Radeon HD 4850 Works With Open Driver

[phoronix]

Phoronix: Radeon HD 4850 Works With Open Driver

In our article this morning entitled AMD Makes An Evolutionary Leap In Linux Support, we briefly touched on the fact that AMD would be continuing in their open-source support for the Radeon HD 4000 generation of GPUs. AMD will release register information for the RV770 and the code to TCore and KGrids will help the community developers when it comes to 2D and 3D acceleration. Though, with not much work at all, this afternoon we have the ATI Radeon HD 4850 512MB running with the open-source xf86-video-ati driver! The Radeon HD 4850 will not currently work with the xf86-video-radeonhd driver as it doesn't rely upon the AtomBIOS in these areas, but in this article we have enclosed our patch and other information for using this open-source driver on AMD's latest hardware.

http://www.phoronix.com/vr.php?view=12503

[d2kx]

Can't wait to get one of those cards

[Chewi]

I'm not particurlarly for or against AtomBIOS but this certainly highlights the difference between the two drivers. I suppose it might not be a priority for them but it'll be interesting to see how long it takes for radeonhd to at least do this much.

[Louise]

I think it have come down to a prestige question for Novell/Sues now, not to use AtomBIOS.

[some-guy]

I wonder if using AtomBIOS removes stability, if it does then radeonhd will probably be more stable in the long run

[NeoBrain]

I wonder if using AtomBIOS removes stability, if it does then radeonhd will probably be more stable in the long run

It's actually very stable and well tested (ATI didn't ever have to do a revision on it AFAIK), which is why I don't understand the point in NOT using it.
However, most people probably just see the problem in AtomBIOS not being Open Source and thus not "free".

[Melcar]

It's actually very stable and well tested (ATI didn't ever have to do a revision on it AFAIK), which is why I don't understand the point in NOT using it.
However, most people probably just see the problem in AtomBIOS not being Open Source and thus not "free".

Yeah, I think that's the main issue the radeonHD guys have with it. However, it's obvious that it allows for greater, shall I say, "simplicity". At this rate, seeing as how radeon is growing faster than radeonHD (and that it supports more cards), I don't see why anyone would choose radeonHD over radeon, unless it offered some crazy extra features or even better 3D performance.

[TechMage89]

In theory atombios is slower than native modesetting, because the instructions have to be run through an interpreter on the card, but that rarely matters, because modesetting doesn't really need to be fast. Also, even with atombios, it looks like sometimes errata on various cards need to be dealt with anyway. That said, I think the simplicity and compatibility of an atombios-based driver is a huge advantage.

I hope someone gets around to writing an atombios parser for the kernel! Kernel modesetting would be awesome.

[Swoopy]

Does anyone know if an AGP RadeonHD 4xxx card will be forthcoming?

[d2kx]

I really don't think so.