A Complex 7970 GPU problem [BLACKSCREEN - SOLVED]

[InSpace]

Hello boys aaand girls

I don't know where to post this so I'm posting this here, you guys have a very active forum and as far as I was reading you guys have pretty professional people here

I have a serious issue that'a arisen recently. I have been getting black screens on my Gigabyte Radeon HD7970 [Exact Model: GV-R7970OC-3GD] which is a GPU with a custop 3 fan cooling system overclocked to 1000MHz on the core and to 1375MHz on the memory, this is a factory overclock.

I have started getting these black screens about two weeks ago with and their intensity has been increasing. It usually happens when I play games or watch youtube videos. Sometimes tho, it does it randomly, when my PC is idle. I have tried a miriad of things to fix this, to no avail. It usually ends up with me not being able to do anything, I can usually still hear the sound, or talk to my friends on comms still, sometimes it crashes my sound as well. and sometimes white stripes appear.

Before I get into what I've already tried to fix this, I am going to post my complete system specs.

CPU: i7 2700k @3.5 GHz
CPU Cooler: Noctua NH-D14
GPU: As stated above - GV-R7970OC-3GD
RAM: Corsair Vengance Mini 8GB with 2 sticks each 4gb @1866 per stick
Mobo: P8Z68-V PRO/GEN3
HDD: Seagate Barracuda 1TB
I will list the monitor as well which may play a role in this: Samsung SyncMaster B2240


I think it's either a hardware issue or driver issue, I have been unable to locate it, and I cannot RMA my card since it is 2.5 years old now, bought it pretty much when the 7900 series launched. And I do not have the money or the means to get a new card now. I am hoping it is a driver issue, because if not, I am screwed.

Things I have tried.

1) Inspecting the card, plugging it out and back in.
2) Update my drivers, all of them.
3) Resetting the drivers to older versions.
4) Updating my BIOS
5) Updating my VBIOS
6) Underclocking and Overclocking the card
7) Changing up the voltages, although not singificantly (-5%/+5%)

After I did all this the problems increased. I cannot play a game for longer than 2 hours without the gpu crashing, and when it crashes and I reboot, the crashes get even more constant, I have to let it sit for a while.

The GPU temps are fine, what I dont know is are the VRM temps, and I dont know how to check 'em.

Also, in GPU-Z the VDCC is stable at 0.803-0.805v, although the VDCC current is fluctuating heavily, is that normal.

My questions to you guys are,

1) How do I check VRM temp
2) Is the VDCC Current fluctuation normal
3) What else can I do to fix this damned Black Screen issue?

In advance, thanks for your help, If I get this fixed, I'm gonna be very greatful.

 

[newtekie1]

What exactly are your temps under load?

Have you made sure all the fans are spinning?

Have you cleaned the card?

 

[Solaris17]

those white bars scare me into thinking its vmem

 

[InSpace]

@newtekie1 The GPU temps are normal, even under the heaviest loads the card doesnt go above 67 celsius
Fans are spinning, yes.
I did

@Solaris17
how can I check my VMEM activity / temp?

 

[eidairaman1]

Power supply?

 

[InSpace]

I hope not, I checked all the voltages and most seem nominal

 

[InSpace]

Also I have a question, My card, HD7970 has OpenCL support yes? GPU-Z is constantly telling me an error about it that it cant detect it or something and when I view the OpenCL isnt checked, I removed my drivers multiple times and even ran Driver Sweeper but it didnt work

is there a way to install openCL seperately without having to install catalyst? I already have that.

 

[TRWOV]

hummm... with all the stuff you describe I'd think that the GPU has damage on the solder pads or maybe the GPU itself. I've got the "not waking up" bug from time to time but not a black screen out of the blue.

OpenCL is built in into the driver. GPU-Z not detecting support for it means that something is wrong.

VCCD fluctuating over load is "normal" but it shouldn't vary a lot, say 1.30-1.32v. If it's doing something like going from 1 to 1.35v that's definitely not normal. Maybe the VRMs are shot and can't provide the card with stable power. Have you tested with another power supply?

