Need help with TDR issues please

[Taneras]

Firstly, my rig:

GPU: Gigabtye GTX 970
CPU: i5-4690
Mobo: Gigagyte G1 Sniper H6 or Z6 (I forget, whichever goes with the CPU).
Memory: 4X2 GB
PSU: Corsair HX750
OS: Windows 7

I've been having lots of TDR issues, probably around 20 within the last week. I can run my PC anywheres between 15 minutes to 6-8 hours before getting one. It doesn't matter if I'm web browsing, playing lightly intensive games or heavily intensive games, I still get these errors.

I've checked my temps, drivers, and uninstalled programs that I had installed right around the time I started having these issues. One oddity, and it only happened once. One of the crashes didn't result in a reboot, the comp went completely off. I pressed the power button and it wouldn't boot. I had to flip the PSU's switch off and on again to get it to boot. That's only happened once. I haven't had any issues with my PSU outside of that.

I have lots of minidump files from the TDR's. I downloaded a program called "Whocrashed" and their analysis of the problem is that its a driver issue for the GTX 970. Oddly enough when I started playing with older drivers, and analysed my TDR's again, all of the errors changed from the driver I had to the driver I just installed. I'm not sure the proper way to uninstall/reinstall drivers, I just went to the control panel, uninstalled the latest driver, rebooted, went on the web and downloaded/installed the driver I wanted.

I'm sure yall will want to look at the minidump files, what would be the best program to open up a .dmp file and display it here?

Thanks again!

 

[Shaun o]

The errors you are talking about will relate to one or two things in the system.
The first suspect of course would be memory, or the memory mdodules used.

You should check in your bios all of the auto detected settings the bios made for the memory by auto detection are correct.

Take all of the memory out of your system first, each stick.
TDL is mainly down to a timing, or memory error.

Read off all of the information on the memory stick (found on a sticker of the memory module), for the voltage, clock speed and timing values at the rated operation speed of each memory stick.

If you have any stick that does not match in voltage, speed or timing values depending on how many ram modules or sticks you have do not use it when running other memory at different timings or clock speed to other sticks of memory in the boards memory slots at the same time.
The last value of the memory in wait state, should in most cases be set to T2 mode.

TDL can also relate to a call request for data from a driver of the system to the cpu and a memory location of the system ram. It can be down to a time out for a requested operation not sent or received from or to the cpu from memory or a location or block of system memory producing the error in question. This can be down to a fact that another bit of hardware in the system is using the same motherboard resource channel, that we call an IRQ number the cpu uses to talk to system memory and the hardware device via the driver installed for it. We will deal with that later but first check the state of all of the memory used in the system.

After you have the settings of all the memory sticks, insert only one into the memory slot of the motherboard.
Enter the bios of the motherboard and verify that the rated speed, along with the voltage, and timing values of the memory have been correctly auto detected.

Make sure if the rated speed of the memory if quoted as Xmp memory mode that the option of XMP is enabled in your bios.
By setting the Xmp mode option in the bios it allows the voltage to be automatically changed from a Stock 1.5v to a 1.65v setting because the memory is potentially overclocked to a higher Mhz speed and requires slightly more power to keep stable when load is put on it.
If Xmp mode is required to be enabled, and is not in the bios settings, it will produce a memory error in most cases, when running in windows and produce a memory error dump file.

Another thing to look for is the memory type mode option set in the bios.
You can get two types of memory buffered, normal pc standard of memory.
Or you can get EEC rated memory.
If EEC memory is used then an option in your bios must be set to EEC mode on in the bios.
If it is turned off, in the bios, and the memory does have the EEC option it will result in errors in windows most of the time.

Test one stick only in the first slot of the board on it`s own and do a Microsoft memory test of about three memory scans to verify the memory on the module is fully working and not the cause of the errors after checking in the bios all settings are correct to what is stated on the sticker of the memory module.

If the stick passes all tests.

Remove it and place another in the first memory slot of the motherboard on it`s own and verify the memory module is good by the Microsoft memory error checking program.
Do this untill each of the memory sticks you have are tested.

If you follow these tips, then post back on the results, you will find the cause of the errors in most cases. But if the errors keep happening after memory testing, the next thing to look at is the cpu, and other hardware that may cause the system to be unstable.

NB: It is also common that if the main PSU is faulty it can cause power problems and memory errors at random times. and can be a sign the main PSU should be replaced as a last resort after testing and error checking of the system.

Please do the checks, and report back on the forum, if the errors still happen.

 

[Taneras]

The errors you are talking about will relate to one or two things in the system.
The first suspect of course would be memory, or the memory mdodules used.

You should check in your bios all of the auto detected settings the bios made for the memory by auto detection are correct.

Take all of the memory out of your system first, each stick.
TDL is mainly down to a timing, or memory error.

Read off all of the information on the memory stick (found on a sticker of the memory module), for the voltage, clock speed and timing values at the rated operation speed of each memory stick.

If you have any stick that does not match in voltage, speed or timing values depending on how many ram modules or sticks you have do not use it when running other memory at different timings or clock speed to other sticks of memory in the boards memory slots at the same time.
The last value of the memory in wait state, should in most cases be set to T2 mode.

TDL can also relate to a call request for data from a driver of the system to the cpu and a memory location of the system ram. It can be down to a time out for a requested operation not sent or received from or to the cpu from memory or a location or block of system memory producing the error in question. This can be down to a fact that another bit of hardware in the system is using the same motherboard resource channel, that we call an IRQ number the cpu uses to talk to system memory and the hardware device via the driver installed for it. We will deal with that later but first check the state of all of the memory used in the system.

After you have the settings of all the memory sticks, insert only one into the memory slot of the motherboard.
Enter the bios of the motherboard and verify that the rated speed, along with the voltage, and timing values of the memory have been correctly auto detected.

Make sure if the rated speed of the memory if quoted as Xmp memory mode that the option of XMP is enabled in your bios.
By setting the Xmp mode option in the bios it allows the voltage to be automatically changed from a Stock 1.5v to a 1.65v setting because the memory is potentially overclocked to a higher Mhz speed and requires slightly more power to keep stable when load is put on it.
If Xmp mode is required to be enabled, and is not in the bios settings, it will produce a memory error in most cases, when running in windows and produce a memory error dump file.

Another thing to look for is the memory type mode option set in the bios.
You can get two types of memory buffered, normal pc standard of memory.
Or you can get EEC rated memory.
If EEC memory is used then an option in your bios must be set to EEC mode on in the bios.
If it is turned off, in the bios, and the memory does have the EEC option it will result in errors in windows most of the time.

Test one stick only in the first slot of the board on it`s own and do a Microsoft memory test of about three memory scans to verify the memory on the module is fully working and not the cause of the errors after checking in the bios all settings are correct to what is stated on the sticker of the memory module.

If the stick passes all tests.

Remove it and place another in the first memory slot of the motherboard on it`s own and verify the memory module is good by the Microsoft memory error checking program.
Do this untill each of the memory sticks you have are tested.

If you follow these tips, then post back on the results, you will find the cause of the errors in most cases. But if the errors keep happening after memory testing, the next thing to look at is the cpu, and other hardware that may cause the system to be unstable.

NB: It is also common that if the main PSU is faulty it can cause power problems and memory errors at random times. and can be a sign the main PSU should be replaced as a last resort after testing and error checking of the system.

Please do the checks, and report back on the forum, if the errors still happen.

Thanks for the indepth and quick response. I will certainly give this a shot when I'm off of work this afternoon and will post the results. Just to make sure though, because I saw you mention TDL, we're talking about TDR issues right? WhoCrashed blamed it on "nvlddmkm.sys" which is the GPU driver if I'm not mistaken.

Thanks again!