CCC errors, hangups, FPS drops

I've hit a sh1t wall again, this time the ATI drivers are going crazy.

ISSUE

For no apparent reason, the games will hard lock the computer after some 30 minutes of playing. Sometimes disappearing textures/glitches occur prior to this.

It starts with extreme frame rate drops to the point where image pauses but sound goes on, the 0.2 - 0.5 FPS "state" will last for 10-20 seconds and it goes back to normal.
Occasionally it will either crash the game or lock the entire machine.
Also Sapphire Trixx has crashed because of those errors, I stopped using it since (it's not a cause, it's also a victim).

ERRORS

In event viewer under ACEEventLog you can find reports (Informations, not Errors) such as:

Code:

0000000004: 2011-03-22 10:32:42:342    FAILED:ADL_Display_SLSGrid_Caps and return value is: -1
 Error      Called by: ATI.ACE.CLI.Caste.Graphics.Runtime.RT_GraphicsAdapter_N::CheckSLSSupported                       processID:02152 threadID:(CCCThreadNew:Runtime) domainName:(CCC.exe                    ) assemblyName:(CLI.Caste.Graphics.Runtime, Version=3.5.4074.42519, Culture=neutral, PublicKeyToken=90ba9c70f846762e)

0000000008: 2011-03-22 10:32:43:949    _IDEMDeviceDFP2Settings_0812.GetDFP2ITCFlag failed with status 2
 Error      Called by: ATI.ACE.CLI.Aspect.DeviceDFP.Graphics.Runtime.RT_DeviceDFP::PrivateRefresh                       processID:02152 threadID:(CCCThreadNew:Runtime) domainName:(CCC.exe                    ) assemblyName:(CLI.Aspect.DeviceDFP.Graphics.Runtime, Version=3.5.4074.42540, Culture=neutral, PublicKeyToken=90ba9c70f846762e)

0000000098: 2011-03-22 15:51:59:122    Removing errant RevertAfterComponentUpdated observers.
 Error      Called by: ATI.ACE.CLI.Component.Dashboard.Dashboard::DerivedRun                                            processID:00712 threadID:(CCCThreadNew:Dashboard) domainName:(CCC.exe                    ) assemblyName:(CLI.Component.Dashboard, Version=3.5.0.0, Culture=neutral, PublicKeyToken=90ba9c70f846762e)

0000000099: 2011-03-22 15:51:59:122    Target: ATI.ACE.CLI.Aspect.DeviceDFP.Graphics.Dashboard.DB_DeviceDFP Method: ATI.ACE.CLI.Aspect.DeviceDFP.Graphics.Dashboard.DB_DeviceDFP
 Error      Called by: ATI.ACE.CLI.Component.Dashboard.Dashboard::DerivedRun                                            processID:00712 threadID:(CCCThreadNew:Dashboard) domainName:(CCC.exe                    ) assemblyName:(CLI.Component.Dashboard, Version=3.5.0.0, Culture=neutral, PublicKeyToken=90ba9c70f846762e)

These are some of them, not all.
The first error is the one that always appears when the application crashes or the computer hangs.
There seems to be a "pipeline" of errors that leads to a crash/lockup, i. e. there is a repeating pattern of errors occurring before the crash/lockup.
Those errors will literally appear in hundreds, I have more than 150 errors in 5hrs uptime of which zero is games (or 3D in general). This leads me to conclusion that it's not only related to 3D (games), but also to UVD and all hardware accelerated content.

FIXES

I've google some of them and seen many people having those issues, it seems unrelated to driver version (happened back on 10.x and older).
Reinstalling drivers doesn't help, some people even formatted and fresh installed system and the issue persisted.

Also changing cards didn't help, which rules out hardware issue with one GPU.

I have found reports suggesting that disabling Overdrive helps, which, sadly, didn't work for me.

What worked (at least in BFBC2) was changing the FOV from original 55 to 85.
I have no idea how this can be related, but I had no crash in this game for hours last night, while I had a couple of crashes and multiple freezes/FPS drops every time I played.

I have discovered that disabling fan profilers such as Sapphire Trixx and SpeedFan postpones the crash/hangup, but it will eventually happen anyway.

