Sign in with
Sign up | Sign in
Your question

Black screen -> Restart on Win7 x64

Last response: in Windows 7
Share
June 17, 2014 6:51:15 PM

I'm having some trouble tracing a bug and hope you folks can help:

My PC has started shutting out to a black screen / resetting, typically while playing games / watching shows. I have 3 monitors and my typical evening flow is one monitor with a maximized, windowed game (Rift, The Forest, Rust - nothing bleeding edge), one monitor with a maximized browser window playing Netflix / Aereo / Amazon and one monitor empty (sometimes off) for pulling new tabs or game guides up in.

A few months ago I had an unexpected restart while playing Rift & watching Amazon. I ran a component monitor and noticed that one of my two GTX 460's was running pretty hot while gaming. After 8 more blackouts in the following weeks I chalked it up to my older graphics cards getting overloaded. Today I replaced my SLI 460's with a single GTX 760 4gb. First run after driver updates I loaded full-screen Amazon in one monitor and tried to load The Forest at Medium settings in the other. Amazon ran fine but I get a blackout/restart when The Forest attempts to render its opening cutscene. I've found now that even on the fresh card I can't run The Forest by itself - my old pair of 460s could run it solo without error. So I guess I've ruled out the GPU. :(  Event logs after these crashes only register "Kernel-Power" failures - no app warnings or Windows notices associated with the pre-crash.

Some specs:
- CyberpowerPC build from 2011 with my own additions (+1 GTX460, +multiple hard drives over time).
- Win 7 x64 Pro
- Core i7 2600k (no OC, 3.4ghz)
- 8gb of 1600 mhz RAM (4x 2gb)
- MSI P67A-GD65 motherboard
-- http://www.msi.com/product/mb/P67AGD65.html
- 800w "XtremeGear" PSU
- GTX 760 4gb GPU (brand new)
-- replaced a pair of GTX 460s in SLI, both from 2011.

Things I've done:
- memtest / Windows Memory Diagnostic (all passes on both)
- Intel CPU testing (all pass)
- Try new surge protector / direct wall power.
- Driver updates for mobo & display adapters.
- BIOS update on mobo.
- Malware scans (just in case)
- SeaTools HD tests (running these now)

My next thought is PSU - I know the Cyberpower defaults get shaky reviews and 3 years is a good lifetime for a PSU that's been worked as hard as mine has. But maybe I just need more RAM? I don't want to keep chasing false leads and I can't afford a full rebuild right now. Any tips for pinning this down, or does this experience sound familiar to anyone who found a solution?

Thanks in advance for any tips or consolation! Happy to answer any questions or provide diag reports if needed.
June 17, 2014 8:53:24 PM

scout_03 said:
use this for the minidump files http://www.resplendence.com/whocrashed


That's an awesome utility - thanks for sharing! It doesn't shed any light on my current issue, though.

Unfortunately, the dump files don't include any reference to my blackout crashes. I even force-replicated a crash to see if it would update. WhoCrashed does report on some sporadic nvidia driver errors that my system was able to recover from over the past year or so, as well as a driver bug from 3 years ago when I used a wireless card for internet. Wow. :) 

Whatever is happening to cause the blackouts isn't giving the computer time enough for a memory dump. I guess that's a revealing detail in itself.
m
0
l
Related resources
June 18, 2014 8:40:06 PM

scout_03 said:
check in the bios if the power lost is on restart so make it not then you will have time to read the code on the blue screen page this power on restart http://www.tomshardware.com/reviews/bios-beginners,1126...


Thanks again for your input, scout_03.

My BIOS setting for unexpected power loss is set as "Power Off". I'm also not seeing any sort of blue screen error - my monitors go directly from screen to black, then I hear the fans spin up for restart.

Updates:
The results of my SeaTools HD tests were limited. I had one drive out of three fail SMART, but it's a secondary drive storing pictures and music - has no data associated with any games on it. My other drives all cleared all tests.

I ran MSI Afterburner with auto logging on during a crash. The logs show no significant spikes in CPU temp or load. I was hoping I might see a major temp spike prior to crashing, as my mobo has a thermal protection shutoff and this would narrow the issue down for me. For anyone interested: Drive download of .hml is here.

I have a Corsair CX 750w that will arrive from Amazon tomorrow. My plan right now is to replace my XG 800 tomorrow afternoon, then buzz through the same tests I've been doing to see if PSU is a contributor. Fingers are crossed that this will be the issue.

Any chance this could be a driver or software issue? My OS is on a dedicated drive so reformatting would be pretty simple - might even be a good time to jump to Win8. I hate to go through all the trouble if it's not part of the problem, though.

Current plan:
- Replace PSU and try to overload.
- Install older nVidia drivers to see if I notice any changes.



m
0
l
a b C Monitor
a b $ Windows 7
June 19, 2014 12:43:49 PM

the drive who fail need to be take out the system ,try to test the gpu in another system and you could also try the gpu in the scond pci-e slot to see what happens on the monitor . if you could stay with win 7 .
m
0
l
June 19, 2014 12:54:03 PM

scout_03 said:
the drive who fail need to be take out the system ,try to test the gpu in another system and you could also try the gpu in the scond pci-e slot to see what happens on the monitor . if you could stay with win 7 .


Thanks scout_03. Good idea on the PCIe slot - I'll swap that and pull the extra drive when I swap the PSU this evening as well.

The new GTX 760 I got was fresh out of a sealed box. Is there enough risk with that to be worth testing it on another system? The issue is the same with the new GPU as it was with the old pair. I don't have a spare desktop that can support this GPU on hand so if that's a real risk I'll have to try and borrow one from a friend.
m
0
l
a b C Monitor
a b $ Windows 7
June 19, 2014 1:04:44 PM

if you had the same issue with the old gpu then it could be the motherboard if you still have that old one try it in the second slot .
m
0
l
June 20, 2014 7:28:21 AM

scout_03 said:
if you had the same issue with the old gpu then it could be the motherboard if you still have that old one try it in the second slot .


A quick update: I replaced my OEM XG 800w PSU last night with a brand new Corsair CTX 750M. Since replacing it last night I've been able to load/play games at max settings with no blackouts. I've also been able to multitask again - for example, one fullscreen instance of Rust at max settings, one fullscreen Netflix video, one fullscreen browser to page to for guides - with no lag.

My fingers are crossed that the PSU was the issue. I'd like to wait a day or so to keep testing before I mark this issue solved, just to ensure it's a real fix. Thanks!
m
0
l
a b C Monitor
a b $ Windows 7
June 20, 2014 8:06:34 PM

if that oem psu still under warranty ask for a rma of it .
m
0
l
!