what is wrong with amd

killernerd

Limp Gawd
Joined
Mar 8, 2012
Messages
215
I made 2 threads about my cf'ed 7970's fucking my system up already but let me explain the situation once more.

I got them the moment they popped up in the local hardware shop. Put them in my system, installed the drivers and whatnot and haven't been able to play a decent game with cf enabled ever since.

BSOD's to no end, display driver crashes, couldn't even go to youtube without crashing my pc, games would make it hardlock and much much more.

My voltages were fine, temps were fine, disabling ULPS didn't do shit and each card on its own runs perfectly.
And by running speratly i mean physically uninstalling one card and testing with the other. Both cards performed beautifully. Add in the second one for cf and it all turns to shit.

I did put them in a friend's system for about an hour. Not enough for decent testing but at least they didn't crash.

So i descided to contact amd support and after 2 months of mailing back and forth with them they finally convinced me that my mb's at fault.

Ok I think, plausible. I checked all the other possibilities. Right?
- the cards on their own perform normally
- voltages were fine
- no problems with the temperatures
- disabled ulps (didn't help)
- tested in another system (though only basic testing was done)
- flashed my bios
- downloaded the latest and greatest drivers for my motherboard
- did a clean install of windows

At that point i was like, what the heck, my current system is just over 3 years old (first gen core i7) so i started looking towards upgrading my pc. And when i do upgrade, i do it decently; No half assed upgrades.

So this is what i got:

Asus rampage IV extreme
core i7 3820
16 gigs corsair vengeance (4x4gig)
cooler master cosmos II
OCZ vertex 3 120gig
corsair hx1050, 1050watt psu

formatted my windows drive, installed windows on my ssd. Installed all the drivers (from the cd this time) and installed the amd drivers.
Played a bit of dirt 3 (which would occasionally crash too), all went well. Untill i started browsing youtube, after about half an hour BAM display driver crashed.

You have got to be fucking kidding me!?!?!?

It recovered so i think, ok w/e perhaps an unlucky strike. Browsed yt a bit more, about 5 minutes later AGAIN.

FUUUUUUUUUUUUUUUUUUUUUUUUUU

I have to shut down the browser this time, artefacting and shit. Open my browser again, go to youtube and i just KNOW that i'm going to get a BSOD when i click that video... and sure enough my pc hardlocks and BSODS. a message about display driver recovery timeout or something (the usual really).

GODMOTHERFUCKINGDAMNITRRRRRRRRAAAAAAAAAGGGGGGGGGGEEEEEE

Conclusion: i just wasted ALOT of money on a new system and now it all turns out to be for fucking nothing. Don't get me wrong though, i like this new system, it's just that i never planned to do any further upgrades since my old system had more than plenty of horsepower left do handle everything i could throw at it. And i haven't even tried anno 2070 yet (the game that gave me the biggest headache).

Damnit, i'm really at wit's end here... I don't know what to do. Why won't it fucking work?
is it the cards after all?
the drivers?
perhaps even the fucking cf bridges?

desperate nerd out.
 
Last edited:
defective xfire bridge? don't install ANY oc software , afterburner can interact weirdly especially if you have "overclock at system startup enabled, power supply, just some random guesses off the top of my head.

What I would do is a fresh windows install, DO NOT mess with ulps, just do not install afterburner period and see how it works. Doing any install with xfire and afterburner fucks my system up, so I use trixx instead. Afterburner caused crashes on my xfire box.
 
psu is a 1050watt HX1050 from corsair, doubt it's having trouble handling this little amount of hardware.

It's a clean install of windows, no OC on cpu or gpu and no OC software at all for that matter.
 
Holy crap sad story if it was me I would have went to Fry's and if they had GTX 680s I would have bought that and seen if if it gives me problems and if not just sell off the AMD cards if it gave me problems I would have taken them back and probly did what you did from there.
On a side note I have a GTX 295 that in SLI mode is problematic and have RMAed twice with no luck sold it to a budy and got a GTX 480 and no problems and he says he has yet to get any of the issues with the GTX 295 I had. I added a second GTX 480 and no SLI issues. I personally think since SLI and CF are not main stream they don't test them and fix them as they should. Just a guess
 
i should have some spare crossfire bridges somewhere... I'll see if that helps. If not, then i'm going to seriously consider putting them up for sale and get 2 680's in stead...