I have, of course, tried every possible fix that I thought of, ranging from driver cleaning and updating, removing anything that isn't necessary from the system to cleaning VCRedist.

I have, of course, made sure every possible hardware related issue isn't causing this, such as heat problems or PSU-related problems.
Let me stress again for those willing to draw conclusions from the fan profilers mention above, this is not heat related.
I have been running Furmark and OCCT for hours with no problems.

I have, of course, tried disabling and tweaking settings in CCC, such as disabling A.I. and adaptive AA. None of those settings seem to have any effect, the problem (as errors suggest) lies deeper.

It has not been present from the beginning, nor is related to new GPU.
It just started happening one day (about a week and a half ago) for no apparent reason, with nothing new being installed. One exception could be an automatic system update.
I've seen reports of this with 69xx, 6xxx in general, 5xxx and 3xxx and even older cards.

Someone suggested that installing the video driver without Catalyst Control Center might help, which obviously makes sense, since those are CCC-related errors. But this isn't really a fix, is it.

If you know this issue and have fixed it, please reply.

If you have this problem too, please reply.

If you don't have a problem, but just checked and you also have those errors in event viewer, please reply.

If you think that nVidia has superior drivers, STFU&GTFO because you're ignorant.

I know it has appeared here on this forum, but there is still no solution.

 

I do have those "errors"

Code:

FAILED:ADL_Display_SLSGrid_Caps and return value is: -1
 Error      Called by: ATI.ACE.CLI.Caste.Graphics.Runtime.RT_GraphicsAdapter_N::CheckSLSSupported                       processID:03352 threadID:(                ) domainName:(CCC.exe                    ) assemblyName:(CLI.Caste.Graphics.Runtime, Version=3.5.4074.42519, Culture=neutral, PublicKeyToken=90ba9c70f846762e)

But, everything is running perfectly.
What you can try is disable this service:

AMD External Events Utility

Run CMD.
Net stop AMD External Events Utility

And try games. See if it still locks up.

I'd wish i can help you further, but, I'm totally lost to be honest.

 

did u try running ur cpu @ stock ?

 

MerolaC I will try that, but I have a strange feeling that will suppress error messages only, not actually fix the cause. But I will try it.
It is visible and accessible in Service Control Manager, and since it's a service, you'd rather use

Code:

sc stop "AMD External Events Utility"
sc config "AMD External Events Utility" start= disabled

than

Code:

net stop "AMD External Events Utility"

lenne yes, I have it running on stock clock, I even tried locking it back to 2 cores, it doesn't matter.

 

Have you tried reverting everything back to default and just cranking the fan up? There is more to cool on the video card then just GPU. You have:
-IC vram
-vregs
-and other components

Also make sure that the screws are fasten correctly (not loose).

 

Thanks a lot for those commands. Today, EXACTLY, today, i was at the place I work trying to figure out with a friend how to set a service to disable through CMD.
Awesome.

I hope you can solve your issue.
And thank you. :nerd:

 

Hmm, I think I may know the problem here. Your Zalman ZM-600HP 600W has 4 12v rails each rated for a max of 16A. From my understanding the 6850 needs at least 25A for the 12 volt rail. If there is nothing else wrong and the video card is in fact not damage (and software is fine) getting a PSU that has at least 25A on the 12V rail should solve the problem. I'm not sure if you prefer Zalman PSUs or not but if you look at their website they do have a moddable PSU that offers 25A per rail: http://www.zalman.co.kr/ENG/product/Product_Read.asp?idx=404.
But I would suggest either PC Power and Cooling, Corsair, etc.

 

6850 is only about ~120W so needs about 10A on the rail.

Watts / Volts = Amps
120 / 12 = 10

 

I played BFBC2 now for like 15 minutes and already had 2 freezes - the sound is there, the FPS is 0, and after it resumes my character is ofc moving by itself - pending query. So the FOV wasn't a fix, which was to be expected.

I have disabled the AMD External Events service mentioned previously.

Code:

0000000040: 2011-03-23 02:53:23:102    FAILED:ADL_Display_SLSGrid_Caps and return value is: -1
 Error      Called by: ATI.ACE.CLI.Caste.Graphics.Runtime.RT_GraphicsAdapter_N::CheckSLSSupported                       processID:01112 threadID:(                ) domainName:(CCC.exe                    ) assemblyName:(CLI.Caste.Graphics.Runtime, Version=3.5.4074.42519, Culture=neutral, PublicKeyToken=90ba9c70f846762e)

0000000051: 2011-03-23 02:53:23:760    FAILED:ADL_Display_SLSGrid_Caps and return value is: -1
 Error      Called by: ATI.ACE.CLI.Caste.Graphics.Runtime.RT_GraphicsAdapter_N::CheckSLSSupported                       processID:01112 threadID:(                ) domainName:(CCC.exe                    ) assemblyName:(CLI.Caste.Graphics.Runtime, Version=3.5.4074.42519, Culture=neutral, PublicKeyToken=90ba9c70f846762e)

Since it started happening, I manually crank up the fans before I play the game.
The GPU fan is variated by MSI Afterburber now that I ditched Trixx, but apparently this doesn't change anything.
I play with VSync so on max settings the GPU load is approx. 80% and it's not getting very hot, staying in the 50ish range.
The GPU and voltage regulation are actively cooled, the memory is not. However it looks nothing like memory overheat, and since the clock are stock and everything is well-ventilated, I rule this out.

As I have previously stated, I ruled out the PSU as a cause as I have recently put new high-end low-ESR electrolytic capacitors in it and I have measured the voltages under stress when the problems occurred. If the PSU was struggling to deliver under load, the sign of it would be a considerable voltage drop, which isn't the case on any of the lines.

The PSU does have four rails, but those are virtual rails, and it can provide more than 30A for the GPU alone, so this is really not the problem. Yes, the sticker says 16A, but the sticker is a communist and they lie. Link

Also the GPU isn't overclocked, as a matter of fact nothing in my system is.

And remember that the GPU is getting up to 75 Watts from the PCI-E alone.

The GPU is also fairly new, I have had it for less than a month. I have taken the whole machine apart and reassembled, reseating the GPU repeatedly.
I wish I could have Riva Tuner working, I could see CPU load, RAM usage and all that stuff at the very moment the problems occur.

I ran out of ideas as to what could be the cause.

 

This is incorrect on a number of reasons. But to get to the gist of it. There is more on the 12 volt rail then just the GPU.

If you cannot find no other reason and you believe that the card is working as it should (drivers and all) the only other step that I can think of is to try a better PSU. One that offers more amperage on the 12volt rail. All you are saying is that the current PSU you have is giving you the power it should at spec. That may not be adequate for you PC. Also, your linked review only suggested over 30amps on just one rail while it dropped to 18amps on the other. Only 2 of the 4 were tested.

 

Dude, get your facts straight.
As I said, this PSU has 4 VIRTUAL rails, that means that they pulled 48A from the PSU 12V and that's it. There are no physical rails in the PSU but one.

And that doesn't even matter. 16A is way more than my card would ever require, and if it was power deficiency, it would be felt all the time under stress - it's not. And what inklimited said is correct for a number of reasons, because if you only count the rail the GPU is on, there is nothing else to be counted towards the 16A limit.

And I have already said that I have tested the PSU under heavy load with multimeter, the voltage doesn't drop even 0.1V, maybe two to three hundredths, but that's within measurement error.
The PSU is able to deliver 700W which is way more than I would ever need even with two 6850 in CF.

 

how come nobody's blaming the card?

does it happen on very weak games? like minecraft, hl1, quake3, etc
dx8, 9, 10, 11, ogl?

 

Because it's not the card's fault.

 

Okay, reevaluation:

I have made a clean install of Win 7 x64 Pro on a different HDD.
Installed only the basic drivers, that is: for the chipset, NIC, sound and GPU (11.2). DirectX, too.

Disabled AMD External Events service. Did not unlock or enable AOD in CCC.

Installed critical system updates only. No SP1 this time.

Same CCC-related errors in Event Viewer occur. I had BSOD F4 on desktop and BFBC2 freezed with looping sound after some 15 minutes of playing.

This rules out any possible conflict with any application. I still think that it might be some system update messing things up. I haven't even installed Flash.

Why not hardware? Because it started happening about 2 weeks ago for a number of people, if it was a hardware issue it would have surfaced earlier and for a reason, not just out of nowhere.

I still have no idea what could cause this. I know for a fact that I have wasted a lot of time setting up new OS for nothing.

 

well 2 weeks ago (march8) was patch tuesday so go through/uninstall/dont install those patches & test?

i dont recall seeing anyone with something like this this month, got some links?

 

Interesting that your hypothesis didn't work. But like I mentioned earlier that review doesn't justify what you think and I still think you should look into your PSU if you think that card is fine. Or just RMA the card But I do wonder if gpuz reads your slot as x16 2.0.

 

I have uninstalled all windows updates except for two, which can't be uninstalled. 64 updates less, the computer ran FurMark extreme burn-in + OCCT Linpack for 4 solid hours and nothing bad happened, except 6700 CCC-related errors in Event Viewer.

I have plugged in a monitoring device into the PC's socket especially for you.
Idle, it draws 100W, under extreme load it goes up to 300W.
The GPU alone draws 140W. The CPU in Linpack draws 80-90W (3 cores).
That is measured on primary side of the PSU, of course. Taking PSU's efficiency of about 80% into account, that means the PC is actually drawing 240 Watts, the GPU alone about 115W and the CPU about 72W.
The combined output of 12V rails is 504W. Maximum combined current on 12V rails is 42A (it obviously isn't 16A per rail multiplied by 4 rails). Overall the PC isn't even touching 50% of maximum power the PSU can deliver. Give the PSU a break, even a 400W PSU wouldn't sweat much, 600W PSU is bored and actually not sufficiently loaded, which decreases it's efficiency.

There are 2-3 threads on this forum in Gaming section about problems with BFBC2.
There is a big discussion on EA BFBC2 UK forum and it has been going for some time.
Even in new AB beta thread someone mentioned the BFBC2 problem.
As for other games, I can't run L4D2 today. It crashes to desktop after about 2-5 minutes, even when in the main menu.
I was getting BSODs F4 yesterday, but I can't reproduce it today to analyze the memory dump. Perhaps getting rid of updates solved that part of this mess.

 

i just stick to the ati section, so i havent seen bc2 stuff outside black & texture flicker (which happens on sli too)

what i'm wanting to solve is your crashing after a few min in multiple games

 

I have fixed L4D2 and it's working rather fine, although there were some strange FPS drops appearing for people who played on my servers. I'm not sure if this is related in any way, so let's say that this game is fixed.

GFWL in Batman: AA refuse to recognize my Steam account so I basically lost all my progress and I'm not very eager to replay it, so I can't elaborate on this game's stability. I have however run the in-game benchmark and it was ok. No BSOD 0xF4 today anyway.

So basically it comes down to just one game now, Battlefield Bad Company 2. Since I removed all the windows updates, this possible cause probably have to be crossed out. I have just played BFBC2 after applying a supposed fix to PB's service found somewhere on the net, and the game, as usual, freezed with looping sound after about 10-15 minutes of playing.
The problems with this title have not, and most probably will never be, addressed by DICE, so apparently for those unlucky to suffer from this malfunction, this is the end of BFBC2 MP.

I'll try some tweaks and things that I thought of now, but I'm nor very optimistic.

ED: I set the fans manually, no monitoring/fan controlling app in the background, then shut down CCC and MOM with task manager. I played for more than an hour and no crash or freeze or slowdown whatsoever. Interesting. Apparently, CCC might indeed be the cause. More testing tomorrow.

 

Time has passed, fresh Windows 7 x64 installation has been made (twice), everything was fine until today.
Out of nowhere, for no apparent reason - crashes, BSODs and whatnot.

I have new PSU, I have been doing OCing for two days and everything was fine.
Today, don't know why, it's a ****storm.

No windows updates, no driver change, no new app. BSODs when booting, hangs when browsing net, WTF.