video card stopped responding but recovered

[dazco]

Thats the error message i have gotten ever since this PC was new. It's a ATI or i guess now AMD HD5850. The screen will freeze for a second then go black. 3 seconds later it's back and i get that error message. I have tried various drivers and thats just not it. Case is well vented and it's not even hot, nor does it tend to happen when it IS hot when gaming for example. Can't imagine the contacts on the card in the PCI slot are contaminated at this point only about 3 months after i built it. It only happens maybe once a week. Any ideas what else could cause this?

 

[TVeblen]

You are experiencing TDR events. The following is my diagnostic procedure for tracking down the cause. I hope it helps.

"Display driver xxxxx stopped responding and was recovered"

Timeout Detection & Recovery (TDR) = "Display Driver Stopped Responding and was Recovered" is a useful feature that started in Vista and is also in W7 that allows the OS to try and recover from a video timeout so that the system does not crash to a bluescreen. Symptoms included a screen flash with the TDR message appearing one or more times or the screen blinking out to black. If the system cannot recover it will crash (Stop Error 116 typical). The issue is that the video card is not responding as expected. The solution is in the: why?

There is no one-size-fits-all solution to TDR errors. But the problem is usually found in the local environment (your computer). Finding the cause is a matter of checking every possible cause and uncovering the culprit through a simple process of elimination. By methodically running down a checklist of diagnostic procedures you should be able to find the cause and can correct it.

There are numerous reports of hardware solutions to TDR's. The most common are:

• Poor Cooling
• Problems with the power supply
• Overclocking Issues
• Bad System memory or incorrect memory timings
• Defective PC Components

The order you do the diagnostics is not all that important. My personal strategy is to do the cheap & easy stuff first, the cheap & harder stuff next, and then the stuff that costs last. But whatever order you do it in you need to check or confirm the following:

SOFTWARE
Poorly written software and games will cause TDRs. But if this were the case it would affect lots of people, not just a few. Check the game's website & forums for patches and tips.
See if other people in the forums are having the same problem and if they were able to solve it and how.
You could also be asking too much of your video card. Check to see if your video card is tested and recommended for the game/program. Test the game at reduced settings.

WHAT ACTIONS CAUSE THE PROBLEM
It helps if you can isolate the actions that trigger the TDR. Most often it will be an application using 3D graphics. But if the incidents occur constantly it would point more towards defective hardware. If it happens more specifically (just when running Game X) it points towards overheating, settings, software, or driver issues.

GENERAL SYSTEM PERFORMANCE
You need to eliminate the possibility that your computer has a global problem. You can use a program like Prime95 to stress test your system. Free Software - GIMPS
You can run the "Stress Test" for a few hours or overnight. This will not tell you what the problem is, but it is helpful to uncover any issues your system has with instability and cooling.

OVERHEATING
Running a video intensive game for hours can generate some serious heat and overheating will cause video errors. You can check your temps by looking at your BIOS readings or use a free program like Speedfan SpeedFan - Access temperature sensor in your computer .
A real easy test is to just pull the side panel(s) off your case (You can also blow a house fan directly into the open case) and see if the problem goes away or gets better. If it does then the issue is definitely overheating. If you are overheating you need to look at installing some cooling upgrades. You want to look at ventilating the case (more or bigger fans), Upgrade your case to a larger gaming case (lots of fans, water-cooling), etc.
There are free utilities like BurninTest PassMark BurnInTest software - PC Reliability and Load Testing that you can use to test your system's cooling capability. Caution is recommended using these types of programs.

VIDEO DRIVERS
Bad drivers happen and they can get corrupted. Before installing or reinstalling any video drivers first completely uninstall all video software and the drivers. (Some people say to run a cleaner program from safe mode, some say this is unnecessary). Never rely on the driver package to overwrite the old drivers. Also: Delete the video driver folder (ex: C:\NVIDIA) in Windows Explorer (or windows may install the same drivers again!).
After uninstalling the old drivers and rebooting Windows 7 will install it's own WDDM 1.1 driver. Check for the video problem while using this generic Windows driver.
You can then install the latest drivers for your card (or try older drivers).
See This Tutorial: http://www.sevenforums.com/tutorials/43216-installing-updating-drivers-7-a.html#post414637

DEVICE MANAGER
Look in Device Manager and make sure there are no problem devices (yellow ! icon). Correct these by loading the correct drivers or disable the problem device and see if the video problem goes away.

POOR CONNECTIONS
Reseat video card and memory modules. Make sure the contacts are clean. Check all the electrical connections.

CHECK YOUR MOTHERBOARD VOLTAGES
In BIOS, check the listed voltages against the manufacturer recommended specs. Reset the voltages to factory defaults and see if the video problems disappear.

MEMORY
Memory errors can cause video problems. Run a program like Memtest86+ for at least 3 passes to see if there are any memory errors. Memtest86+ - Advanced Memory Diagnostic Tool .
You can also test for a bad memory module by installing one stick in Slot 1 and testing, and then switch it out for the next stick, etc.
When populating all of the RAM slots on a motherboard it is sometimes necessary to go into the BIOS and increase the voltage to the RAM slightly to obtain a stable system.

OVERCLOCKING
Overclocking can be a trial and error process. The clocks and/or multipliers you set or change for CPU, Memory, or GPU could be unstable. Eliminate this as a possibility by resetting these to their BIOS defaults to see if that clears the video problems. The simplest way to do this is to "Restore Bios Defaults", or “Clear CMOS”.

UNDERCLOCKING
Some people have reported that by going into the video cards control panel and "down-clocking" the cards performance settings they were able to clear up the TDRs. Since W7 does not seem to tolerate any hiccups in the GPU, this would allow you to run a poor performing card in the W7 environment.
So for instance, you could set the GPU clock from a 777 MHz factory setting to 750MHz, and the ram clock from a 1126MHz factory setting to 1050Mhz, or similar small change for your particular card.

BIOS
Check you motherboard manufacturer’s website for an updated BIOS. An updated BIOS may correct an unstable condition, particularly if it says the newer BIOS corrects memory errors or has bug fixes. You could also try loading the BIOS defaults.
Caution is recommended when updating (flashing) a BIOS. The safest way to do so is from the update utility within the BIOS. Follow instructions carefully.
While you are there, check the motherboard manufacturers forums to see if others are having issues with the same board.

WINDOWS POWER MANAGEMENT
Eliminate Power Management settings as a possible cause, especially if you are working with a laptop. These settings could be particularly important if the issue is in playing games.
Go to Control Panel > Hardware & Sound > Power Options. Under "Select a Power Plan" you will find that "Balanced" is the default setting.
At the bottom you will see a Down arrow next to "Show Additional Plans". Click that and select "High Performance". See if the TDR issue is affected.
Alternately, you can click "Change Plan Settings" next to the "Balanced" plan and change the setting to "Never" put the computer to sleep (This is the default on a desktop) and/or change when the display is turned off as a test.

POWER SUPPLY
You need to know that your power supply is delivering sufficient power. Power supply problems are the most common cause of video problems, especially using high end cards.
Check the power supply's amperage ratings. Be sure it has the ample amperage for your video card and the rest of the system.
You can test the supply with multimeter to measure for a steady 12v to the card's power connectors. (The only true way to test a power supply would be to use the very expensive diagnostic equipment used in labs). But for us regular folks: I tested my power supply by hooking up my multimeter to the PCI-E connectors that I was using to power my video card (I used a spare pair from the power supply to run the card while I was testing). I then observed the meter while I used the computer, first watching the voltage, then the amps, to see if there was any drop-off or erratic behavior while booting or using the computer. My readings were rock solid. So I declared my power supply good.
Otherwise you need to replace the supply to eliminate this possibility. Or borrow one from another computer.

VIDEO CARD
I suspect that a video card must perform flawlessly to operate in a Windows 7 environment and run the most recent games. If you tried all the above diagnostics and no problems were found then that leaves you with only one possibility: a defective video card. Some brands and models have the problem more consistently than others. You could check their forums for clues.
You could try your card in another computer running W7 to see if the problem goes along with the card.
You could try a different card in your computer. I bought an inexpensive card to use. My TDR's disappeared using a "lesser" card. Or borrow a card from another computer.
Otherwise RMA or replace the card.

**********
TDR complaints have come from PC owners running virtually every PC configuration. They occur regardless of which video engine, manufacturer, driver, or system used. They are too numerous to write off as a random problem, but at the same time if people are getting their systems to run correctly using the same hardware and software that you are then it follows that your problem must be solvable.

More Info Here:
Timeout Detection and Recovery of GPUs through WDDM
NVIDIA Statement on TDR Errors - NVIDIA Forums
27116: ATIKMDAG has stopped responding error message

 

[dazco]

Thanks. theres a lot there, plus it only happens every so often so i won't know whether i have found it anytime soon. But i'll go thru that and see what i can find. Thanks.

 

[TVeblen]

That's the best you can do with this issue really - just plug away at it!

The majority of times we see this problem here, if it can't be corrected by a clean install of the driver, it ends up being RAM, Power Supply, or the Card itself is defective.

But if you are like me you will want to eliminate all other possibilities before shelling out the bucks!

Good luck!