12.3 was even worse for me.
 
I've seen this with various drivers on single cards going back to the 4850,5870, 6970, and my 7970. When I was googling for a solution, I found it in an Nvidia related thread (googled by the BSOD #), so it's not manufacturer dependant. For each of your browsers, open the front page on [H], right click on an animated ad, you need one that has the "settings" option available, click on that, then disable hardware acceleration. It may need a reboot, and you may need to do it on each update to Flash. Enjoy your comp !
 
Hi,

I had the same problem in my CrossHair IV Formula and three 6970s that just caused constant crashes, bsod, and hardlocks. Was using a 1200 Watt Corsair AX Gold PSU even but found that they were completely stable in an X58 system I had after swapping video cards around. So it could be the cpu/motherboard as it was in my case. Which is why my CH4 is forsale right now on Ebay! No more AMD for me!

Good luck..
 
My single 7950 OC does the same as your cards do. To make mine run at stock speeds without the driver recovering I have to go into Catalyst Control, click Performance, click Graphics Overdrive, scroll down to Power Control settings and slide it all the way to the right which is +20%.

If I don't do this my card recovers the driver every 15 secs or so in a continuous loop especially if I enable anything such as AA. Even after doing that text in games will sometimes just stutter like I'm sitting in a car with a high powered stereo and trying to stare at the rear view mirror. But it only does that if I cut on AA so I just learned to leave it off.

Hope that helps.
 
My friend had these problems a few generations ago with an Nvidia card and his motherboard. The card worked fine in another computer of nearly the same specs with a different motherboard/ram configuration.
 
I've seen this with various drivers on single cards going back to the 4850,5870, 6970, and my 7970. When I was googling for a solution, I found it in an Nvidia related thread (googled by the BSOD #), so it's not manufacturer dependant. For each of your browsers, open the front page on [H], right click on an animated ad, you need one that has the "settings" option available, click on that, then disable hardware acceleration. It may need a reboot, and you may need to do it on each update to Flash. Enjoy your comp !

hmmm. I don't recall it working wonders but w/e i'll do it anyway.

Hi,

I had the same problem in my CrossHair IV Formula and three 6970s that just caused constant crashes, bsod, and hardlocks. Was using a 1200 Watt Corsair AX Gold PSU even but found that they were completely stable in an X58 system I had after swapping video cards around. So it could be the cpu/motherboard as it was in my case. Which is why my CH4 is forsale right now on Ebay! No more AMD for me!

Good luck..

what's the chance of both my old and new system being broken?
 
dude, did you try using just single 7970?

if nothing happens with one then its the gpu or drivers.


cant believe you bought all that stuff before trying a single gpu first.

He can't believe you posted that without reading the first post !
 
That sucks man. One more reason I'm done with multi-card setups, CF specifically. Nothing but a goddam headache.
 
If it always happens on YouTube, then the issue is most likely Flash based

Yes, SteadyVideo enabled is fine. If you have it checked so that it works with internet video it messes with my system. I forgot about that!
 
The only other thing that I can think of, based on your previous diagnostics, is the HDMI audio driver. I've read about there being conflicts with certain HW combinations using Realtek onboard sound.
 
The only other thing that I can think of, based on your previous diagnostics, is the HDMI audio driver. I've read about there being conflicts with certain HW combinations using Realtek onboard sound.

SSD's are known to cause these kind of problems even more.
 
That sucks man. One more reason I'm done with multi-card setups, CF specifically. Nothing but a goddam headache.

Much agreed, although it would be such a shitty turn of events if it indeed was related to the audio after going through all that
 
Forget everything, and just install 12.1, not 12.2, not 12.3 12.1! try it with clean install and report back! they are by far the best recent driver AMD released.