 

[Ja.KooLit]

last resort before hardware issue. Reformat. Back up all your important stuff. Then completely reformat your windows installation. Then from there, install all updates and drivers and so.

If still doesnt resolve your issue, then you can rule out the driver issue. It is hardware failing. Probably PSU is dying. what exactly is the model of your PSU? you may say it is normal voltage but what about during load? You have to consider capacitor aging thus affect the performance of the PSU.

 

[de.das.dude]

dont these cards have 3 year warrant atleast? you should be able to avail it. its your best bet since problems sound like a h/w issue.

 

[AsRock]

If you live in the US at least the card should have a 3 year warranty with it.

http://www.gigabyte.us/support-downloads/customer-service.aspx

 

[qubit]

Unfortunately this does sound like a hardware fault with your graphics card. Those white stripes are very telling of video memory failure.

Likely there's one or more bad solder joints somewhere on the card, but they could be anywhere and impossible to fix without specialized equipment, which would be uneconomic anyway.

Many cards have a three year warranty. Are you sure yours doesn't?

 

[GreiverBlade]

the only time i had that issue was with a 7870ghz ed. and it was a known issue called BSOD (Black Screen Of Death no the Blue one ) the only solution i found was to downclock the card around 950mhz core (meme untouched) which also incidentally solved the coil whining issue the card also had ... (Club3D never again ... )

 

[InSpace]

!!SOLVED!!

I did two things, dunno which one solved them but I'm betting on the second one

1st - I unplugged the GPU and cracked it open, it was a pretty bad ordeal since it had a custom cooler and that was very hard to put back together, I didn't see no damage in there, although quite a bit of dust on left and right corner of the heatsink. That shouldn't be a problem tho because my card's heatsink is massive and the dust took only about 20-25% of it. Then I put back together, inspected the PCI slot, dusted off that as well and plugged it in. That was also not fun because of my humongous CPU cooler Noctua NH-D14, also when I put my PC "back on its feet" something cracked inside that sounded like a PCB cracked, that scared the ever living shit out of me and I was anticipating what happens when I turn it. Maybe I've made it even worse.

Pictures is here! Prepare to feast your eyes, I made this PC 2.5 years ago when I was about to turn 17 and It was my first build. there is literally no cable managment, I just made sure they're not hitting the fans. Also, I dust my PC about twice a year only, and alas, it still wasn't problem, it will be explained below.

2nd - So, When all this started, I just associated the issue with my GPU, which should have been, or be accurate when I thought about it at the time, but in retrospect, it was a gamble, turned out to be jackpot tho. After I did that I started reinstalling drivers, dozens of older versions, no drivers at all etc. BTW the "no drivers" worked but it puts the card into literally 2D mode, which makes playing games and doing other 3D related activites like 3D modelling impossible, so naturally, that wasn't a solution, and so the battle went on, all nighters searching for solution, trying out different things, fixing things I didn't bother doing before a problem was before me like updating bioses on various things, cleaning the PC components, stuff like that. In doing so tho, especially with the catalyst drivers I made my PC even more confused, You see, it had so many drivers still in its registry, or somewhere else for that matter that it was still probably dragging data from the old ones, which were overlaying with the new ones and as I installed and tried new and new drivers, the PC grew even more and more unstable. Well I was getting desperate, I didn't even count this as a solution but I did it anyway because I didn't know what to do and apperantly it can help. I downloaded DriverSweeper. I did a sweep, restarted the PC, and logically windows started to look for a driver for the card itself, I didn't even try to stop it because I was losing hope and started to stop caring, abit. Then I turned on a game and played it for 1 hour. hmmm this is weird. 2 hours wow is this fixed, nah can't be I went longer without a crash. 6 hours almost forgot about the whole ordeal, being consumed by the game. But when I rememberd it, I was extatic. 24 hours, yes I stayed up 24 hours playing a game, and no crash what so ever. I was pretty happy with it and went to sleep. The next day I played for about 4 hours and everything went smooth as well. It's fixed.

