HD 6990 Artifacting on my X58 MB; AMD acknowledges issue

[happy medium]

Suggests your X58 MB is starting to show its age and cannot deliver STABLE current over the PCIE lanes at max load

One way to test that would be for him to put the gtx590 in there.
If it runs the 590 it will surely run a 6990.

Apoppin?

 

[apoppin]

One way to test that would be for him to put the gtx590 in there.
If it runs the 590 it will surely run a 6990.

Apoppin?

You haven't been reading my posts.

i have been running (since the HD 6990 failure) the GTX 590, the HD 6970 and HD 6970 CrossFire. And to double-check stability, i put my CPU overclock back to 4Hz and even overclocked my cards.

Not a hint of instability with ANY other graphics configuration. But stick the HD 6990 in and it artifacts on the desktop - with my system completely at stock. Stick it in the Phenom II system and it works.
:|

 

[happy medium]

Sorry I missed the 590.

Well in that case ,its not a power delivery problem.

edit:
"And everyone seems to forget that the HD 6990 worked for months in the same MB."

There are 3 things left that could be causing this problem.

1.drivers-, new chipset drivers or new video drivers
2.windows update
3. have you flashed to a newer bios recently?

 

[apoppin]

Sorry I missed the 590.

Well in that case ,its not a power delivery problem.

No worries there is a lot in here. i have been pulling out my hair for over a week trying to figure this out.

And power delivery issues would show up in a 3D app; not on the desktop when power draw is a minimum. And don't forget, AMD said "this happens sometime"

i believe it all adds up to a defective card. Whether it is common or not, only AMD and their partners know.

But i am going to add to my troubleshooting list for video cards (in general) when all else fails = to try it in another MB

 

[Ghiedo27]

Is this something that's creeped in with a driver update? It would be weird that it would discriminate against Intel, but maybe it's something that corrupts over time? Just throwing something else out there. Have you tried a relatively old driver?

 

[apoppin]

Sorry I missed the 590.

Well in that case ,its not a power delivery problem.

edit:
"And everyone seems to forget that the HD 6990 worked for months in the same MB."

There are 3 things left that could be causing this problem.

1.drivers-, new chipset drivers or new video drivers
2.windows update
3. have you flashed to a newer bios recently?

Not 1 nor 3; but who can say about Windows updates?

Is this something that's creeped in with a driver update? It would be weird that it would discriminate against Intel, but maybe it's something that corrupts over time? Just throwing something else out there. Have you tried a relatively old driver?

Well, as you know, when you put in a new video card, Windows picks the drivers. So i have had Cat 11-8 on them. And when i put HD 6970 back in, i installed 11-4 (i believe the HD 6990 launch drivers) several times so it would be the default drivers that windows installed.

i also ran 11-4 with my Phenom II system and then installed 11-8 just to make sure the new drivers had no issues. They were fine in that PC.

i ran HD 6990 last successfully with Catalyst 11-6 - then it got put away for a couple of months while i was doing my 3D Vision evaluation.

 

[Ghiedo27]

Sorry, I'm a Win XP caveman :awe:. You can't disable the auto update and force 11.6?

I'm shooting spitballs here, but it just seems plain weird that you're having a Vram type problem with a dual frame buffer card only in a triple channel memory MB that manifests (apparently) right when the driver installs. It just screams memory optimization gone bad to me.

Again, I'm just throwing stuff out there. I'm going to go back to watching UFO conspiracy documentaries.

 

[apoppin]

Sorry, I'm a Win XP caveman :awe:. You can't disable the auto update and force 11.6?

I'm shooting spitballs here, but it just seems plain weird that you're having a Vram type problem with a dual frame buffer card only in a triple channel memory MB that manifests (apparently) right when the driver installs. It just screams memory optimization gone bad to me.

Again, I'm just throwing stuff out there. I'm going to go back to watching UFO conspiracy documentaries.

Let's put it this way, it doesn't work with 11-4, 11-6 nor 11-8; i didn't bother with trying 11-5 nor 11-7. It used to work fine with 11-4 and 11-6, so that is what i tried; and 11-8, of course.

If there were memory optimizations gone bad on my Intel PC, wouldn't you say it would manifest itself with my HD 6970 CrossFire? ... or my GTX 590 perhaps?

imho it points to my HD 6990. And my best guess is that its in the power regulating ICs as someone on ABT forum threw out.

But seriously, your guess is now as good as mine. i have never ran across anything like this and i thought i'd share. After more research, it may become an article. Unfortunately, it doesn't appear that i will get the chance to try another HD 6990 in my X58 MB. That would settle it but AMD hasn't got back to me in 10 days.

Maybe i'll run across someone that has issues with their HD 6990 in his AMD MB and we can trade.

 

[skipsneeky2]

Had a 6990

While it could run completely stable while stress testing it for hours without issues.

Moment i gamed problems came up and this was on a msi x58 mobo with a i7 950 so i do believe the Op and his email.

Either the drivers or the card itself locked up my system and it wouldn't boot even in safe mode so i was forced to reinstall windows and the next day i just returned the card and got my money back.

 

[apoppin]

Had a 6990

While it could run completely stable while stress testing it for hours without issues.

Moment i gamed problems came up and this was on a msi x58 mobo with a i7 950 so i do believe the Op and his email.

Either the drivers or the card itself locked up my system and it wouldn't boot even in safe mode so i was forced to reinstall windows and the next day i just returned the card and got my money back.

Well, my job is to work on these issues. i am still thinking that there may be some way to reset as AMD suggested since the HD 6990 works in the Phenom II system.

i'll try to fool it and run it overnight in my Phenom II system and then put it in my Intel PC on a fresh install of Window 7, RAM with different timings, and the release drivers for it. And of course, i'll switch out cables and the PSU again.

That is about all i am going to do for it.... next up, a 16 gauge single shot.
:biggrin:

i'll update you if there is anything new. Meanwhile i will look for suggestions here. And now i got to get back to writing while Win 7 sets up. Thanks for your time, your suggestions and your interest
:thumbsup:

 

[hdfxst]

maybe i missed it but is this a clean install?

 

[busydude]

You might have a point - except that it artifacts on the desktop the second Windows sets up the drivers for it.

Can you just pop in a Linux live CD and check if that boots OK?

 

[apoppin]

maybe i missed it but is this a clean install?

It is the same install that worked the last time it was in my PC two months ago. Nothing changed and nothing changed for any other graphics card.

As i mentioned in my last post, as a LAST resort, i will do a clean install of Win 7 on the Intel PC (using my HD 6970) while the HD 6990 runs all night keeping my room warm and looping Heaven in the AMD PC.

i will use a different brand of RAM with different timings than my Kingston and i will also change out the PSU (again) and the cables. Then i'll install my HD 6990 into the X58 MB and report back here.

And while all that is going on, i will be working on my 3D Vision evaluation on my notebook till i pass out

Can you just pop in a Linux live CD and check if that boots OK?

It will load up Windows - until it installs the driver. Then it artifacts badly. Works great in the BIOS and probably in safe mode. However, this clean install of Win 7 may tell me something. Perhaps there is some conflict.

 

[hdfxst]

do you have afterburner and trixx installed?This gave my 5850 nice little squares all over the desktop.I mined for 3 months and thought i killed the card

 

[Silverforce11]

Disable Aero or any fancy GUI effects, prolly Aero boot up on load causing a spike in GPU load.

It may work with the 590 but who knows how they are configured to draw power, max from MB or max from 6/8 pins first?

A VC cannot be defective if it works fine in another system. Doesn't make any sense.

 

[3DVagabond]

imho it points to my HD 6990. And my best guess is that its in the power regulating ICs as someone on ABT forum threw out.

So, why doesn't the fault in the power regulators show on your AMD board? Sounds like some sort of incompatibility, not faulty hardware. Faulty hardware would always be faulty. Not referring to intermittent shorts, which this obviously isn't either. It could be a particular combination of a few things that trigger it.

 

[apoppin]

So, why doesn't the fault in the power regulators show on your AMD board? Sounds like some sort of incompatibility, not faulty hardware. Faulty hardware would always be faulty. Not referring to intermittent shorts, which this obviously isn't either. It could be a particular combination of a few things that trigger it.

Who knows?

i set up the clean install (no overclocking programs installed; minimal system). Transferred the HD 6990 from my Phenom II where it works fine, to my Intel PC and booted it up. It worked again(!) and scored normally in Vantage and looped Heaven at the same mid-80s C that it ran in the Phenom II system

Upon a reboot - back to artifacting.
:'(

So basically, there is no difference. i can move the card to the Phenom II system and it will work. Move it to the Intel PC (even my same 'old' HDD) and it will "reset" and function apparently normally. Reboot it and it BSODs.


This is a first for me. i am unsure what to do. Maybe i will just retire the card until i upgrade to Ivy Bridge this Winter - and retest it then. Although i am tempted to grab the shotgun (but what if it is a incompatibility that crept in somehow just with my MB?).
:hmm:

A VC cannot be defective if it works fine in another system. Doesn't make any sense.

And a MB cannot be defective if it runs every other VC perfectly - except for one. i am agreed - it makes NO sense.

i am sorry but i do not want to spend any more time on this issue; i am so tired that it hurts ... and now i begin my day. If AMD ever responds, or when i get my new Intel CPU/MB, i will resurrect this thread.
:|

 

[zebrax2]

Maybe the card bios?

 

[apoppin]

Maybe the card bios?

i asked AMD about updating its BIOS but they didn't reply
--now that i think about it .. why the heck didn't i think of this before ? Good idea.
:thumbsup:

Yes, i'll flash it to a newer HD 6990 .. maybe to one with a warranty
:biggrin:

Give me a day or so to recover. i still got this massive article to complete and a headache to match it.

Any suggestions for "best BIOS"? Definitely not a reference AMD 6990 BIOS. Probably one of the latest would be best.

 

[biostud]

try running computer without overclock?

 

[apoppin]

Of course.

The next thing i am going to do is update the card's BIOS. If it still doesn't work in my X58 MB, i will switch it out for X68 when Ivy Bridge is released. And i'll retest then. It could just be some weird incompatibility.

 

[happy medium]

Mabe try just installing just the video driver without CCC?

 

[Chiropteran]

AMD knew exactly what i was talking about and pointed me FIRST to changing my MB

Sure, I don't dispute that. What I dispute is your interpretation.

You say: AMD knows the 6990 has issues in some motherbaords, so it's AMD's fault!

I say: AMD knows, because it's a known flaw with some motherboards when running their card. Maybe the nVidia cards that draw similar power don't exhibit the problem because they include a work-around to compensate for these faulty motherboards. The motherboard could still be flawed.

 

[apoppin]

Sure, I don't dispute that. What I dispute is your interpretation.

You say: AMD knows the 6990 has issues in some motherbaords, so it's AMD's fault!

I say: AMD knows, because it's a known flaw with some motherboards when running their card. Maybe the nVidia cards that draw similar power don't exhibit the problem because they include a work-around to compensate for these faulty motherboards. The motherboard could still be flawed.

You are not disputing my interpretation as it is the same as yours.
:whiste:

i have stopped faulting the card because it appears that there is some definite incompatibility between it and the MB. i can't blame AMD for not making HD 6990 work in every single MB nor can i ask Gigabyte to RMA my board because it no longer functions with a single videocard.

As i said before, a BIOS update may include a workaround. If so, great. If not, then the HD 6990 goes into its box until i get Ivy Bridge this Winter (and i will use it in Bulldozer to compare against Ph II architecture, so not all is lost)

i'll resurrect this thread when i get yet another MB

i appreciate all of your comments and suggestions. No point in trying *everything* .. if a BIOS update doesn't fix it, i will be patient. Thanks again!

 

[apoppin]

UPDATE, as promised:

No idea what happened or what went wrong. After a few days of moving the HD 6990 back and forth and changing cables, RAM, PSUs in and out .... i asked Gigabyte to get involved and got a RMA number for my X58 MB

Immediately after getting the RMA number, the HD 6990 began working again and it has been fine for the last week - and it has been working on the exact same setup that it failed on previously. i have no explanation.


Anyway, i have finished benching with HD 6990 for my upcoming SLI vs CF showdown at 1920x1200/2560x1600 and 5760x1050 and i am running TriFire-X3 (HD 6990 + HD 6970) without incident (except for my damaged hearing).
:whiste:

_______________

Oh yeah, i finally heard back from AMD .. it was something like Pfft
J/K'ing