Sign in with
Sign up | Sign in
Your question

My graphics card is Dying??

Last response: in Graphics & Displays
Share
June 27, 2012 10:28:42 AM

Hey everyone, need some urgent help.

I had a pc made about a year and 2 months ago, mainly for gaming. Over the last year the only thing iv changed is the PSU and the Case, which required me to take everything apart and rebuild into the new case.

The last few weeks, iv been suffering really bad issues with my computers performance while gaming, and over the last 2 days, it has got to a point where i cant run Battlefield 3 at all, as the framerate is jumping from 20 down to 2/3 at times. Now im no tech expert, so i would REALLY like so help on how i can go about finding and fixing the problem.

Iv been cleaning out the fna on my GPU and heatsink, and even though i keep them well ventilated, the GPU temp goes over 60 degrees when i try to run BF3, not that i can anymore as it just freezes up. temps at idle are as follows:

GPU: 50 degrees
CPU: 38 degrees
MB: 28 degrees
Core+ HDD: 30 degrees each

Not sure what to make of the temps, as im no hardware analysis expert, but the GPU seems like its way too hot for idle.

i run a ATI 5770 with the stock heatsink and fan, and i think this GPU is dying/overheating. I am a complete novice at these things, and have never replaced a graphics card heatisnk/fan or applied thermal paste, but if that is the cheap solution to this problem, i would be happy to try it.

If you need anymore info on my pc, please ask, as i am very hopeful someone on these forums can help me out. Right now, i cant play any games, as the framerate just dies and im scared it will kill my already dying GPU.

I, not sure if this is related, but i seem to suffer from VERY slow processing speed the first time i startup my computer in the morning. If i run any program at all, it will just *not respond* and so will task manager when i press ctl alt delete. I then have to wait 15 minutes, and when im back, everything is responding and the computer runs good for the rest of the day. Well, at least it did before this issue with my graphics card. Every morning i go through this, and it has been happening for the last two weeks, even before the GPU started acting up.

Once again, thanks to all in advance for taking the time to helping out a fellow computer user, and i hope to get this ting solved.

More about : graphics card dying

June 27, 2012 10:34:41 AM

Well those temps arnt that bad on the gpu side as far as i can tell but what temps are your cpu when running bf3 the cpu could be throttling, makeing it appear like your video card is running bad..60 c is probly right when running bf3 my 2x480s hit 55 to 60 on water..
June 27, 2012 10:45:05 AM

thanks for the quick reply man, ill run BF3 and tell you the temps. You obviously know much more than me, but yes, as you said, CPU may be "throttling"
Related resources
June 27, 2012 10:48:52 AM

I have just bought a new gtx 680 and that temp at idle is around the 35 to 40 mark however under load it has hit temps of 69 so i would probably relate it to your cpu temp if that gets to hot i have experienced lag and allsorts of crashing due to a cpu being to hot, this can affect the in game play making it feel very laggy which would explain the drop in frame rates. i used to run a 5750 before my gtx 680 and i would experience lag in some intensive parts of the game, try lowing the resolution and see if that affects the frame rates and how much by. hope this helps.
June 27, 2012 10:56:58 AM

ok robustus so i launched BF3, with everything set on LOW, on an empty server, and the framerate was hopping all over the place, when i shoot something, it just drops to like 2 frames a second. Here are the temps while playing:


HDD: 31
CPU: 51
MB: 28
GPU: 73
Core:44
a b U Graphics card
June 27, 2012 10:58:27 AM

Here is my basic checklist, run furmark for atleast 5-10 minutes and watch temps, if they get around 90c, shut down furmark as thats most likely whats causing your instability.

You can also run Prime95 and watch temps with another program like HWmonitor, again run prime95 for atleast 5-10 minutes.

While you are running furmark you might want to open up GPUz (download it first) and watch the temps of your VRMs (they might be called VRREGs or voltage regulators or something like that. (you can also set it up to log the temps into a text file incase your pc crashes you still know what your temps were at)

One more thing, try turning your gpu fan to max (you can do this in catalyst control center (your video card drivers)) and see if fps quit jumping around

Oh AND I had a problem with Origin scanning or doing some weird thing and it took up 90% of my cpu while playing I would check task manager and make sure nothing is going on there.
June 27, 2012 10:59:43 AM

DG8ball said:
I have just bought a new gtx 680 and that temp at idle is around the 35 to 40 mark however under load it has hit temps of 69 so i would probably relate it to your cpu temp if that gets to hot i have experienced lag and allsorts of crashing due to a cpu being to hot, this can affect the in game play making it feel very laggy which would explain the drop in frame rates. i used to run a 5750 before my gtx 680 and i would experience lag in some intensive parts of the game, try lowing the resolution and see if that affects the frame rates and how much by. hope this helps.




DG, im sure there is a cause for my huge framerate loss, and thats not the fact that my graphics card cant run it. on LOW settings, im having these problems, and that's not even on a actual server, iv been trying on empty servers. its a 5770, and if i set graphics to default, its sets itself to high, so why i would be having these huge lag/framerate issues at LOw settings is beyond me atm...
June 27, 2012 11:01:10 AM

mouse24 said:
Here is my basic checklist, run furmark for atleast 5-10 minutes and watch temps, if they get around 90c, shut down furmark as thats most likely whats causing your instability.

