ATi Support on infinityOS

[Armin]

Yeah, I know it's an attractive idea, but the interfaces are *really* far apart.

Even something like this probably took a lot of work, and the interfaces were a *lot* easier - basically fuel, power to the engine controls, throttle, and a big red button for the (apparently home-made) afterburner. The T58 was converted from turboshaft to turbojet apparently, so no need for reduction gears or complicated drivetrain.

That's what I meant! I'm a mechanical engineer, that's how we work! Also I (still) love to play with lego, so I naturally tend to such solutions.

[darkphoenix22]

Eh, in the short term though (and this is something you can do without code), ATi should reposition the OSS Radeon drivers as the "desktop" Linux drivers and fglrx as the "workstation" Linux drivers.

This will stop all the requests for ATi to open source the fglrx drivers and will stop distros like Ubuntu from encouraging users to install the fglrx for purposes which they were not made for. You will also look much better than Nvidia in the process, for having an "official" open source Linux driver, even if it isn't complete yet or contains all the features you want.

Your users will also be more comfortable using the OSS Radeon drivers if they are the "official desktop Linux" drivers. I know I made some assumptions about them just because they are open source and default, mostly because of the experiences I had with the "open source" nv drivers.

You should tell distros to make a clear distinction between the target user bases and feature sets of both drivers (and for godsakes get Ubuntu to stop letting users install fglrx through Jockey :P).

You'd be surprised at how much simple pretty words and press releases can accomplish.

[Kano]

I really don't get what you have got against fglrx for video, when you use mplayer vaapi thats the only way to get video accelleration. gl output is fine for low bitrate but not for high bitrate like bluray m2ts. be sure to use vsync on in amdcccle. xv is just not ideal for ati, but thats known for long time. bitrates below 15 mpbs should be no problem for software mode with many cpus, but above 30 this is different.

[Kano]

Btw. if you have got compiz problems just get rid of it, like

metacity --replace &

or use the "official" way

[RealNC]

Kano, you mentioned "it works OK if you use that and then use this but then configure with that and this as a workaround..."

The problem is that it's fglrx that needs to change, not the software we use. fglrx must adapt to the Linux Desktop, not the other way around.

I hope you can see the problem here.

[bridgman]

We're still a few months away from the open source drivers being a viable alternative to fglrx for all of our GPUs, but that has been the plan since we restarted open source driver support in 2007 :

How will the existence of a full-featured open source driver affect the goals of the Catalyst Linux package? Will it continue to be a comprehensive driver package, or will it become more focused on workstation applications?

Combined with next question. See below.

Will the new open source driver ever completely replace Catalyst on Linux? How will the two cooperate, and will AMD do anything to inform customers of the choice?

I see the two drivers complementing each other. Having an open source driver makes it possible for distributions to keep the graphics driver in sync with their own development efforts, provide a great "out of box" experience, and to fully support their corporate customers.

The Catalyst Linux driver (fglrx) will focus on released and stable OS distributions, offering ISV-certified support for workstation users and providing a feature and performance upgrade for consumer users. OEMs who pre-load Linux on their products will use fglrx

In general, consumers will have access to both of these drivers and will be able to choose the one which best meets their needs.

http://www.beyond3d.com/content/interviews/43/2

Between then and now the open source driver community has made huge progress, going from a "2D mostly" userspace stack to kernel modesetting, GL 2.x, Redirected Direct Rendering, and a bunch of cool TLAs

The spring 2010 distro releases will be the first to really make use of the new stack, although Fedora (and probably others) have been shipping with KMS enabled for a year or so in order to prove out the new technology.

[darkphoenix22]

I honestly don't think desktop Linux users should be using fglrx as they clearly were not and are intended for desktop use.

The Radeon drivers work great for Linux desktops and have support for 3D and Mesa in the latest versions.

[darkphoenix22]

We're still a few months away from the open source drivers being a viable alternative to fglrx for all of our GPUs, but that has been the plan since we restarted open source driver support in 2007 :

http://www.beyond3d.com/content/interviews/43/2

Between then and now the open source driver community has made huge progress, going from a "2D mostly" userspace stack to kernel modesetting, GL 2.x, Redirected Direct Rendering, and a bunch of cool TLAs

The spring 2010 distro releases will be the first to really make use of the new stack, although Fedora (and probably others) have been shipping with KMS enabled for a year or so in order to prove out the new technology.

Ah you need to get ATi to make an epic announcement. Get it on Slashdot and Digg. And be clear to Ubuntu that the fglrx drivers are not intended for desktop use.

[bridgman]

Kano, you mentioned "it works OK if you use that and then use this but then configure with that and this as a workaround..."

The problem is that it's fglrx that needs to change, not the software we use. fglrx must adapt to the Linux Desktop, not the other way around.

I hope you can see the problem here.

I don't think Kano is suggesting for a moment that fglrx is perfect and never needs to change (in fact he's the last person I would expect to say that ) but if the driver does provide the functionality you want when used "this way" then why not use it ?

I don't think there is a stone tablet anywhere saying that Xv is "the one true video output to rule above the rest", however it has been popular and was broadly implemented even on drivers or GPUs which did not have a working GL implementation. As overlay hardware gets replaced with shader hardware the trend is from Xv to GL, but I would like it if we supported both well. Today we support GL better than Xv. I recognize that not all players can work with GL today, but if the option is there for the using then why not use it ?

[bridgman]

Ah you need to get ATi to make an epic announcement. Get it on Slashdot and Digg. And be clear to Ubuntu that the fglrx drivers are not intended for desktop use.

We need to get caught up with new GPU support first. Patience, grasshopper.

Remember that your desktop needs may be very different from someone else's desktop needs, and so while fglrx may not be suited for *your* desktop needs it may actually be the driver of choice for another user's desktop.

The important thing is that we close the gap between the drivers so that users don't feel they need to switch between drivers (or OSes) to do all the things they want. We're pretty close to that point but not quite there yet.