Vid Card Help (on behalf of Nomad)

**Ferris Bueller** · 03-09-2011, 06:44 PM

So Nomad, whom many of you know is my brother, fried his video card last week (8800GTX). I replaced it with a HD5750 because its all he could afford. I understand that its not a great card and yadda yadda yadda. Can we move past that please?

Now, he's into playing world of tanks and BC2. The old 8800 worked just fine with both. The 5750, which is a much beefier card, is having problems. About every 15-20 seconds in-game, the game will freeze for 10-15 seconds, followed by a 5 second black screen, and then it will fix itself...for 15-20 seconds. Alt+tabbing back to desktop presents me with a message that "the display drivers have stopped responding but have recovered".

I scoured google and found some articles on how to fix it. Tried the following to no avail:
-Scrubbing and reinstalling drivers.
-Reinstalling games.
-Reinstalling latest .net, C++ 2010.
-Manually setting minimum GPU clock settings.

Nothing has worked. I'm kinda pulling my hair out.
System specs:
E6300 1.86GHz core2duo
4GB RAM
500mb 7200RPM HDD
XFX ATI HD5750 1GB GDDR5
Win7 x86 (32Bit) + SP1

PLEASE limit responses to trying to help. He cant afford to upgrade hardware. He cant afford a new rig altogether. I just want to fix the problem that this card is having. Thanks guys.

**Wicks** · 03-09-2011, 06:51 PM

PSU, PSU connection to card, overheating, those are the things that immediately spring to mind and run it at default. Also try switching resolutions.

**LpBronco** · 03-09-2011, 06:52 PM

Have you downloaded the latest Catalyst Control Center? This has a few features that might make it easier to make sure that your actually running the correct drivers and check your dxdiag and make sure the ATI drivers are listed.

**MacLeod** · 03-09-2011, 06:52 PM

That would happen to Nomad, he probably deserves it

I tried googling the issue and turned up with nothing much more then what you have already done. Dice seems to have inherent issues with ATI cards (like my old one) and it may just be an issue with the card itself. Is it possible to get a replacement card to see if that would help?

**Namebot** · 03-09-2011, 06:54 PM

Symptoms exist in multiple games?

Maybe the reason his 8800gtx(maybe this wasn't even broken) fried is because his power supply took a a dump and is now producing much below its rated power? Might try swapping a a different capable unit in and seeing if that fixes the problem.

**Acreo Aeneas** · 03-09-2011, 07:28 PM

As LpBronco already pointed out (#1):

1. Update your card's drivers to the latest 11.2. Get it from http://ati.amd.com/

2. If he has eVGA's Precision or MSI Afterburner utilities installed, he needs to uninstall them and download and the install the latest versions (or just leave them uninstalled). These programs are known to cause various games to crash or suffer weird problems since most of the older versions are incompatible with SP1 for Windows 7.

3. Check the power connections to the video card. Perhaps the 6-pin PCI-E power adapter is loose. It never hurts to make sure.

**Ferris Bueller** · 03-09-2011, 07:40 PM

@Bronco: Latest drivers were installed today when I started trying to troubleshoot. Same with CCC. Although I have to say that CCC2 is crap for making profiles.

@MacLeod: Negative on the replacement. Already called to check.

@Namebot: Symptoms exist in BC2 and WOT, which are the only games he has installed currently. Have done fresh reinstalls of both, just to check options. Checked the power supply. 12v rail showing good, power levels check out. The old card died because of chronic overheating for a year+. Nomad got help from a stupid person and OC'd the card to run BC2 on ultra high settings and just cooked it.

@Acreo: checked connections about 20 minutes ago. Same symptoms. Drivers are up to date as of today.

On the up side, I think its an unstable clock on the GPU. Scaled back the GPU clock and mem clock from 750/1150 to 650/1050 and now I'm only getting the problem once every 20-30 minutes. Sad though, considering it's rated to run WAY above those marks.

**LpBronco** · 03-09-2011, 08:01 PM

I've heard of issues moving from Nvidia to ATI in the past (and the biggest reason I haven't ordered a 6970 as we speak). The Nvidia files are scattered all over the place and I don't remember but after a driver sweep I think what you have to do is let windows install a generic driver and then install the ATI drivers. You can also try to install an older ATI driver first and then upgrade. Make sure Precision or any other Nvidia utility is also removed

**Acreo Aeneas** · 03-09-2011, 08:36 PM

Originally Posted by Ferris Bueller

@Acreo: checked connections about 20 minutes ago. Same symptoms. Drivers are up to date as of today.

