AMD 2010 Catalyst Driver Year In Review

[Apopas]

I probably have to go through my .config file to see if there's anything strange in it, I might have missed something when updating from .35.

Probably. I made a clear installation and maybe that's why I avoided problems.
BTW even hibernation and suspend works like a charm. For example my laptop has uptime 1.5 day now only because I had to power it off 2 days ago since I didn't have along my charger.

[evolution]

I really liked most of this year's Catalyst Releases, although I've to say since Catalyst 10.8, some / most people who've a r600 card with a UVD1 decoder can't decode H264 videos properly anymore, only VC-1 ones... (On the Windows side, it's even worse and I can't decode both H264/VC-1 encoded videos without using about 50% of my CPU (which corresponds to 100% usage in a single-threaded application)).

By another side, currently we've some 'patched' support for newer kernel releases (2.6.36/2.6.37), latest stable Xorg releases are working with recent Catalyst versions and it even supports my HD4650 AGP card out-of-box without applying the "AGP patch", like it happens on Windows.

Cheers

[Azpegath]

I really liked most of this year's Catalyst Releases, although I've to say since Catalyst 10.8, some / most people who've a r600 card with a UVD1 decoder can't decode H264 videos properly anymore, only VC-1 ones... (On the Windows side, it's even worse and I can't decode both H264/VC-1 encoded videos without using about 50% of my CPU (which corresponds to 100% usage in a single-threaded application)).

By another side, currently we've some 'patched' support for newer kernel releases (2.6.36/2.6.37), latest stable Xorg releases are working with recent Catalyst versions and it even supports my HD4650 AGP card out-of-box without applying the "AGP patch", like it happens on Windows.

Cheers

Seems like I perhaps could go back to Catalyst every now and then, whenever I play games =)
Perhaps Minecraft fares better with Catalyst than video-ati-drivers.

[AdrenalineJunky]

Really seems like those benchmarks are missing some context... like, what settings were used, did unigine take advantage of the opengl 4 features when used with a compatable driver? Was tessalation enabled when used with a compatible driver (for the record I read kano's assessment, just saying.. it should be in the review).

Benchmarks without the proper context loose a lot (if not all) of their meaning.

[deanjo]

Perhaps Michael can update the PTS suite to report the settings used on the various tests.

[Michael]

Really seems like those benchmarks are missing some context... like, what settings were used, did unigine take advantage of the opengl 4 features when used with a compatable driver? Was tessalation enabled when used with a compatible driver (for the record I read kano's assessment, just saying.. it should be in the review).

Benchmarks without the proper context loose a lot (if not all) of their meaning.

You can find out all of the settings by running the phoronix-test-suite yourself or looking at its package settings. Simple as that, totally reproducible.

[Michael]

Perhaps Michael can update the PTS suite to report the settings used on the various tests.

You can find them on OpenBenchmarking.org when publicly available.

[deanjo]

You can find them on OpenBenchmarking.org when publicly available.

It would be nice if they were embedded into the graphs on the bottom of them.

[Michael]

It would be nice if they were embedded into the graphs on the bottom of them.

With the OpenBenchmarking.org embeddable graphs its a matter of being a click away from it.

[deanjo]

I'm talking about something like this Michael, rough mockup: