Graphics Card TDR problem not resolved by RMA replacement - Advice?

[drdrewnatic]

Hi all, first time posting on the forum!

 

SPECS:

CPU: Intel Core i7-9700k - No overclock but some Asus default optimizations ie. Multicore Enhancement are on auto

GPU: Asus ROG Strix Geforce RTX 2070 OC - No overclock apart from factory - Studio Driver v511.65

MOTHERBOARD: Asus ROG Maximus XI Hero (Wifi) - BIOS Version 2004

RAM: 2x8GB G.Skill Trident Z RGB 3200 CL16 - XMP

PSU: EVGA G3 650w Power Supply

STORAGE: Boot: 500GB Samsung 860 Evo SSD, Storage: Seagate 2TB 7200RPM Hard Drive

COOLING: Cooler Master Hyper 212 RGB Black Edition

CASE: NZXT H500

OS: Windows 10 Build 19044.1566

DISPLAY: (2) BenQ BL2420PT Monitors - Connected via HDMI

 

For about a year or two now, I've had many issues regarding my GPU and GPU driver. Eventually after getting fed up with all these problems, I deiced to contact Asus product support as my card is still under warranty. They said that I had to RMA it to fix the problem. After sending it to them, they could not find any issues with the card apart from a wobbling fan. I got back a replacement graphics card and when I put it back in my system, it still has the same problem as before. This leads me to believe that the issue is not coming from the GPU. Am I right?

 

ISSUES:

1. Occasionally, while no load is being put on the GPU, both my displays would freeze and eventually go black. After about 30 seconds to 5 minutes (or a force restart depending on how patient I am), the system would eventually blue screen (while still displaying nothing) and restart without any issue.  The following dump file would be created: (see Google Drive Link). The occurrence of this would be sporadic. The system could go months without any issue while other times it could be an interval of a week.

 

2. More recently (around a year ago), when I turn on my system, the post screen sometimes would not display on any of my monitors and the no VGA debug light on my motherboard turned on. Once the system fully booted into Windows, the lock screen would display however there would be no keyboard or mouse input. After a couple of seconds the display would go dark for a little bit and then display the sign-in screen. Around five minutes after logging on, the system would become very 'laggy' and eventually unresponsive causing me to force shutdown the system.

 

3. If the system would boot correctly with no issues, it could occasionally 'hiccup'. The mouse cursor would be replaced by pixels scattered in the vicinity of the cursor; then, both displays would go dark for a second before returning to their original state. Promptly after trying to shutdown the system, both displays would lose output but the system would never turn off, causing a force shutdown.

 

3.1 This only happened twice but it was very odd. While watching a YouTube video on my primary display, it suddenly was replaced by the static you'd see on old tv's. It only happened those two times and never since.

 

4. Sometimes when I wake up my displays after they've turned off, the displays speakers would no longer show up in the list of audio devices. After running the Windows troubleshooter, they reappear.

 

TROUBLESHOOTING STEPS:

• Re-installing GPU drivers including a complete wipe using DDU
•  Updating BIOS to latest version
• Installing GPU in different slot
• Running the system with minimal components (1xRAM, 1xHDD)
• Testing memory with MemTest86 - 0 errors
• Re-installing Windows
• Stress testing system (CPU and GPU) for one hour - temps normal could be better
• chkdsk C: /f /r, DISM /Online /Cleanup-Image /RestoreHealth, SFC /Scannow commands in command prompt
• RMA'd original card
• Turned XMP off
• Increased TDRdelay value in registry

 

From looking at various forum posts from the past, it appears that this is a TDR display error. I have found a way of recreating the issue which will make it quicker and easier to tell if something works or not. Hope I'll find a solution soon. Thanks.

 

GOOGLE DRIVE LINK: https://drive.google.com/file/d/1mdYU2Y--kWukGhtiBTJ80WenhG-3JsSP/view?usp=sharing

[Frizz]

12 hours ago, drdrewnatic said:

Hi all, first time posting on the forum!