On the up side, I think its an unstable clock on the GPU. Scaled back the GPU clock and mem clock from 750/1150 to 650/1050 and now I'm only getting the problem once every 20-30 minutes. Sad though, considering it's rated to run WAY above those marks.

Hold on here. You didn't mention you have the thing OC'd. Whatever you do, do not mess with the VRAM clocks! #1 cause of problems is when people decide to OC the VRAM on their HD58xx cards.

Leave them on defaults. I've already gone through a slew of troubleshooting when I decided to OC my own HD5830's VRAM. There's also a slew of similar threads over at OCN about this exact same issue. In the end, everyone found out that if you OC the VRAM, you're going to get nasty weird problems that don't seem to have a solution.

**Ferris Bueller** · 03-09-2011, 08:53 PM

Thats the thing, it wasnt overclocked to begin with. Those were the default settings in the profile (which I had to go in and manually lower). Hell, overdrive in CCC is still locked cause I havent touched it. So it's now running 100mhz below default on both the mem and gpu clocks.

**Acreo Aeneas** · 03-09-2011, 09:07 PM

Originally Posted by Ferris Bueller

Thats the thing, it wasnt overclocked to begin with. Those were the default settings in the profile (which I had to go in and manually lower). Hell, overdrive in CCC is still locked cause I havent touched it. So it's now running 100mhz below default on both the mem and gpu clocks.

Actually, I have a better question:

How did your bro's old 8800GT fry? Was it a power surge that went through his PSU?

If it was a power surge, you need to replace the PSU as well, not just the old graphics card.

If it wasn't a power surge, and you highly suspect it to be the PSU, replace it immediately.

**Ferris Bueller** · 03-09-2011, 10:30 PM

Nah, wasnt a surge. He got some amateur help OC'ing the 8800 and fried it from ridiculously high temps over a long period of time. So I'm pretty sure the PSU is out of the equation.

**Acreo Aeneas** · 03-09-2011, 11:27 PM

Hmm, had to look around the net as I was having this nagging feeling I've seen this before. Apparently I have since this particular issue seems to be randomly occurring to both nVidia and ATI/AMD owners of various models.

The issue revolves around a stability issue with power saving modes inherent in the built-in firmware of each card. Basically, when you're not doing anything intensive, the card clocks down to save power and when it clocks down, sometimes it doesn't go back to 3D clock speeds properly and that's what causes the problem. For us AMD/ATI owners, it's known as "Powerplay".

The only known workaround I could find involves using AMD's Overdrive in CCC (Catalyst Control Center), basically overclocking utility built into the drivers. This should effectively disable Powerplay on the card. This may not your problem. I'm assuming you're using the latest drivers.

Here it is:

1. Right-click anywhere on your desktop.
2. Click on "Catalyst Control Center".
3. Click on "Performance" on the left-hand side.
4. Click on "AMD Overdrive".
5. Click on the red and gold lock icon to enable Overdrive (there'll be a warning message, click OK).

What you see now are slider bars that let you customize the effective clock speed of your GPU core and VRAM. The sliders at where they are now should be the default values for your card. (*** In your case since you already downclocked the VRAM/GPU, I'd assume you know what the original effective clocks are for your card. ***) I'd write them down in case you forget them in the case that you need to revert them back.

Assuming you don't know what the original effective clocks were/are, I'd go look it up on the box/manual that came with that card. Or visit XFX's website and see if it's listed.

Try inching both sliders up until you add just 5 mhz to your GPU clock and 10 to your VRAM clock. This is to ensure that Powerplay is "turned off". We're going to ignore my previous post here about overclocking the VRAM as this small of a OC shouldn't cause any problems.

6. Click "Apply".

Now go ahead and launch up BC2 or whatever it is he plays and give it a few run throughs to see if the problem is solved or not.

Note: Ideally, if we left them at original clocks, it should automatically disable Powerplay since Overdrive is on.

Anyways, come back and tell us how it goes.

I can't say I've had this problem ever and glad it never came up for me.

**Mindkill** · 03-09-2011, 11:44 PM

I had a similar problem with my nvidia GTX 275 card in Black Ops and BC2, except I never saw the error message about the driver. I was able to stop it from happening by lowering the anti-aliasing to 4x.

After I installed the newest nvidia drivers, I was able to set the AA back to max with no issues.

**LpBronco** · 03-10-2011, 06:05 PM

CCC on my 3850 AGP ,running on my P4 rig, has an overclocking utility that basically looks for the best settings for your particular system. I got great results and a very stable overclock with about a 13% gain in clock speed.

I would still look to purge any Nvidia drivers and files that are left on the machine to make sure there isn't any conflict remaining with the new drivers. Let us know what works and I'll have to think again about that 6970.