AMD Catalyst 7.12 Linux Driver -- The Baby's In Surgery

[happycampers]

They've got the source code. They know.

All kidding aside, just because you have the source code doesn't mean you know how it all works. I've seen plenty of projects that were poorly documented and some poor devs had to come in and try to make sense of it all -- not too easy when you are talking about hundreds of thousands or millions of LOC. Most devs I know can't hold that information in their heads, so they might be an expert at one little piece of the puzzle, but not the whole thing. I would suspect, judging by the regressions, that AMD devs are now paying the price for previous underfunding and underdocumenting.

Just because they're unable to fix things quickly (yeah, that's it...) doesn't mean they're incompetent or idiots.

I was being a little tongue-in-cheek, but again, if they are dealing with an under-documented mess, it could very well be that the devs don't even know how their own changes will behave under nominal settings, much less how the driver will behave when users start poking around in their xorg.confs. I'm not talking about speed of development here; it's the number and scope of regressions that would indicate that even the devs don't have enough information to consistently predict how their changes are going to impact overall behavior

The devs aren't the problem. AMD not moving fast enough to clean up ATI's dirty laundry's the problem.

I agree, but that's easier said than done...

[Redeeman]

amd really should start pulling themselves together, they are on really serious thin ice.

everywhere around me, people are cursing the ati drivers.

i JUST thought, with the new stuff in fglrx and such, that i might be able to buy ati now, in wait for the free drivers, however i realize now that this will be impossible. With nvidia, i know what im getting into(with instability and shit), with amd i have NO clue how bad they are gonna screw me over - even if i expect it to be bad, all i need do is come here, and realize its alot worse than one thought possible.

and this isnt just affecting their graphics card business, as a result, hell will pretty much have to freeze over, dry up, and then freeze over again, before there will sit a new AMD cpu at my desk, and i have _NEVER_ bought intel before.

i really wish intel would throw out a graphics card capable of duallink dvi - i dont really care about performance as long as it does good xv, and has basic 3d support.

and while my sales alone arent gonna affect them, im the person _ALOT_ of people are asking for advice when in need of both entirely new computers, graphics cards and such, so i do affect abit more than just myself.

[LucasG15]

Dang it! 7.11 and 7.12 both give me blank screen screen when X attempts to start. Additionally, the computer becomes unresponsive. I have to do a cold reboot. Initially, I used the easy method by using the graphical installer than tried the method 2 which involves create the packages. Both ways had the same results. The only thing that works right now is the open source RadeonHD driver. (just updated to 1.10 w/o problems)

I have a sapphire X1950pro with ubuntu 7.10.

[smlbstcbr]

I support the motion about what should xorg.conf have, to avoid such problems, however, it may be handy to us all to register in AMD customers service and to flood them with our feedback, it will be really useful. Devs will not look at this forum, so lets show them by the (relatively) long way (a couple of minutes registering won't hurt anybody)

[Kano]

That driver has beta status when you call it friendly. None of the bugs I found have been fixed, so for me it is no bugfix release. The release notes are even missleading, not 1680x1050 is not supported the driver has problems with non-vesa modes. Sometimes EDID info could be used as Michael mentions for his display, but there must be a huge bug in the EDID use in this driver. Now the xmas break will come and I guess nobody will work on it till next year. So maybe I even switch back to 7-11 as default driver (and 8.39.4 for firegl), not sure about that now. Or maybe ATI could prepare a bugfix for the bugfix release and call it even 7-12 My script is prepared for any nameing scheme now *g*

[hobophobe]

I don't blame your developers. I blame your management. A soldier can only be as good as his orders.

My next computer will have integrated Intel graphics, which means it will have an Intel processor. Sorry nvidia & AMD/ATI. If you ever get your acts together I'll buy another graphics card.

I'd just rather not have to worry about whether my computer will be functional next week, or if, instead, some manager is too busy staring at his secretary's thighs and thinking about his holiday bonus check to make sure his customers get quality drivers.

[butdie]

sounds like ati is making fun of developers too

their releasing cycle, their random version numbers

so, it's more than just making fun of users

That driver has beta status when you call it friendly. None of the bugs I found have been fixed, so for me it is no bugfix release. The release notes are even missleading, not 1680x1050 is not supported the driver has problems with non-vesa modes. Sometimes EDID info could be used as Michael mentions for his display, but there must be a huge bug in the EDID use in this driver. Now the xmas break will come and I guess nobody will work on it till next year. So maybe I even switch back to 7-11 as default driver (and 8.39.4 for firegl), not sure about that now. Or maybe ATI could prepare a bugfix for the bugfix release and call it even 7-12 My script is prepared for any nameing scheme now *g*

[sloggerKhan]

1680x1050 ????

Looks like I'm going to be reverting to last version of fglrx... WTF WITH NO WIDESCREEN SUPPORT?????

[yoshi314]

ati should just release the specs and give it up. get their devs to help with radeonhd instead.

why embarass yourself further with such awful drivers. this is seriously getting beyond unreal. even crappies chinese/taiwanese vendors write better drivers for their hardware.

[pedepy]

this is seriously retarded..

I dare one ati dev to come up and explain how this 1680x1050 even occured in just a month. This just doesnt make any sense. Whatever code you guys are working with must be horribly written.

I was gonna try 7.12 but I'm not even gonna risk getting it even close to my box. No way..

AMD hello?....

(oh and this is really starting to weigh down on Phoronix & Michael's credibility ... remember that Gamespot story from a couple weeks back? Read the articles about new driver releases and you'll see what I'm talking about.....)