Computer crashes during gaming but not during stress tests

Toggle sidebar Toggle sidebar
Status
Not open for further replies.
G

Guest

Guest
This has been going on for a couple of weeks now my computer would crash and I would have to hard turn off the system. When I play higher graphical games I usually crash, but currently I am running both prime95 and furmark's burn-in 1920x1080 tests at the same time and it's not crashing like it does when I play games

and my temps(DURING BOTH THE STRESS TESTS):
cpu around ~93-96° C
gpu at 66°C (fan is at max speed while test is on)
(THESE ARE WITH BOTH FURMARK AND PRIME95 AT THE SAME TIME)

My temps while gaming are lower than these, and it still crashes

I've tried clean installing drivers (new and old)
I've ran mem test over night no errors
I've ran prime95 by itself for 5 hours and it was like 60-ish tests no errors
I've ran furmark burn in test multiple times for 15 min each

Screenshot of the three most recent WhoCrashed reports:
http://i.imgur.com/iq8FzUM.png


parts:
Mobo: ECS Z77 H2-A3
Ram: G.SKILL Sniper 8gb
Gpu: EVGA GTX 660 ftw sig 2, 3 GB
Cpu: i5-3570k @ 3.4
Psu: Corsair CX 750M 80 + bronze (just got this recently)

any help?
 
Sort by date Sort by votes
A

anthony8989

Distinguished
Feb 2, 2013
653
2
19,165
Wipe your graphics driver and re-install using the latest drivers from nVidia. The problem is the definitely the graphics driver.

On a side-note - please get an aftermarket CPU cooler for your i5. That thing is running way too hot. You'll shorten the life of the CPU if you keep running it at 90+ Celsius for prolonged periods of time.
 
Upvote 0 Downvote
J

junkeymonkey

Polypheme
BANNED
Nov 11, 2013
17,389
4
66,960
just to take a stab at this --- remove the drivers for the card and get it back on the generic vga driver built in windows. [theres a footnote on NVidia about not uninstalling the driver from device manager] once your back on the windows driver shut back down and remove the card from the slot and blow out any dust from it take your shirt tail and wipe the ''golden fingers '' once its cleaned up reset the card and hook it back up and get back to desktop and you should still be on the windows driver if so download the NVidia driver to a folder on your desktop and install it from there
http://www.nvidia.com/Download/index.aspx?lang=en-us

also you may want to get this from Microsoft it's one of the first things I get from them when I need to game as you can see its there #1 download

http://www.microsoft.com/en-us/download/details.aspx?id=35

 
Upvote 0 Downvote
17seconds

17seconds

Champion
Jul 27, 2010
10,584
4
51,965
Beginning with the Fermi generation, Nvidia long ago implemented throttling in it's GPUs specifically when Furmark and/or Prime95 are being run. This was intended to prevent damage and reduce RMAs. In effect, it made those applications obsolete for stress testing. Certainly running both simultaneously would compound the issue.

As you are finding out, only through gaming or other stress tests can you get a true measure of your system stability. I recommend you try MSI Kombustor which comes bundled with Afterburner.
 
Upvote 0 Downvote
G

Guest

Guest


Will do sorry for the late reply was pretty busy all weekend
 
Upvote 0 Downvote
G

Guest

Guest


So I did what you said uninstall drivers clean out pci slot wipe the "golden fingers" then re install the latest drivers
I started up HWmonitor and saw that my idle temps went down a good amount.
Then I started steam and played deus ex HR at max settings. the temp of my gpu didn't go above 60 C
I'm gonna test more recent games like Watch dogs and AC IV soon
And I noticed one of the little golden like teeth of fingers was chipped is that an issue I should be worried about
 
Upvote 0 Downvote
G

Guest

Guest

Yeah just found a pic its the same thanks for clearing that question I had
 
Upvote 0 Downvote
G

Guest

Guest
ok so I played deus ex for about an hour and my computer crashed again no DMP file after I rebooted
I guess I should just RMA the card to EVGA and get another one?
 
Upvote 0 Downvote
17seconds

17seconds

Champion
Jul 27, 2010
10,584
4
51,965

That game should be very stable by now. It can't hurt to call EVGA and test out their reputation for customer service.

Just out of curiosity, what's your GPU ASIC score?
GPU-Z > right-click on the Title Bar > Read ASIC quality...
 
Upvote 0 Downvote
G

Guest

Guest


71.9%
 
Upvote 0 Downvote
17seconds

17seconds

Champion
Jul 27, 2010
10,584
4
51,965
ASIC is a measure of your GPUs quality level or "binning". They sort through each GPU and assign it an ASIC score. Those with higher ASIC scores supposedly are more efficient, which means they can run with less voltage and are able to run at cooler temperatures. That in turn means that they can also overclock better.... in theory. Supposedly, a "superclock" video card is going to use a binned chip with a higher ASIC score than a reference card, but that's not always the case.

I gained some interest and insight into ASIC scores when I had to RMA an old GTX 580. The original card had an ASIC score of about 74%. The new card came at 82%. The new card proved to me that ASIC quality was as advertised because it overclocked better and ran cooler. But the real thing that told me that ASIC matters is that it idled at a lower voltage. I looked it up, and sure enough discovered that manufacturers set the card's BIOS to run at a lower voltage when ASIC quality is higher.

So I'm a big believer in ASIC quality scores. Low to mid-70%'s is typical/average, but you want to get in the upper 70's/80's or more for one of the better chips. I even returned a new GTX 780 Ti because it had an ASIC score of about 68%, and I could tell it just wasn't a good quality chip in terms of heat and overclocking. Luckily, the new card was in the mid-70%'s, so I could live with it. 78% is pretty good, so I would say you're lucky.
 
Upvote 0 Downvote
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom