Graphics artifacts, black screens, and device hung

bravodown

Distinguished
Dec 5, 2012
73
1
18,545
Ive been having problems with my graphics for a few weeks now. Ive been working a lot and went on vacation during this period also, but I've spent many hours trying to fix the problem, which includes:
-artifacts in different forms (short green horizontal bars stacked in columns across the screen, large colored pixels, weird screen distortions and other colors of bars, etc, hard to describe all of them).
-screen resetting and going black (as though my computer was shut off but monitor still powered on; the blue on light turns to the no-picture yellow light).
-computer freezing during black screen or artifacting.
-not being able to load past windows loading logo without screen going black at times.

My system specs:
CPU: Intel i7 quad core 3.5ghz (3.9 turbo)
GPU: GeForce 670 FTW 4gb (now GeForce 660 ti that is recognized as a 670 FTW from EVGA RMA replacement)
Mobo: Asus z77 sabertooth
memory: Corsair Vengeance 16 gb DDR3
storage: x2 256gb OCZ Vertex 4 SSD
psu: Corsair TX650W 80 Plus Bronze
monitor: LG Flatron L196WQ-BF

The problem seemed to start when I was playing Battlefield 4. In the past my screen might have flickered off briefly but it always came back after about 1 sec and then i played on without issue. I just thought this was some random glitch and it happened rarely enough that it didnt bother me much. I didnt have any problems with any other games, except sometimes Planetside 2 would stutter briefly but also rare and didnt give me much pause for concern. The computer has otherwise ran flawlessly since ive built it back in late 2012. Then about 3 weeks ago or so, I started getting the errors after the game flickered off in BF4, resulting in the game just closing or freezing up and also receiving the following messages:

- (1) - Display driver stopped responding and has recovered Display driver NVIDIA Windows Kernal Mode Driver, Version 340.52 stopped responding ans has successfully recovered

or

- (2) - DirectX function "GetDeviceRemovedReason" failed with DXGI_ERROR_DEVICE_HUNG ("The device is hung which is typically caused by issues in the graphics driver or alternatively the application"). GPU: "NVIDIA GeForce GTX670, Driver: 33788

At first the problem only seemed to be with battlefield 4. But then it started happening during Hearthstone, Firefall, and other games. The screen would even flicker while just sitting idle at desktop giving me the first error message posted above. At times the computer would freeze up requiring me to hardboot system. The screen would go black as I got to the windows logo load screen. Id have to reboot again, enter safemode, reinstall graphics drivers, then I would be able to load into windows.

With the 670 i tried everything EVGA asked me to do, including going back to older drivers, switching sli slots of graphics cards on mobo, uninstalling/reinstalling directx, updating bios, and other things that ive read about online that i cant recall at the moment. Nothing has seemed to work. After trying everything to their satisfaction, they told me to go ahead with the RMA process. As noted in specs above, I received my new card (660 ti) but it is recognized as 670 in device manager and also gpu-z recognizes the card as a 670 with 4gb. Ive tried current and older graphics drivers with both cards to no avail.

Everything ran fine at first. Windows loaded without the screen going black. I was able to play games like ghost recon phantoms and firefall without issue. Everything seemed fine, so I decided to play some planetside 2 because I hadnt played in awhile (I didnt play this while my other card was acting out so I cant compare results with replacement). After about a minute of playing the screen will start to distort with artifacts (different from the type of artifacting i saw generally from the original card, different in color, shape, etc). Then Planetside 2 will usually freeze up requiring a hard reboot unless i close the game shortly after seeing the aftifacts and stuttering. I thought this might be an issue with just planetside. So i tried playing Battlefield 4 for the first time with the RMA replacement card. Game ran fine for a couple minutes before I received error (1) listed above again. I should note that I have only experienced errors inside bf4 or planetside 2 with the replacement card and have not had trouble with windows loading or screen flickering while idle at desktop, whereas the old card had problems sitting idle on desktop or even loading windows as mentioned previously.

Ive had ppl mention PSU as possible problem or the motherboard itself. The motherboard comes with a pretty serious warranty (5 years) so I'd liek to think it wasnt the mobo but you never know. The PSU i have is sufficient for my setup, or so ive been told by multiple sources including EVGA tech support.

I've been meanign to reformat and start from clean slate anyway, so after the RMA card not performing I decided to do that hoping it might help. After all the updating and reinstalling after reformat, I tried planetside 2 and the same artifacting happened, and computer freezing. I havent tried bf4 yet since reformat but I'm guessing i'll get the smee errors (im downloading it now).

Does anyone have any ideas of what might possibly be wrong? I'm going to call EVGA back and discuss this with them (because there is always a chance they sent me a faulty replacement) but I wanted to ask the community for any suggestions or ideas, I am at a complete loss right now as to what I can do.

Thanks.

EDIT: 8-26-14 I talked with corsair about my TX650 psu. The agent had me look at voltage levels in bios. From what i read back to him, he said one of the levels seemed lower than it should be. He said there is a chance it could be affecting my ability to play higher intensive games but he had doubts at first. After suggesting a few possible solutions (installing older video drivers, reseating card, etc etc), I told him I tried everything. He then said i could go ahead with the rma if i think it might be my psu. Ive also noticed that the fan in psu wasnt running. I tried running some of the problem games (like planetside2) and played until the system started to glitch or crash, looked at fan and it still wasnt running. I'll know in about 5 days when my replacement arrives if the psu is causing these problems.
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545
The PSU came and didnt change anything. It seemed to work well at first, but after running planetside 2 for about 2-3 minutes it did the same artifacts and system crash. I've also tried changing registry for TDR, which didn't help either. Now the computer will boot up to the windows load screen, show massive artifacts across entire screen, a bunch of small horizontal green lines all over screen, and then pc will either load into windows but with only default screen (super low resolution, device manager shows yelow triangle over the geforce 670) or the screen goes black and doesnt load windows, restarts and goes to safe mode options. Load into safe mode, uninstall all geforce drivers, restart. Windows loads fine in normal mode and installs standard vga drivers. I install any version, old or new, of geforce graphics drivers, restart on prompt after installation, and get the same artifacting at windows load screen.

3-4 weeks later, many 'solutions' attempted including two RMA (video card and psu) and still having issues. The fact that i dont have any problems until i install the geforce drivers tells me its a software issue. What else could it be? Is there some sort of conflict with windows updates and geforce drivers? I disabled automatcc windows updating and changed setting so that windows doesnt automatically look for and install device drivers. This didn't help ether.

Any suggestions would be greatly appreciated. Ive read countless forums with people facing similar problem but none of the so-called 'fixes' have worked for me. Maybe they sent me a faulty replacement gpu? Calling evga again, will update post again if/when i find a solution.
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


I did RMA the card and am having the same issues. My original card is a GeForce 670 FTW, the RMA replacement they shipped me says 660 ti on the front and 670 on the back.... The EVGA rep said this was odd. He told me to go ahead with another RMA but then he talked with a supervisor about it, and the supervisor blamed my cpu or mobo. All my mobo and cpu temps are fine and other than this I havent had any other issues. I've reseated the card on my motherboard and still have same problems. Would the cpu or mobo really cause this issue? I have read in some forums that people had to RMA 2 or 3 times before getting a card that worked. He did say they wouldnt leave me hanging and would send me another RMA if i desired but he wanted me to check thes things first as to not waste more time waiting on another RMA.

Either way, both the old card and RMA replacement are giving me similar problems. They are still under warranty.
 

iAboodz

Honorable
Mar 20, 2013
582
0
11,360
Artifacts and driver failing like you are having is maybe 80% GPU related. But CPU can cause that too if it was overclocked and is not a god chip to handle over clock. It could also be your motherboard but I highly doubt that. It's said that updating your bios may help (I'd start with that). And I'd ask for another RMA too because that 660 ti/670 ftw seems worrying.
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


Yeah I also think it is strange that it says both 660 ti and 670 on the same card, as did the tech rep i spoke with from EVGA. I want to send it back for a different rma but they want me to send the first card back.

When I first called EVGA before doing RMA, he wanted me to update bios drivers first. This didn't change anything.