Display driver has stopped working is not present only on 12.1 for me i tried EVERY driver possible.
 
What was the release drivers again? Didn't Kyle/Brent say that every driver but the release driver causes a BSOD in crossfire. This issue is supposed to be fixed in the official 12.5's. I'd test another power supply if possible. The power supply is definitely adequate or even overkill for that setup but could be defective and only showing this when it's stressed a bit more. Also run a cpu only burn in test and verify stability.

Sadly my only gripe with AMD this gen is their drivers and the mini dp aspect of their cards. Getting their hardware much cheaper than nvidia is sometimes worth the trade off, this gen however there isn't much of a price difference to justify tollerating AMD's fuck ups. My best advice OP, is if your done with this crap, sell your cards and get 2 670s or 2 680s. Unless the card's arrive DOA which can happen with any card, they should fix your issue and you will be gaming on a stable rig for months/years to come.
 
nvm, reading new responses

ok this system does indeed use realtek audio drivers for the onboard audio. When i check my audio devices in their respective manager i see that each card has 6 audio devices, for 12 in total.
The main annoyance is that it takes ages for the manager to open, if it opens to begin with. Disabling said hdmi audio devices solves that issue though. (could this be the culprit then?)

steadyVideo is not enabled, it only fucks things up even more.

I used every driver release from the moment the cards came out, including 12.1. It wasn't any better than my current drivers or at least not enough to be noticable.

Why would i test another psu? This one is only a couple of months old and the voltage readings are fine.

Not sure what the bsod code is tbh. I don't really pay attention to that, it's suffice for me to know that it's caused by the gpu.
either:
atikmpag.sys
or
failed to recover after display driver timeout (or something along those lines).

I'll write down the code if it crashes again (which it probably will).
 
Last edited:
In before a bunch of AMD fan boyz get in this thread and post "I have no problems with mine, must be user error..." :rolleyes:

Seriously dude, get rid of AMD CF, nothing but headache. You will spend more time troubleshooting than actually enjoying your PC. Playing driver roulette with AMD got old for me really quick. If you must, stick with only 1 7970 and sell the other because you will sleep alot better at night.
 
What was the release drivers again? Didn't Kyle/Brent say that every driver but the release driver causes a BSOD in crossfire. This issue is supposed to be fixed in the official 12.5's. I'd test another power supply if possible. The power supply is definitely adequate or even overkill for that setup but could be defective and only showing this when it's stressed a bit more. Also run a cpu only burn in test and verify stability.

Sadly my only gripe with AMD this gen is their drivers and the mini dp aspect of their cards. Getting their hardware much cheaper than nvidia is sometimes worth the trade off, this gen however there isn't much of a price difference to justify tollerating AMD's fuck ups. My best advice OP, is if your done with this crap, sell your cards and get 2 670s or 2 680s. Unless the card's arrive DOA which can happen with any card, they should fix your issue and you will be gaming on a stable rig for months/years to come.



As far as I know, Brent complained about crossfire / eyefinity which apparently had some issues. Crossfire single screen works fine for most people, at least for me and a bunch of other people....I honestly think most people are fine and the problems are the exception.

I don't like defending AMD much anymore, clearly they have issues. Nvidia definitely is having a better track record with out of box experience and working on exotic setups with no issues, have to hand it to them. If AMD has any business sense they will increase the size of their software development team, good hardware doesn't mean a damn thing without proper software to enhance it.
 
Last edited:
In before a bunch of AMD fan boyz get in this thread and post "I have no problems with mine, must be user error..." :rolleyes:

Seriously dude, get rid of AMD CF, nothing but headache. You will spend more time troubleshooting than actually enjoying your PC. Playing driver roulette with AMD got old for me really quick. If you must, stick with only 1 7970 and sell the other because you will sleep alot better at night.

Why be obnoxious with a response like this?
 
Why would i test another psu? This one is only a couple of months old and the voltage readings are fine.

Well, is your voltage readings normal on your system (not just the cards) when your rig crashes? Have you kept logs of this?

The real question is why wouldn't you test another PSU.

I can give you one and many reasons why you should

1. To keep your sanity

Unless you want to keep posting shit like this

FUUUUUUUUUUUUUUUUUUUUUUUUUU, GODMOTHERFUCKINGDAMNITRRRRRRRRAAAAAAAAAGGGGGGGGGGEEEEEE
 
my HD4890's in cf worked fine for 3 years on end, then some bad stuff happened and they melted, which was the only reason for me to upgrade to 2 7970's.
 
Try entirely disabling onboard audio. Also, post what the BSOD code is. Also,try running memtest86 to verify your RAM is good. I know its unlikely but I spent a month tracking down random BSOD and hardlock problems before and it was a problem with one stick of memory.

I would also try another crossfire bridge.
 
And yes try another PSU. Yours should be fine with that system but it could be defective.
 
Well, is your voltage readings normal on your system (not just the cards) when your rig crashes? Have you kept logs of this?

The real question is why wouldn't you test another PSU.

I can give you one and many reasons why you should

1. To keep your sanity

Unless you want to keep posting shit like this

hmmm, your point is valid good sir. it's just that windows 7 and voltage monitoring programs do not like each other.

speedfan says that my +12v is at 6.71
my bios says it's good (can't remember the exact value)
 
hmmm, your point is valid good sir. it's just that windows 7 and voltage monitoring programs do not like each other.

speedfan says that my +12v is at 6.71
my bios says it's good (can't remember the exact value)

Understood, just keep in mind a bad psu could have voltage droop during stressful sittuations that can cause issues. That PSU is a bad ass unit and can hadle tri fire let alone your setup. However it could be defective. See if a buddy of yours can let you test their psu just for the hell of it. You have to know someone that has at least a decent enough psu to throw in the mix, just so you can check that off the list of troubleshooting steps. Look at the PSU in my sig. If I was having similar issues, I wouldn't think twice about testing another unit. Bad PSU's can happen to the best manufacturer and are very hard to diagnose at times.

Kudos to Rizen mentioning the ram. I second his recomendation there. A bad stick of ram could definitely contribute/cause to this kind of issue
 
what power supply?

This doesn't sound like a power supply problem.

defective xfire bridge? don't install ANY oc software , afterburner can interact weirdly especially if you have "overclock at system startup enabled, power supply, just some random guesses off the top of my head.

What I would do is a fresh windows install, DO NOT mess with ulps, just do not install afterburner period and see how it works. Doing any install with xfire and afterburner fucks my system up, so I use trixx instead. Afterburner caused crashes on my xfire box.

Unlikely, Crossfire bridges either work or they don't. They won't allow you to enable Crossfire and act weird. Power supply isn't the issue. This isn't how a power supply problem manifests itself. I do agree that all that extra software needs to be uninstalled or thrown out until the system is working.

Deff sounds like PSU is running out of Juice.

How so? Power supply problems DO NOT MANIFEST THIS WAY. I've said this in 1,000 threads. When a power supply has insufficient power, the system will usually randomly reset, power off, or fail to POST at all. They do not cause instability in the traditional sense. You do not get BSOD's from their failure. And it's very likely that an insufficient PSU could cause the graphics card to alarm as it would when you don't connect all the power cables and start up the system.

psu is a 1050watt HX1050 from corsair, doubt it's having trouble handling this little amount of hardware.

It's a clean install of windows, no OC on cpu or gpu and no OC software at all for that matter.

That's good.

What slots on the motherboard are you using for your cards? I'm assuming you are using the primary PCIe x16 slot for the first one, but what of the second?
 
i went along with the lay-out suggested by the manual, so it *should* be using the 2 x16 slots.
They are slot 1 and slot 4 or slot 1 and 3 if you don't count the black one.

130989.jpg



also, i do have a spare 850w corsair psu collecting dust somewhere. But as i and Dan_D have said, it's unlikely that the psu is at fault here.
 
Pretty sure you have to use the top 2 ....

Thats the way every motherboard i've owned is, the slots closest to the CPU have the most pci-e traffic....
 
Back
Top