Although, I do not have a catalyst installed, just some driver that windows downloaded, and GPU-Z is still not picking up OpenCL. I would try to install new drivers, or something like that, but I'm scared. Hey if it's not broke, don't toutch it ey?!

Thank you all for responding to this thread, and even tho your forum is great and people here a highly professional and friendly, I hope I won't have to come here again, because if I do, it will be because I have another PROBLEM! Goodbye guys.

 

[Blue-Knight]

!!SOLVED!!

I love this word. It is so good when we have a problem SOLVED.

I hope I won't have to come here again, because if I do, it will be because I have another PROBLEM! Goodbye guys.

Then I hope you never come back. Goodbye!

 

[de.das.dude]

you didnt install any AMD beta drivers did you? They fucked up my system completely. i had to reinstall windows. nothing worked. registry, sweeper, etc.
affected the 7790 and the 7970.

 

[blacky white]

Can anyone identify this? My monitor is not detected on Generic Non PNP Monitor? Can anyone help me?

Thank you

http://www.techpowerup.com/gpuz/details.php?id=azfy

 

[eidairaman1]

With that pick above either you have that machine out in a garage or in a damp area where rust or dust can happen. Desktops are meant for climate controlled areas only.

To the guy asking. Does the monitor even come on?

 

[mx500torid]

That pic looked very ugly dust everywhere and I think I see a OCZ power supply. I think I would check that out first then do some housecleaning. Gotta be moisture getting to that comp.

 

[InSpace]

@de.das.dude Yes I did I am not sure when tho, Maybe they were the cause, I think it was version 14.11.2 beta, stay da fuq away!


@eidairaman1 No, It is my room, problem with it is, it has 3 carpets and a 30 year old bed sitting around, that circulates the dust atleast 3 times as much as it would be normally. )

@mx500torid It is fixed, it was a driver issue. and yes that is an OCZ ZX-Series 850W PWU. Never had problems with it although under the computer the dust accumulates at 6 times the rate, and even tho I have filters everywhere on every fan, I just cant stop it so I have to clean the the PWU filter frequently, that is the only thing I clean frequently on my beast.

 

[de.das.dude]

@de.das.dude Yes I did I am not sure when tho, Maybe they were the cause, I think it was version 14.11.2 beta, stay da fuq away!


@eidairaman1 No, It is my room, problem with it is, it has 3 carpets and a 30 year old bed sitting around, that circulates the dust atleast 3 times as much as it would be normally. )

@mx500torid It is fixed, it was a driver issue. and yes that is an OCZ ZX-Series 850W PWU. Never had problems with it although under the computer the dust accumulates at 6 times the rate, and even tho I have filters everywhere on every fan, I just cant stop it so I have to clean the the PWU filter frequently, that is the only thing I clean frequently on my beast.

well... ever since it happened to me i dont install beta drivers. furthermore i always look out for a month atleast, if a certain driver is worth downloading. Drivers seem to always go wrong for 7790s and 7970s.

 

[st2000]

got the same issue on my old 550ti, replaced it with 280x and problem is still here, so i tried a lot of things but the working solution is to change hdmi on tv(switching it between 1-4 ports)
problem is more ofen in 2d mode after turning TV on, so dont know if it's cable issue or TV issue(also never had this problem after 3-4 hours working TV with 280x,only with 550ti)

 

[InSpace]

Well, Do the driver sweeper thing, or try to reinstall completely.

 

[InSpace]

The problem is back, although in a weird way, whenever I sweep the drivers, its ok for a while, like a day or sometimes more, and then it starts having these crashes.

I think the drivers creates either too much files over time or a file thats messing with something...

Gonna completely reinstall windows after weekend

 

[Devon68]

Yeah my pc didn't want to turn on today (just one long beep), but yesterday it was working fine. So I just smacked it and it started working.