ATI Radeon HD 4850 512MB

[phoronix]

Phoronix: ATI Radeon HD 4850 512MB

Last week we exclusively shared the steps AMD was taking to make an evolutionary leap in Linux support with same-day support for their brand-new Radeon HD 4800 series, Linux drivers shipping on the product CD, some manufacturers showcasing Tux on the product packaging, and their proprietary Linux driver reaching a feature parity with their Windows driver. We had also shared that the Radeon HD 4850 works with open-source xf86-video-ati driver since day one. Now that we have had time to complete testing of the Radeon HD 4850, today we are sharing the first Linux results from this brand-new ATI graphics processor. Before you think the Windows and Linux performance is equal for the Radeon HD 4800 series, this isn't the case, at least not yet.

http://www.phoronix.com/vr.php?view=12534

[satan0rx]

Edit: and I know I'm comparing apples to oranges, since Intel's IGP will never do more than 2D in Linux, while ATI's open source drivers should offer 2D, 3D, possibly hw video decode and a perfect suspend/hibernate experience in the future. It's just that getting updated about suspend/resume advances can be as important as updates on 3D or hw video decode.

ebbeh??!?
I got an Intel based notebook with gl960 (= stripped gm965) and it does 3D. compiz fusion works (albeit xv + compiz fusion does not ...), and I can play quake III at 1280x800 ~40fps. It is too slow for nexuiz, ok,
but it does 3D (if you are not a hardcore gamer)

[miles]

Yes, that is true, but in general, the suspend/resume issues are triggered not by the Hardware, but by the SW ecosystem described above.

Also note that some of the suspend hacks (like reposting cards, etc) will may work on some cards and not. But as I mentioned before, they are mainly hacks that should generally be avoided.

One reason people using Linux tend to prefer Intel's IGP is because the suspend/resume works in most (all?) distributions without having to hack it in yourself. Oter IGP and dedicated hw can often resume, but to a blank screen or a messed up one. Having similar success as Intel's IGP would be an important feature to know before purchasing the hardware

Edit: and I know I'm comparing apples to oranges, since Intel's IGP will never do more than 2D in Linux, while ATI's open source drivers should offer 2D, 3D, possibly hw video decode and a perfect suspend/hibernate experience in the future. It's just that getting updated about suspend/resume advances can be as important as updates on 3D or hw video decode.

[Kano]

Try removing the file or using

dpkg-reconfigure xserver-xorg

for some language selections. For standard xservers you don't need much in there.

[leo11877]

Usually you use -p1.

I think that worked. Here are the steps that I followed

sudo apt-get install build-essential git-core configure-debian automake autoconf xorg-dev libtool

git-clone git://anongit.freedesktop.org/git/xorg/driver/xf86-video-ati

cd xf86-video ati
patch -p1 < ~/rv770-id.patch
./autogen.sh --prefix=/usr/
make
sudo make install
make clean

sudo nano /etc/X11/xorg.conf

find the device section. In the line "Driver" substitute "ati" for whatever is there now....i.e., vesa or fglrx

However, when I made changes to XORG file and rebooted, the screen turned white and I had to reboot and come into Recovery mode to reset the xorg.conf file.
One thing I noticed is that after I reset the file, I can now use full 1280 X 1024 resolution. However, ATI catalyst control center still doesnt recognize the card or start. Can someone share how to find if the patch worked? and also send the Xorg.conf file contents?

Thanks
Leo

[Kano]

Usually you use -p1.

[Michael]

Leo:

What's your current directory when running the command? It should be one directory lower than xf86-video-ati. Otherwise just type the relative path to those files.

[leo11877]

Hi
I am trying to patch and build the Open source ATI driver for my new 4850 card, as per Michael's instructions. When run the command

patch -p0 < ~/rv770-id.patch

this is the output
can't find file to patch at input line 4
Perhaps you used the wrong -p or --strip option?
The text leading up to this was:
--------------------------
|diff -Naur xf86-video-ati.orig/src/ati_pciids_gen.h xf86-video-ati/src/ati_pciids_gen.h
|--- xf86-video-ati.orig/src/ati_pciids_gen.h 2008-06-19 14:36:59.000000000 -0400
|+++ xf86-video-ati/src/ati_pciids_gen.h 2008-06-19 14:38:40.000000000 -0400
--------------------------
File to patch:


I got the git code from

git://anongit.freedesktop.org/git/xorg/driver/xf86-video-ati


The patch file is on my home dir, the code is also at the same level (in the xf86-video-ati folder)

Any help would be highly appreciated.

Thanks
Leo.

[brouhaha]

Feature parity?

Feature parity? Does that mean that fglrx finally supports screen rotation?

[mtippett]

True, but if the machine he's using suspends and resumes nicely already, changing the graphic card with the model he wants to test (an HD4850 or 4870 here) wouldn't it give a fair indication about the card itself?

Yes, that is true, but in general, the suspend/resume issues are triggered not by the Hardware, but by the SW ecosystem described above.

Also note that some of the suspend hacks (like reposting cards, etc) will may work on some cards and not. But as I mentioned before, they are mainly hacks that should generally be avoided.

Regards,

Matthew