Things I've tried (not necessarily in this order):
- regedit for TdrDelay (didnt help)
-updating bios (didnt help)
-uninstalling/reinstalling new and old nvidia drivers (didnt help)
-reformatting (didnt help)
-reseating card in all the different pci-e slots (didn't help)
-RMA video card (didnt help)
-RMA PSU (didnt help)

I really do not think it is my mobo or cpu, highly doubt this, considering many other people with different setups are having similar problems (according to multiple forums) and there doesn't seem to be a definite solution. Some of the things I've tried have worked for other people, and it seems other people are in my position, stuck with no readily available or acknowledged solution.

When i have time later in the week, i want to bring my pc to a buddy's place to try some of his amd cards in my computer and my card in his. I'm also going to run some cpu tests and check my memory just to be sure.

How can such a debilitating issue pop up like this out of nowhere and seem to affect so many people?

I also am going to try to lower clock speed and memory speed on the card with msi afterburner but right now im having trouble even loading windows normally once the graphics drivers have been installed. I spoke about this with the evga tech rep and he said this is not a solution, which i agree, but rather a work around. I have msi afterburner but couldnt see how to lower speeds. Anyone know how?
 

iAboodz

Honorable
Mar 20, 2013
582
0
11,360
Just slide the core clock, voltage or memory clock slider to the left until you see the minus"number" and reduce as you please. I'd say -50 MHz on gpu core and -100 to 200 on mem and leave voltage as it is. Try lowering the voltage a bit it may help too.
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545
Ok thx. Ill give it a try when I get home tmrw morning. When i looked at it before I saw the bars for various things with a number value to the right of the bar. But there was no way to adjust the bar (kinda like when a windows options is greyed out and you cant click on it). Do I have to create a profile or something befor ei can start adjusting clock values?

 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


Ok thx. Ill give it a try when I get home tmrw morning. When i looked at it before I saw the bars for various things with a number value to the right of the bar. But there was no way to adjust the bar (kinda like when a windows options is greyed out and you cant click on it). Do I have to create a profile or something befor ei can start adjusting clock values?
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


Ok. Today I decided to reformat again just in case there was a funky windows install. Everything is fine until i install any nvidia drivers. Running off of the standard vga is fine (altho i cant run any games). I also installed precision x. i installed nvidia drivers and started having problems again (artifacts at windows load screen- 'matrix' style screen, sometimes windows wont even load - the screen just goes black and i have to hard boot machine, and if windows does load it looks terrible). Windows prompted me for a system restore before loading because I had to hard boot system. So i did that and restarted. Was able to get into windows but it seemed really slow (cursor was constantly showing a state of loading but nothing seemed to be loading and simply right clicking on desktop would take about 1 min for the options to load up). I restarted again and things seemed a bit better.
I tried using precision x to reduce core and mem clock speeds by 100. I was able to play planetside 2 for a bit to test it, altho i was only getting about 20-30 fps (i normally get 50 fps in high activity areas and in low activity areas of the game i normally get 90-120 fps). So i closed the game, opened precision x, returned to default settings and tried playing planetside again. Same thing, just really crappy low fps and game was laggy. If you turn on the fps monitor in the game, itll show next to the fps value either cpu or gpu, indicating which is being more stressed. Normally, it goes back and forth between cpu and gpu. But today with the constantly very low fps for my system, it only showed gpu.

The issue seems to progressively get worse no matter what I do. Perhaps there is something wrong with my motherboard afterall. I'm going to be giving ASUS a call when i get a chance. I'm also going to run Prime 95 to test cpu based on a suggestion, but i may have to completely remove nvidia drivers to get a decent windows load. The fact that the problem seems to be getting worse despite any software changes (including reformatting), and the fact that i've already done an RMA on both video card and psu, would this point to a motherboard problem? My mobo (z77 sabertooth), has red lights that are supposed to turn on to indicate problems on the board, but no reds lights turn on except when the computer first boots, all the lights blink on then turn off right away. Luckily, all my parts are still under warranty for another 3-4 months for the 2 year warranty parts, and my mother board has a 5 year warranty so if there is a problem it wont be a financial problem.
 

iAboodz

Honorable
Mar 20, 2013
582
0
11,360
I have an Asus board my self and lights do turn on first too and then off. That does not indicate a problem. The reason why you get those problems after you install drivers is that the windows recognise the GPU as just a graphic chip to run the monitor and it would probably not have any gpu power at all, just what it's needed to run windows. The problems occur when you install drivers, which is exactly what happened to my GTX 580 which is retired now (artefacts, driver crashes, bsods and freezes). Exactly your case. I'd really just ask for another GPU RMA. You can do the CPU testing with P95 it's a good idea too. but i don't think it's related.

