Computer shuts down while gaming with no bluescreen

Toggle sidebar Toggle sidebar
T

TheQuantumLlama

Honorable
Feb 8, 2014
23
0
10,510
After playing a graphically intensive game for a bit (anywhere from 2 min to 30 min), my computer shuts down with no warning at all. No BSOD, my computer just shuts down and after about 3 seconds restarts. Using WhoCrashed, I found a couple instances of a "VIDEO_TDR_ERROR" where atikmpag.sys is referenced as the most likely culprit. I don't know how to diagnose this problem as it does not cause a BSOD, although i can reliably cause it by starting a highly intensive game such as Shadow Warrior or Skyrim.
I have tested my CPU with Prime95 for about 3 hours, the load went up to 100% at some points but the temperature never topped 79 C.

Specs:
Intel Core i5-4670K (not OC'd) w/ Cooler Master Hyper 212 EVO
MSI Z87-G45 Gaming
Radeon XFX 6950
Crucial Ballistix Sport 8GB
Thermaltake 500W
 
Sort by date Sort by votes
I've been thinking the same thing about the cooler but I'd rather solve one problem at a time. Is there any way to test my PSU? I've tested it with a multimeter recently and all the voltages were good, I doubt this is the same as when it's under load but I don't know of any way to test it while its under load.
 
Upvote 0 Downvote
I don't know of any easy way to test your psu under load. I guess you can try having the multimeter hooked up while your system is under load.

Try running furmark (gpu stress test) and monitor your gpu temps as well. See what happens. Normally, driver and software related issues BSOD or freeze.

My old psu, when I tested gave 12v on the multimeter, but it woudnt turn on my PC at all.
 
Upvote 0 Downvote
I just downloaded it and tested using the default settings and my computer instantly shut down and then restarted, which it normally doesnt do, usually it just shuts down and then stays off. Not sure if that difference means anything. My GPU temps didn't even have time to display.
 
Upvote 0 Downvote
Ehh that's a pretty low quality psu, so its not too strange that it would be giving you issues. I would get a new PSU. There really isn't any other component that would cause a shutdown of the system, when its under load, especially when the temps are okay.
 
Upvote 0 Downvote
The antec HCGs are great units, and 750w is more then enough for your current setup. For dual 770s or 280x's I wouldn't get less then 850W. Unless you want a super loud psu and the psu failing after a couple months.

Personally, I would use a 1000w for such a system. This is pretty good for the price:http://www.newegg.com/Product/Product.aspx?Item=N82E16817139057

But again 850w also gets the job done:http://www.newegg.com/Product/Product.aspx?Item=N82E16817438018 or http://www.newegg.com/Product/Product.aspx?Item=N82E16817139011

if you want, you can get even higher quality platinum rated psus.
 
Upvote 0 Downvote


simply put your cpu is overheating

 
Upvote 0 Downvote
Would 750 be enough to run just one 770 or 280x? Or should I go for the 850? Also, how did you come up with the numbers you gave me? I've used a couple psu calculators and they all give radically different numbers, one on Newegg gave me a wattage of ~1500 for dual cards but then this one, http://extreme.outervision.com/PSUEngine, gave me a minimum recommended wattage of ~700 for the same dual 280x setup.
 
Upvote 0 Downvote


When I'm gaming and before it shuts down my GPU goes up to 80 or 85 C max (I forget which), but I can be sitting on the menu of a game with my GPU temps at 70 or 75 and still crash.
Edit: Missed you said CPU, leaving the GPU part there in case it is relevant. My CPU doesn't even top 75 C while gaming (and that's being generous). I tested with Prime95 and my max was 79 C with no crash.
 
Upvote 0 Downvote


85 degrees sounds very high sounds the gpu is damaged do you have another gpu to test?
 
Upvote 0 Downvote


750w is more than enough for a single card setup and even a midranged 2 way CFX/SLI config
 
Upvote 0 Downvote
I just tested again to make sure I was giving the correct GPU temps and found that it's gotten even worse in the past month or two that I've been trying to solve this problem. Just sitting on the home screen of Warframe got me up to 83-84 C, when I went into a game and did something graphically taxing my temps got up to 89 C before my computer shut itself off. And no, I dont have another GPU to test with.
 
Upvote 0 Downvote
@DarknessRising

1. His CPU is NOT overheating. If you read the earlier posts, I had the OP run furmark and it crashed instantly. Furmark does not test the CPU at all.

2. Having temps in the 80s for a GPU is common, especially AMD cards with reference coolers. And when GPUs overheat, they throttle, they do not have the ability to shut down the system.

3. You are right that a 750w psu can "run" an SLI system. However, this is based on the peak output of the psu. Keep in mind PSUs are NOT designed to run at 100%, rather to run at 50%-60%. This is also the range in which psu's are the most efficient in providing power. Running a PSU close to its limits SIGNIFICANTLY effects its lifespan and it also becomes super loud.

@TheQuantumLlama
GPU calculators are often contradictory, as you saw. Some provide recommended PSUS and try to push sales of a specific PSU, while other provide exactly the power your components are going to use (i.e. they add up the manufacture specified wattages for each component). I use the numbers provided by the latter method of calculation, and add about 100-200W for some headroom to ensure your PSU runs comfortably and not at peak capacity.

A 750W is more then enough for a single 770, while 850W+ is necessary for two. A decent 600W can also easily run a 770.

Also in terms of 12v amperage, most new psu's (good quality) can provide almost all their wattage on the 12v, so youre safe there.

 
Upvote 0 Downvote
Regardless of what might be safe for the card, I don't really want it throttling itself, so I'm going to try and get a new one under warranty because it has heated up in the past couple months and go from there. If I still have a problem with shutting down, I'm going to buy an Antec HCG 850W PSU, that should be enough to run whatever i decide to throw in. Thanks both of you for all your help, anything I'm missing?
 
Upvote 0 Downvote


take out the graphics and boot without it see if that fixes your problem
 
Upvote 0 Downvote


looking at the model of the card there is no way that 80 degrees is okay
 
Upvote 0 Downvote
Power Supply Calculators are great i use eXtreme Power Supply Calculator its a very good way to know how much wattage you'll use i recommended at least a decent branded 600w 80+ for a gaming computer Certified psu like Antec,NZXT, Enermax, Corisar, XFX, and EVGA for a good cheap psu i'd get the EVGA Super Nova NEX750B i use it my myself and you could use it for 2 way CFX/SLI config for mid ranged gpu
 
Upvote 0 Downvote
The OP is RMAing his card anyway, but please let us know how you found that 80c is bad for this card. It doesn't hurt to post some links to back up your claims.

A simple google search yields a consensus that up to 95c is safe for this specific card.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom