Vertical Lines of Death

[dcarpenter85]

Greetings, this is my first post on these forums. I decided to come here with my problem because this community seems to be one of the most knowledgeable when it comes to GPUs so here is my problem.

I recently built a system (parts listed below) and while it has been running great for the most part, when I am playing some (not all) games I get a vertical stripe lock up and I am forced to hard reboot the system. I have an XFX RadeonHD 5770 XXX edition (875/1300 clocks). I have read that this may be caused by the stock clocks being too high, so I clocked them back to the reference 850/1200 and was still getting the vertical line crash. The weird thing about it is is that I can play some games for hours on end with no crashes at all. World of Warcraft, Dragon Age: Origins, Mass Effect 2, Burnout Paradise and and Team Fortress 2 have never produced the crash and those are the games I have spent the most time playing since building the system. Games that seem to be more prone to causing it are Left 4 Dead, Crysis Warhead, Lord of the Rings Online, and Modern Warfare 2. There are a few other games that the crash has occurred in once or twice but has never happened again.

Something else that is odd, if I am using Skype to voice chat with someone while playing a game and I get the vertical stripe lock up, I can continue to talk to them even though I am locked out of my PC. Meaning I have vertical stripes on my monitor yet I can still hear the game sound and I can still communicate both ways via Skype.

I am not overclocking my system at all, everything is stock. Memtest86+ after multiple passes shows no errors, and prime95 after running for 6 hours showed no problems as well. My card runs Furmark flawlessly, no crashes. I am really at a loss as to what is causing this problem. If anyone has a clue as to what is wrong or suggestions on how to fix it I would greatly appreciate it.

System Specs:
Gigabyte P55A-UD3 w/ latest bios + drivers
Core i5 750
XFX RadeonHD 5770 XXX Edition w/ Catalyst 10.7
4gb G.Skill Ripjaw DDR3 1600 @ 1333 9-9-9-24-1T
650 Watt Antec EarthWatts PSU

 

[jedimasterben]

I'm honestly at a loss, as this was supposedly a driver issue that has long been fixed. I had the same issues, but they disappeared after Catalyst 10.3 came out.

Where are you setting your clocks at? Try MSI Afterburner (it doesn't matter that your card is from XFX), as it by default doesn't allow the clocks to go back down when the card is idle. Set the clocks manually to the reference clocks and then see if that helps you out any.

 

[dcarpenter85]

I'm honestly at a loss, as this was supposedly a driver issue that has long been fixed. I had the same issues, but they disappeared after Catalyst 10.3 came out.

Where are you setting your clocks at? Try MSI Afterburner (it doesn't matter that your card is from XFX), as it by default doesn't allow the clocks to go back down when the card is idle. Set the clocks manually to the reference clocks and then see if that helps you out any.

Okay I will try this and do some testing, thanks.

 

[dcarpenter85]

Used MSI Afterburner like you suggested to lower the clocks and I still got the vertical line lock up about 30 minutes into playing Left 4 Dead

 

[dcarpenter85]

https://docs.google.com/leaf?id=0B2Li7XODSScWYzBkY2IyNWQtODJmNS00MmY0LWI3ZmYtZWY3YmQ5MDZjZmY3&sort=name&layout=list&num=50

I uploaded my GPU-Z sensor log to google docs so you guys can take a look at it, nothing out of the ordinary seems to be happening when the card locks up. The last entry in the log is when the card locked.

 

[bystander]

From my experience, this is related to the vram overheating, which has no sensors to be monitored. Even with the new drivers, I've seen this happen, so I went back to my original fix:

If you downclock the memory to 50 below stock, I've never had an issue.

 

[dcarpenter85]

From my experience, this is related to the vram overheating, which has no sensors to be monitored. Even with the new drivers, I've seen this happen, so I went back to my original fix:

If you downclock the memory to 50 below stock, I've never had an issue.

I will set it to 1150 and test. If this is the problem I will probably just use my warranty on the card and get it replaced. I shouldn't have to downclock 150mhz to make the card stable

Also, how would this explain some games causing the crash and others not? Could it just be that some games are using the memory more rigorously than others, causing it to overheat?

 

[bystander]

I will set it to 1150 and test. If this is the problem I will probably just use my warranty on the card and get it replaced. I shouldn't have to downclock 150mhz to make the card stable

Also, how would this explain some games causing the crash and others not? Could it just be that some games are using the memory more rigorously than others, causing it to overheat?

Probably.

Just so you know, the memory clock has far less impact on performance than the GPU clock. If you overclock the GPU by 50, and underclock the memory by 50, I find it more stable, and I gain performance.

 

[dcarpenter85]

Okay thanks for the info, I will test out the lower clock and post my results

 

[dcarpenter85]

I tried setting the memory to 1150 like you suggested and Lord of the Rings Online crashed after ~20 min with the vertical stripes

I also tried upping the voltages via MSI Afterburner but still vertical striped at 1.2 volts.

I guess I will just RMA with XFX and cross my fingers. I can't imagine it is anything besides a faulty card. I don't have another GPU to test in my system though so I can't totally rule it out.