Good luck
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


I may have made some headway. After looking at the problem from a motherboard perspective, i started reading different forum postings more related to asus mobos. Someone mentioned changing their bios profile from manual to XMS. I just tried doing this and lo-and-behold i was able to actually load windows without artifacts after installing nvidia drivers. I did, however, see some artifact glitches briefly while just sitting at the windows desktop screen, but i updated nvidia drivers, restarted and havent see them again... yet.

I noticed before i changed to XMS that my memory frequency was running around 1371 mhz or so. After changing to XMS it is now at 1600 mhz (which is the natural frequency for my sticks). I'm guessing this had something to do with the problems I was having.

I don't think i'm out of the woods yet, i didnt have much time at home to play around with it so I will post again with another update after I have more free time but at least some progress, even if only a little, has been made.

Why would my memory frequencies have been lowered like that? I have never messed around with them before. Could an update automatically make a change like that? Also, could an update of sorts cause physical dmg to my hardware? Now that i know the memory was running at a low frequency, I'll have to look at other hardware like graphics card to see if its running as it should. Also, i have a i7 3.5 (3.9 turbo). In bios under XMS screen, it said my cpu frequency is 4100...is that too high for the turbo, should it be 3900? If the settings were messed up like this should I reset CMOS or anything like that?
 

iAboodz

Honorable
Mar 20, 2013
582
0
11,360


I have no idea :/ That is weird.
You made me reconsider my fallen 580 now :/.

Do more testing try to play a demanding game and see what happens. If your rams underclocking was the case, which I do recall having too cause all asus boards do that i guess, I'm trying to run my 580 again. It may work why not.
Please do those heavy testing and get back to me if possible
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


Ok, so i finally had some free time at home. I can still load windows flawlessly, no issues there. I think setting the bios profile to XMP (which corrected my memory frequency) fixed that issue.
I tried running Planetside 2, which has been one of the high graphics intensive games i was having problems with. I'm still getting graphical artifacts on screen and the game stopped responding. However, the game did NOT shut down on its own and give me the error i had been getting before (the one about drivers stopped responding and have recovered, windows kernel mode etc..).
I noticed in cpu-z and in my bios that my cpu is running at 4100mhz. Is this normal for an i7 3770k 3.5ghz (3.9ghz turbo). 4100 mhz seems higher than the turbo should allow. Maybe this is causing problems? I have no idea. I never really took note of my cpu speeds before this so i cant say if its always been this way or not.
Also, that funky replacement RMA graphics card they gave me makes me uneasy (saying 660 ti on the front and 670 on the back, even the tech rep said this was very odd). I think i should do another rma, everyone has been telling me to do so because of the funky card labeling.
So, problem isnt resolved yet but i thought id give a quick update to where im currently at. At least i can load windows normally now. And i havent had the drivers not responding error yet either since switching to xmp. At least i'm seeing some progress, even if only a little. Will update again when more happens.

EDIT: Okay i spoke too soon. After writing this, I received the kernel mode error again with artifacts just sitting at windows desktop. I also got the error again when trying planetside 2 again....
 

iAboodz

Honorable
Mar 20, 2013
582
0
11,360


4.1 is okay with the 3770k I think the XMP mode just overclocked the cpu which is fine nothing to worry about. I would do that RMA though as I'm sure that 660 ti or 670 is not okay at all
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


Yeah im definitely calling evga agian for another RMA. It will just take a few weeks to give them first card back and to get the 2nd rma card.
I did want to try prime95 to check my cpu for stability before calling them back. Switching to XMP resolved the issue they told me was mobo side (the artifacts or black screens at windows load screen), so thats been resolved. But I also told them I would try prime95 to test cpu before requesting the next gpu rma.
I have never used Prime95 before and as a result I dont really know what to look for. Do i just let the stress test run for a few hours and if it doesnt show any failures then the cpu is good? Is there anything else i should know about prime95 before using it?
 

