Gfx card problem?

Associate
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
Ok, since building my new pc a few months ago, i've had a few Physical Memory Dump's, that happens just after i log in. But i dont think that had anything to do with it.

But the graphics card had stopped quite a few times while in mid game. Never when im not gaming, i'll hear fans power down and then my moniter goes into standby mode while the fans on the pc still running. Once i reboot and get back into windows Catalyst Control Center (CCC) will have some error and say its stopped working.

Now, i actually typed those 2 wee bits out and decided to search for the problem. so i seen someone say about upping the virtual memory, so i did. It lasted around 10-15 mins. then it stopped. i did notice the fans were still running when i opend the case up, also this time, no CCC error. Its not overheating because the temp inside the case is only around 35c.

I cant test to see if it is the gfx card by plugging the DVi into the mobo because ive got a RIVE. I'm gonna install the latest beta driver and see if it helps.
 
Soldato
Joined
23 Nov 2002
Posts
9,597
Location
Near Bristol
You say it isn't overheating, but have you actually monitored the GPU temperatures while it is under load? Use something like HWMonitor and see what they peak at.

What is the spec? Are you on the latest motherboard BIOS? Is the system overclocked? Does it happen in all titles?
 
Associate
Joined
21 May 2013
Posts
1,973
It's possible the card might be faulty. I put a 7850 in the girlfriend's PC a while ago, and upon running any kind of game or 3d app for a while it would just black screen and not respond (screen went in to standby mode like you describe). The only way I could get a game to run properly was to massively underclock the card, albeit this caused choppy performance. Returned the card for a replacement and the new card was perfect.
 
Associate
OP
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
Ah thats the program haha, couldnt think of what that was, cheers downloading now. to be fair im only goin by my tower temp. But i'll check that now.

3820 (stock)
RIVE
16gb corsair dominator platinum
7970 ghz edition.

Everything is stock, and i havent updated the bios. so i'll install beta drivers, update bios and then i'll test the temps.
 
Associate
OP
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
Ah wow, that one lasted a whole 3 mins haha. Is their anyway i can monitor the temps while playing? i see it up to 62c but whenever i came out to check, the temp would drop. i went back in and i heard the fans speed up, and that was it.



EDIT: Ok i tried a stress test program, it gets to 70c, then cuts out. It's been fine the first 2 months of having the card, the last month its just been playing up now and again. but not its just annoying. will do another test to make sure 70 is our magic number...

EDIT2: yup! doesn't like anything too stressful.
 
Last edited:
Soldato
Joined
23 Nov 2002
Posts
9,597
Location
Near Bristol
What power supply do you have?

70c isn't really high enough to cause problems. It is likely hardware instability caused by a faulty component, usually thats the PSU or the part in question, the GPU but don't rule anything out. Do you have another known working system you can try the GPU in?
 
Associate
OP
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
I've got a Seasonic 860w Platinum.

I dont have another system, but i guess i could take it round to my friends house and try it in that, maybe tomorrow night.

Cheers for the help guys, will let you know how it goes.
 
Associate
OP
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
Son of a man nipple! It survived! Got up to 67c then went back down to 63c for the rest of the time. Soooo are we looking at the memory? Since I had some memory dumps? Or do you guys think psu?

Really annoyed, was hoping it was the card.
 
Soldato
Joined
23 Nov 2002
Posts
9,597
Location
Near Bristol
Try running Memtest86+, if it does 2 passes with 0 errors then you can likely rule out memory. If you get loads of errors then you'll likely need to replace it or at the very least alter the settings to match what they are rated for.
 
Associate
OP
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
Well its currently still going, but I wasn't even at 2% and it says 'Hardware problems were detected' any point in continuing? I even put it on extended test, takes ages!
 
Soldato
Joined
23 Nov 2002
Posts
9,597
Location
Near Bristol
Sorry I should have said, ideally you want to be running the bootable DOS version of Memtest86+. There is a bootable USB flash stick creator, the one with *NEW* next to it here: http://www.memtest.org/#downiso

That way all of the memory is tested and nothing can interfere. It will take time yes, do it when you know you won't be using the computer, the longer it runs the better but 2 passes should be enough to establish that the memory is fine. The time taken for 2 passes depends on the amount of memory and the speed, set aside 2-3 hours. That being said, I'd monitor it for the first 10 minutes, that way if you get loads of errors right away it saves running it when you know something is wrong.

If you have to burn a non-rewritable disc then I'd suggest Ultimate Boot CD (UBCD) instead as it saves wasting a disc for just one tool: http://www.ultimatebootcd.com/
 
Associate
OP
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
Oh sorry haha, I forgot to say, I tried making a bootable USB stick but it didn't work, so I done the windows one? Windows memory diagnostics. It reboots into dos too? I'll cancel and try the USB stick thing again.
 
Associate
OP
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
I got it cheers, their was some other option, so I had to manually boot from the bios. If I'm reading this right, I've only got 1405 errors at 60% through. Lol?
 
Soldato
Joined
23 Nov 2002
Posts
9,597
Location
Near Bristol
lol, sounds like the memory is faulty! Most come with lifetime warranties so you should be ok.

Would pay to double check that it is running at the corrects speed, voltage and timings. If you remove a module it should state on the side what it is rated for. Then double check that this is shown CPU-Z/Memtest86+. Voltages can be checked within the bios.
 
Last edited:
Associate
OP
Joined
30 Oct 2010
Posts
2,085
Location
Sunny Scotland
Right so no point carrying on with the tests haha, I'm up to 2814. Will see if I can see if I'm running at the correct times and voltages, never been the best at it haha.
 
Back
Top Bottom