AMD Catalyst 11.5 Linux Driver Released
Phoronix: AMD Catalyst 11.5 Linux Driver Released
The AMD Catalyst 11.5 Linux driver was released yesterday afternoon. Catalyst 11.4 was only released about two weeks back, but the monthly Catalyst update for May has already arrived, so does it not bring much?..
I updated my script yesterday and this driver will on the soon to be released Kanotix 2011-05 iso as well (even for live!!!).
Did not test it with U, but should work when amd did not break the packageing
Thanks Kano, I was waiting for it xD
I would be really surprised if this version works with my card. The last version which was able to start X without crashing was released at least a year ago.
I am using the 11.4 driver currently with a Radeon 4870. I recently started using this card again in one of my secondary machines. It had previously been my main card but the drivers were absolutely horrible at that point. It's amazing how far they have come. The desktop is fast and responsive and tear-free is a welcome addition. Yes, they should have had it all along but I will take late over never. Monitor detection is no problem and the Catalyst Control Center has seen some great improvements.
The one thing I can't get working is the HDMI audio out. The card outputs to DVI and uses a supplied converter to attach the HDMI. I don't know if this is a driver issue as I never tried to get it to work in the past. I am running Ubuntu 11.04. Aside from the audio issue, which I plan to investigate with my Radeon 5570 the Ati performance has been terrific.
I plan to test the 11.5 drivers when they hit the X-Swat PPA. I'll update this post when I have them installed.
KDE4 effects with fglrx
I hope it resolves the problems with KDE4 desktop effects. I recently (3 days ago) bought an AMD 6850 (after years with Nvidia cards) and if I activate KDE 4.4.5 desktop efffects in my Debian testing with fglrx 8.84.5 (fglrx-glx 1:11-4-2 package from unstable) sometimes the windows and menues shows empty...
Well suspend/resume seems to be working again with this version of catalyst, that's certainly a welcome surprise. ATI whatever you do, please do not regress this feature again, ever, thanks!
First off, I'm not trolling, I'm seriously curious as always...
Originally Posted by malkavian
I don't understand how these bugs happen, I mean aren't drivers put through a series of tests (just like Firefox has for their browser tests) to make sure that all the types of requests for the GPU are being done and being done properly? Since if this was a KDE bug, it would be showing in all drivers, so it just surprises me when this or that app reveals a bug in a driver. It's almost like the policy is, "We know there are probably lots of bugs in our drivers, but until some user complains or a specifc app which utilises one of our untested code paths breaks and reveals a bug, we'll just hope we won't have to fix them".
Does AMD just need more testers, or perhaps they don't have a gauntlet set up to run their new drivers through which tests every feature? Or is it because Linux doesn't get them enough money to have testers, but Windows does?
Really curious as to what's wrong, but regardless, I hope more focus is put on the open drivers anyway. The open source AMD drivers are getting quite good, and even the open source nvidia drivers are doing fairly well, though I'm pretty sure the open source AMD ones are a lot more advanced right now. I was shocked to find 1080p fluid video playback under an older nvidia card with the open source drivers, though.
@Yfrwlf: What do you think of this?
Originally Posted by AMD
Nice irony, huh?
All I know is that you shouldn't need tweaks for specific apps. If a specific app needs a tweak, it's the app's fault for being broken, not the driver. You don't see CPUs having lots of random glitches with computations, it seems a lot more cut and dry. The CPU can do X, Y, and Z, and here we tested it for doing X, Y, and Z with our driver, and it all works, here you go, a perfect driver....but with GPUs, it just seems like a total mess. Like with CPUs, if there's a problem found with one of the types of requests made through the driver to the card, it should be caught there, on a low level. I don't understand then why something would only be found out by a bug appearing in a much higher level in the application stack, like KDE Qt panels turning white, and it being a problem in the driver. That just seems out-of-control to me, but maybe Xorg drivers are out-of-control because Xorg is?
Originally Posted by alazar
I know I'm not the only one wishing drivers were simpler though. ^^