iAboodz

Honorable
Mar 20, 2013
582
0
11,360



Nope not really, just let it run for a couple of hours and watch out for the temps. If no fails then your processor is allright.
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


I ran Prime 95 for 4 hours 26 minutes. All 8 workers did 56 tests with 0 Errors and 0 Warnings. The only thing that stood out from the everything else was that on worker 5, once about 1 hour into the test it did a test 1 then a test 2. this was the only time I saw test 2. I don't know what this means but if they all finished with 0 errors and 0 warnings then im guessing my cpu and memory are solid. Temps were good. CPU ran 43-44 C the entire test except when it first started it hit 49 C but went back down right away. Memory temps ran 47-48 C the entire test.



 

iAboodz

Honorable
Mar 20, 2013
582
0
11,360


Great !
Then it is all up to that RMA now :)
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


I'm hesitant to say the issue has been resolved but it very well may be. Still waiting on that other RMA but ive been trying a couple other things in the mean time. I tried resetting bios to defaults, didnt work. And i tried disabling nvidia streaming service which has caused system crashes for some people, but this didnt help either. I noticed a new nvidia driver was released a few days ago and i thought it would be worth a try installing it. During the install, after it uninstalled old drivers the screen just shut off like it had no signal. So i rebooted, and attempted new driver install again. Ever since I havent had the problem. Not a single glitch, kernel mode error, nothing. Before I couldnt run planetside 2 (or any game for that matter) more than 1-2 min without artifacts, glitching and system crash. I played 3 separate sessions all over 20 min without a single glitch. I'm hoping this new driver somehow resolved whatever issue I had. As i said im still hesitant to say its resolved but its looking good, just gonna give it some time to prove itself.

As happy as I am that the issue seems to be resolved so far, I am equally perplexed by the entire last couple months of trouble shooting. If it was in fact the nvidia drivers, why didnt reverting to various old drivers (including the drivers on nvidia driver disc that came with card) fix the problem? Even after not one but two complete reformat and fresh windows installations, the problem persisted no matter the driver version (this led me to believe it to be hardware issue). Yet problems only occurred while nvidia drivers were installed which lead me to believe it was a software/driver issue. Can drivers somehow cause changes to bios or some internal setting on the video card itself that would persist after software changes? This entire issue came completely out of the blue (i hadnt installed new programs, changed any system settings or bios settings, or updated any drivers for anything in my computer for at least days before and after the problem started). And now it suddenly works after the newest driver (344.11). I'm convinced it was a driver issue now but it still doesnt make any sense to me.

I may never fully understand what happened but as long as it continues to run smoothly I guess it doesnt matter all that much :) I'll keep some updated posts if i learn anything more. Thanks for all your help and suggestions :) it was nice to be able to bounce ideas off of someone else and get some input.
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545
Just thought I'd post an update. Waiting on mobo rma. The asus tech rep was about 99% convinced that my mobo is faulty, just wants me to check another harddrive since the boot device led was on before sending the rma. Oddly enough, I've looked before and this light wasnt on but it was on while i talked with tech rep. I looked again a bit later and it was off. I'll repost updates if a new mobo helps.
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545


RMA motherboard arrived. SAME. FRIGGIN. PROBLEM. the second video card rma will be arriving on monday. if this doesnt fix it i dont know what will. RMA the cpu even tho it ran prime95 for hours with no errors or warnings? RMA the memory even though ive tested booting with individual sticks with no different result? I feel so demoralized right now....i feel physically sick to my stomach....
 

bravodown

Distinguished
Dec 5, 2012
73
1
18,545
Finally got my computer running again. Thought I'd post an update. Overall, I did RMA in this order: video card, power supply, motherboard, video card again. I now works again after 2nd video card rma. Either my original video card was dead and evga sent me a faulty replacement on first rma. OR both original video card and mobo had issues and replacing both was required (although this is less likely). Either way, solved.
 

TRENDING THREADS