HD 7950 - black screen computer crash

ABBDVD

Distinguished
Feb 23, 2010
71
0
18,630
Hi
I set up a brand new system and now after installing windows, graphics drivers, ... the graphic card seems to fail under light load already. I can start the computer up and it runs relatively stable just browsing... but when running a game like League of Legends, BF3 or a stress test (I used MSIKombustor 2.3) the graphic card fails in less than 5minutes. The monitor just turns black and the system still seems to run but there's nothing you can do anymore. I hard reboot the computer after a couple minutes with nothing happening changing.

My system setup is the following:

-i5 2500k
-Cooler Master Hyper 212 Evo cpu cooler
-Carbide 400R Tower
-Gigabyte GA-Z68X-UD3H-B3 motherboard
-Crucial m4 SSD 64GB (contains the "system" except for the "Users" folder which I moved to my HD with an 'unattended xml install tweak')
-Samsung HD103SJ F3
-Super Flower Golden Green Pro SF-650P14XE (650W PSU)
-Corsair CMX8GX3M2A1333C9, 2x4GB, DDR3-1333 RAM
-Windows 7 Home 64bit
-SyncMaster S24A450MW 24'' monitor (1920x1200px)

And the GPU: Sapphire HD 7950 OC Dual Fan 3GB

I didn't overclock anything (yet). The BIOS settings are default except for some little things as boot order, no MB start screen, ...
I got the newest Catalyst Control Center installed and the MB drivers.

I let GPU-Z log me what happens.... so I let the stress test run:
GPU Core clock is constant 900MHz
GPU memory clock constant 1250MHz
temperature went up to a maximum of 44°C
Fan speed max 25% (1210RPM)
GPU load up to 99%, last log entry it jumped from 90 to 39
full 256 MB memory usage
dynamic memory usage 48MB
12.31V
VDDC 1.086V
MVDDC 1.602V
VDDC current (quite varying) between 11.8 and 46.5
VDDC current IN between 1.1 and 4.3

Link to the whole log: Download Link to the log
(click the "small" download button in the middle...)

I run prime95 also for like 10min. The temperature and everything seemed ok. This isn't reliable but I'd say this at least hints that it's not really the CPU failing here.

So my resulting question is of course: Can someone help me?
I don't really feel like handing my GPU or Motherboard back if it can be prevented and maybe there's something I should/could check to fix this(?)

Thanks for any help!

Regards
ABBDVD
 

ABBDVD

Distinguished
Feb 23, 2010
71
0
18,630
This could be a hard drive problem as well, had my PC freezing like yours a while a go, defragged the drive, repaired the bad sector and everything has been fine since.

What version of CCC are u using anyways?

AMD Catalyst 12.3 (according to the built in update functionality this is the newest version)

Should I run a checkdisk? I'll manually start the windows defragging now anyway.
I did reinstall Windows about 7 or 8 times in the beginning as I had a little trouble getting it right with the Users folder on the F3 Spinpoint. So this could have a point.
 
it may not be the video card but your power supply may not be holding up the load. or it sending to much ripple when it loaded.
you can tell if it the power supply by tryingto run prim95 it more of a mb/cpu test if the system power and memory are fine it will pass if not if it locks up then you make have an issue with the ram and not the video card.
 

ABBDVD

Distinguished
Feb 23, 2010
71
0
18,630


Ok, I'm running the "maximum heat, power consumption" test in prim95 at the moment (5min without trouble so far).

According to reviews I read the SF-650 should hold up well though and at least theoretically 650W is plenty.
I also took care to take two power supply pins from different cables for the GPU.


Update: After 20min it's still running fine. The CPU's only at 42°C (should be some nice overclock potential).
The defragmentation was probably a good idea as my F3 HD was quite fragmented. I'll let the stress test run a while and hope you have more good inputs :) Later on I'll try the Kombustor test again and see if something's improved because of the defragmentation.
 

ABBDVD

Distinguished
Feb 23, 2010
71
0
18,630
I agree that it's more likely to be a PSU problem,

and dont bother stress testing the CPU anymore, I doubt it's the cause.

I remembered that win7 has an event manager showing errors...
All the errors I got so far are labeled Windows Kernel event ID 41.

- System

- Provider

[ Name] Microsoft-Windows-Kernel-Power
[ Guid] {331C3B3A-2005-44C2-AC5E-77220C37D6B4}

EventID 41

Version 2

Level 1

Task 63

Opcode 0

Keywords 0x8000000000000002

- TimeCreated

[ SystemTime] 2012-04-20T14:20:11.663609100Z

EventRecordID 2796

Correlation

- Execution

[ ProcessID] 4
[ ThreadID] 8

Channel System

Computer JPC

- Security

[ UserID] S-1-5-18


- EventData

BugcheckCode 10
BugcheckParameter1 0xffffffffffffffff
BugcheckParameter2 0x2
BugcheckParameter3 0x0
BugcheckParameter4 0xfffff800030db256
SleepInProgress false
PowerButtonTimestamp 0

Those EventData are not always the same and in most cases it shows BugcheckParameters are = 0x0 (BugcheckCode = 0)

If nothing helps I'll replace my PSU. Do you think the PSU just doesn't give enough power?
 

ABBDVD

Distinguished
Feb 23, 2010
71
0
18,630
I believe the PC is running fine now. I set some setting back in the BIOS and apparently that did it. I could play the first 10min BF3, run Kombustor for 10min and run a PCMark7 test. I'll let prime95 run a while too as last time after about an hour the PC crashed... forgot to mention that.
Anybody got a good idea for an alternative stress test?
 

Neyzon

Honorable
Sep 14, 2013
1
0
10,510


Hi I'm new here & I have similar problem w/ my PC HD 7950 I bought it less than month ago & mainly using it for crypto mining. It was doing fine until 3-4 days ago when it began to be unstable - screen goes black after around 1~2 hrs of mining & had to restart my PC.
 

TRENDING THREADS