"Ask ATI" dev thread

[mitch074]

And now, for more of the same thing...

I installed 9.11. There is a bit of progress: I get an Xorg.0.log file. Too bad, after the system hangs, it contains only corrupted data - so it's no help.

Maybe by Xmas...?

[mirv]

Maybe your card is broken.

[mitch074]

if my card was broken...

Maybe your card is broken.

If my card was broken, it wouldn't work with Catalyst 9.8, or it'd be glitchy. However, it works well with 9.8 (apart from known, documented 9.8 bugs).

I'm open to suggestions, try again.

[mirv]

Don't completely dismiss the possibility. If you can try another one, do so (not that people have spare cards lying around normally). You have no idea what's been done under the hood of fglrx drivers, and if the radeon drivers do the same, and all under different (I'm assuming clean slate) installations, it's kind of high on the list of things to check for.

[mitch074]

Don't completely dismiss the possibility. If you can try another one, do so (not that people have spare cards lying around normally). You have no idea what's been done under the hood of fglrx drivers, and if the radeon drivers do the same, and all under different (I'm assuming clean slate) installations, it's kind of high on the list of things to check for.

I don't completely dismiss the possibility; however, if versions:
9.6;9.7;9.8 work and pass all tests, along with fairly recent xorg-driver-video-ati radeon releases, I would dare say that if 9.9;9.10;9.11 don't work, then it's far more likely it's a regression hitting my model (a Sapphire reference implementation of the earliest RadeonHD 4850 cards, NOT a factory-tweaked one) and not a failed card.

Especially if said card, using solid state capacitors and fed through a brand new 80+ Bronze PSU, now works at far lower temperatures (40-50?C on core, RAM and VRM, instead of 70+) than what it's rated for.

All I can see as 'failed' would actually be my motherboard, which may be getting a bit flakey: it's an Asus A8N-VM CSM with an Nvidia 6150 chipset (PCI-E 1.0 compatible). But then, has anything changed in PCI-E management in fglrx, starting with 9.9? That might be it.

I'll test with 9.11 on WinXP and Win7, but they worked without a hitch with other driver releases.

[baskin]

I have a similar issue, so i suspect that your card is ok as well as mine (two of them).

I'm getting a freeze system as soon as X is loaded on two systems running opensuse 11.2 with HD4670 and catalyst 9.11.

The same systems run fine with catalyst 9.10.

On another system with opensuse 11.1 and HD3200, catalyst 9.11 works fine.

I suspect a drivers problem specific to HD4670 (or more cards) that appear with 9.11 version of the driver.

Also 9.11 works on the above systems on Win7 and WinXP.

[mitch074]

Well, actually, I managed to 'fix' the bug: it seems that the latest fglrx drivers don't like Nvidia chipsets... I replaced my motherboard (along with CPU and RAM) with an AMD 785G-based solution and now, more recent versions of fglrx seem to work.

Now, I understand that Nvidia is a competitor for AMD in chipsets and graphics space, but at the time I bought that previous mobo, AMD chipsets were non existent, and the only way one could use an AMD CPU with good motherboard integration was with an Nvidia chipset... So don't forget us poor 'legacy' hardware users on your regression tests.

[baskin]

Well, actually, I managed to 'fix' the bug: it seems that the latest fglrx drivers don't like Nvidia chipsets... I replaced my motherboard (along with CPU and RAM) with an AMD 785G-based solution and now, more recent versions of fglrx seem to work.

Now, I understand that Nvidia is a competitor for AMD in chipsets and graphics space, but at the time I bought that previous mobo, AMD chipsets were non existent, and the only way one could use an AMD CPU with good motherboard integration was with an Nvidia chipset... So don't forget us poor 'legacy' hardware users on your regression tests.

My motherboard is also with NVIDIA chipset, but it is new and i can't change it.

[LostinSpacetime]

Weird Bug with Skype video

Hello. I was wondering if I'm the only one who notices this effect because I couldn't find anything in the web. My card is HD3200 using the latest Catalyst driver with Ubuntu 9.10 and Skype 2.1.

Ever since I have this computer I have a this weird bug. In the video call window, when my image is displayed over the received image, the whole line (with the height of my image) is replaced by another line from the received video. See pic:



Anyone encountered the same behaviour?

[mirv]

My motherboard is also with NVIDIA chipset, but it is new and i can't change it.

I have an nvidia chipset on board (nforce3), but don't have issues with it and fglrx. I do, however, use the forcedeth driver as a module instead of compiling it directly into the kernel (old bug that I'm not sure affected just me or not, but it never worked properly if I compiled it in).