features of the drivers... and it goes even deeper you would not even care if nvidia would bring you a complete own xserver and a complete own kernel... if it would work good for you... you seem to not care about freedom at all...

so get it, freedom is a feature it self for most people using such software (at least that who know what they are using and what gpl stands for), so its worthless to discuss further. Of course if you care about fps the closed source nvidia driver is better than the opensource radeon driver, but nobody questions that...

[GreatEmerald]

We have had developers working full time on the open source stack since late 2007... perhaps your definition of "ignore" is different from mine ? What makes you think that we are improving catalyst instead of opensource ?

Does that mean that AMD does not improve Catalyst? That is not likely, although I suppose that the shared code with Windows does help a bit there.

Like I mentioned, it would be ideal if there was only one driver, with proprietary extensions if they were necessary. That would make it so that less effort would be needed to maintain both. But, given the size of fglrx, that's probably not realistic, as it would require a lot of restructuring and code review...

[Ancurio]

Do you mean a million on top of the >1.5 mil we've already spent ?
We have had developers working full time on the open source stack since late 2007... perhaps your definition of "ignore" is different from mine ? What makes you think that we are improving catalyst instead of opensource ?

Please, for the love of deities, don't feed the trolls...

On another note, something I never quite understood was how you can provide separate 2D acceleration,
when the openGL model is completely 3D oriented (ie. you have to emulate 2D with orthographic projections etc.).
Could someone explain please? =O

[liam]

Phoronix: Radeon Driver Enables Full 2D Acceleration For HD 7000

A commit to the xf86-video-ati driver this morning by AMD's Michel Dänzer says it enables full 2D acceleration for the Radeon HD 7000 "Southern Islands" GPUs...

http://www.phoronix.com/vr.php?view=MTI2MjY

Speaking of Glamour, I'd like to see some power consumption numbers that compare the various 2d schemes. Particularly, I'd like to see, a W/frame value for the schemes. I'd suspect the intel driver blows everything else away.
Hmm, has anyone else wondered why the intel driver is so much better than the amd driver? The intel team isn't that big (perhaps bigger than the amd team, but not bigger than the amd + red hat team + community contributors). Sure, the intel gpu is simpler, but it performs better than low end discrete gpus from either company, and they keep improving them pretty significently. I wonder if the architechtural path Intel has chosen is the smarter one, or if they are going to require major changes if they want to reach higher performance levels...