Sudden GPU Power surge and black screen

[tolufan]

Hello everyone, first of all I have very little to no technical knowledge so please forgive me for using potentially wrong terminology. My problem started when I bought this new laptop which has Geforce 840m as the graphics card. I have been getting a frustrating black screen problem recently. What happens is, for instance when I'm running a game, the screen suddenly goes black but everything else seems to be working. I did some research and downloaded Techpowerup GPU-Z to monitor the GPU. And I went through the logs after the "crash" and what I saw was this:

Date , GPU Core Clock [MHz] , GPU Memory Clock [MHz] , GPU Temperature [°C] , GPU Power [W] , GPU Load [%] , Memory Usage (Dedicated) [MB] , Memory Usage (Dynamic) [MB] ,
2015-01-23 22:04:25 , 1050.0 , 800.0 , 55.0 , 0.3 , 1 , 0 , 129 ,

2015-01-23 22:04:26 , 12750.0 , 800.0 , 45.0 , 0.2 , 1 , 0 , 133 ,

2015-01-23 22:04:34 , 400.0 , 800.0 , 44.0 , 32764.3 , 0 , 0 , 133 ,

2015-01-23 22:04:35 , 1050.0 , 800.0 , 51.0 , 0.1 , 0 , 0 , 134 ,

2015-01-23 22:04:36 , 1100.0 , 800.0 , 44.0 , 0.1 , 1 , 0 , 144 ,


Sorry for the messy copy&paste, but my point is that GPU power momentarily rises to unnaturally high levels and causes the GPU to shut itself down I guess. What could this mean? I checked the other black screen threads but they mostly revolve around temperature and overheating, mine seems to have a power problem rather than overheating as the temperature stays fine.

I would appreciate any help possible. It's been only two months since I bought this thing and I don't want to lose 20-30 days with the tech guys if it's possible to find a fix around here. Thanks in advance!

 

[gilbadon]

totufan, I think your error has to do with software, not hardware. I think your hardware is fine, at least from what you had given us here, everything looks normal.

For instance, that HUGE value you are seeing for power 32654.3W. That could power your neighborhood, so that value is not true. What the graphics card actually read (oddly enough) was a negative value.

See, how computer science works, is that number is an integer. Once it goes into the negatives, it counts back around to the highest number the integer can hold (in this case 32767). So in reality, your GPU measured a negative 2.7W use. This makes a little bit of sense (the sensor is a bit off of coarse) because the core clock dropped to a measly 400MHz which shows it is not being used much.

The point I am getting at through that whole mumbo jumbo of nerd speak is, power usage is not your issue and because the other readings are low as well, it seems your GPU is not a culprit and is not even being fully utilized. You may need to simply update your graphics display drivers and give it another go.

 

[tolufan]

Hello gilbadon, thanks a lot for the lightning-fast answer! Another reading of the "crash" tells me that it's 9979 this time, but I'll apply the same logic to that as what you explained about the integers.
As for drivers, at the beginning I had a driver version of 332.21 IIRC and I replaced it with the latest version, but even that did not help. That's why I thought it was perhaps a hardware issue.
I'm running Windows 8.1 on my PC. When I was installing my new driver for the GPU, I chose a setting like "remove the old driver completely" to make sure it was gone and the new driver was fresh. But I think it somehow failed. Should I perhaps try getting rid of Windows 8.1 altogether and start from scratch by installing the new drivers and all?
Thanks in advance once again!

 

[gilbadon]

Try using the program ddu, forgot what it stands for, ... Driver uninstaller or something. Just Google search ddu lol. It is the best tool to start driver installs from scratch.

Also check if windows is generating "memory dump" files. If it is, the answers may be in there. Unfortunately I do not know how to read them myself, but someone else will.

I'm off to bed for the night. If you need more assistance let me know and I will get back to you asap.