You can also run Prime95 and watch temps with another program like HWmonitor, again run prime95 for atleast 5-10 minutes.

While you are running furmark you might want to open up GPUz (download it first) and watch the temps of your VRMs (they might be called VRREGs or voltage regulators or something like that. (you can also set it up to log the temps into a text file incase your pc crashes you still know what your temps were at)


im very new to all this hardware testing, but i shall try my best. So i should run all three of those programs seperately and record what i see? Right now im getting my temp values from speedfan
a b U Graphics card
June 27, 2012 11:03:26 AM

mahfuz27 said:
im very new to all this hardware testing, but i shall try my best. So i should run all three of those programs seperately and record what i see? Right now im getting my temp values from speedfan


Yup run furmark and prime95 separately, speed fan does a fine job but I found sometimes it states my gpu temp is at 999c :/ 

Regardless furmark has its own gpu monitoring so its not an issue :)  I also editted the above message before you could reply.
June 27, 2012 11:06:50 AM

ok, so before i go and stress test with furmark, i want to add: i just checked task manager, and im not sure what to look out for that may be eating my CPU, but i should note, there are 12 svchost.exe prcesses running. In total, they are taking up around 240,000 k memory, is this....normal? and is there a way i can post my processes so you may be able to tell me if there is anything useless running?
June 27, 2012 11:10:37 AM

also mouse, i got furmark installed now, but there are quite a few options, what settings should i run it on for 5-10 mins?
a b U Graphics card
June 27, 2012 11:15:36 AM

You can sort things by cpu usage, just hit the cpu thing at the top of your processes tab :) 

Those 12 svchost are interesting, I don't think they are a problem but we might need to come back to those if the tests don't prove useful.

As for options make sure its at max native screen resolution (match it to whatever your desktop is set to) and max Anti Aliasing then I usually tick dynamic camera but I dunno how much it helps stress it. then just hit burn in and your off to the races, and yes its supposed to be laggy :) 

Also watch it to see if you see any artifacting or general oddities with the picture.
June 27, 2012 11:31:00 AM

ok done it and yes, it was definitely laggy xD

But the GPU didnt go over 75 degrees, the temp kinda levelled off there, so does this mean its not an issue with the GPU?

If that is what it means, then i think it may be the CPU, as the whole computer itslef tends to slip into "not responding" fairly frequently, and i assume we blame the CPU for that?
a b U Graphics card
June 27, 2012 11:35:01 AM

mahfuz27 said:
ok done it and yes, it was definitely laggy xD

But the GPU didnt go over 75 degrees, the temp kinda levelled off there, so does this mean its not an issue with the GPU?

If that is what it means, then i think it may be the CPU, as the whole computer itslef tends to slip into "not responding" fairly frequently, and i assume we blame the CPU for that?


We certainly ruled out the GPU over heating. Now all we need to do is grab Prime95 and see if your cpu is overheating.

The not responding thing and SVChost you mentioned has me wondering if its a virus or malware but we can save those for later as those scans usually take a while
June 27, 2012 11:39:48 AM

ok prime95 then, but...i cant seem to find it on the net.

When googling it, i keep coming up with this program that sorts prime numbers xD

Would you be so kind to give me a download link? :) 
a b U Graphics card
June 27, 2012 11:42:25 AM

Yeah thats basically what it does, its rather good at stressing a cpu.

http://www.mersenne.org/freesoft/

just choose windows 64 or 32bit depending on what your running

You don't need to create an account or anything extra like that. (skip to step 2 in the 2nd set of instructions)
a c 84 U Graphics card
June 27, 2012 11:48:36 AM

in Prime95 remember to check the 'round off checking' option under the 'advanced' menu, so it actually tests for system stability as well. You might catch some instability problems with it, even if the cpu doesnt overheat while the test is running.

For stability testing run it for several hours (without errors), for temperature testing a much shorter test run is enough (like 10-20 minutes)
June 27, 2012 11:56:38 AM

Well the temperatures don't seem too bad really. I would be tempted to take the card out and give it a good blast of air though, amazing how much dust accumulates, especially around heatsink fins. A few blasts from an aerosol duster is good but failing that a few good blasts from your own lungs should clear most of it!

A bit concerned what the huge delay is to get to a responsive system. If you can't get straight to the task manager because of it to see if anything is pegging out the cpu it might prove useful to run msconfig. At the end there is a tab with tools and there is a system log there which might shed some light on the matter.

Does it do this when booting into safe mode? (hit f8 before windows gets to start up) might hint at a driver problem.

Wouldn't do any harm to uninstall your graphics card drivers and reinstall them I guess.

What cpu are you running? enough memory and hard drive space?
June 27, 2012 12:23:34 PM

ok so iv done the prime95 testing, and the cpu does not seem to go over 62 degrees. As for the matter of my whole system being unresponsive frequently and slow, i have no idea why. I definitely have enough RAM, 8GB, and my CPU is a phenom x6 so i really have no idea why everything is so slow. As for my graphics card drivers, can you give me a step by step on how to uninstall and re install them? Im very new to such tasks. :) 
June 27, 2012 12:26:22 PM

And i havnt seen if its unresponsive in safe mode, will give it a try in the next few hours
!