 

SPECS:

CPU: Intel Core i7-9700k - No overclock but some Asus default optimizations ie. Multicore Enhancement are on auto

GPU: Asus ROG Strix Geforce RTX 2070 OC - No overclock apart from factory - Studio Driver v511.65

MOTHERBOARD: Asus ROG Maximus XI Hero (Wifi) - BIOS Version 2004

RAM: 2x8GB G.Skill Trident Z RGB 3200 CL16 - XMP

PSU: EVGA G3 650w Power Supply

STORAGE: Boot: 500GB Samsung 860 Evo SSD, Storage: Seagate 2TB 7200RPM Hard Drive

COOLING: Cooler Master Hyper 212 RGB Black Edition

CASE: NZXT H500

OS: Windows 10 Build 19044.1566

DISPLAY: (2) BenQ BL2420PT Monitors - Connected via HDMI

 

For about a year or two now, I've had many issues regarding my GPU and GPU driver. Eventually after getting fed up with all these problems, I deiced to contact Asus product support as my card is still under warranty. They said that I had to RMA it to fix the problem. After sending it to them, they could not find any issues with the card apart from a wobbling fan. I got back a replacement graphics card and when I put it back in my system, it still has the same problem as before. This leads me to believe that the issue is not coming from the GPU. Am I right?

 

ISSUES:

1. Occasionally, while no load is being put on the GPU, both my displays would freeze and eventually go black. After about 30 seconds to 5 minutes (or a force restart depending on how patient I am), the system would eventually blue screen (while still displaying nothing) and restart without any issue.  The following dump file would be created: (see Google Drive Link). The occurrence of this would be sporadic. The system could go months without any issue while other times it could be an interval of a week.

 

2. More recently (around a year ago), when I turn on my system, the post screen sometimes would not display on any of my monitors and the no VGA debug light on my motherboard turned on. Once the system fully booted into Windows, the lock screen would display however there would be no keyboard or mouse input. After a couple of seconds the display would go dark for a little bit and then display the sign-in screen. Around five minutes after logging on, the system would become very 'laggy' and eventually unresponsive causing me to force shutdown the system.

 

3. If the system would boot correctly with no issues, it could occasionally 'hiccup'. The mouse cursor would be replaced by pixels scattered in the vicinity of the cursor; then, both displays would go dark for a second before returning to their original state. Promptly after trying to shutdown the system, both displays would lose output but the system would never turn off, causing a force shutdown.

 

3.1 This only happened twice but it was very odd. While watching a YouTube video on my primary display, it suddenly was replaced by the static you'd see on old tv's. It only happened those two times and never since.

 

TROUBLESHOOTING STEPS:

• Re-installing GPU drivers including a complete wipe using DDU
•  Updating BIOS to latest version
• Installing GPU in different slot
• Running the system with minimal components (1xRAM, 1xHDD)
• Testing memory with MemTest86 - 0 errors
• Re-installing Windows
• Stress testing system (CPU and GPU) for one hour - temps normal
• chkdsk C: /f /r, DISM /Online /Cleanup-Image /RestoreHealth, SFC /Scannow commands in command prompt
• RMA'd original card

 

From looking at various forum posts from the past, it appears that this is a TDR display error. I have found a way of recreating the issue which will make it quicker and easier to tell if something works or not. Hope I'll find a solution soon. Thanks.

 

GOOGLE DRIVE LINK: https://drive.google.com/file/d/1mdYU2Y--kWukGhtiBTJ80WenhG-3JsSP/view?usp=sharing

Try a CMOS clear. See if XMP is pushing  your system too hard.

 

I suggest that as your PSU is the bare minimum recommended PSU.

 

Also have you tried monitoring your temps?

[drdrewnatic]

8 hours ago, Frizz said:

Try a CMOS clear. See if XMP is pushing  your system too hard.

 

I suggest that as your PSU is the bare minimum recommended PSU.

 

Also have you tried monitoring your temps?

I cleared the CMOS. No issues yet.

 

GPU Idle: 28 °C

GPU Stress (Furmark): 68 °C

CPU Idle: 27 °C

CPU Stress (Prime95): 89 °C

 

Last time I tested the temps I don't remember them being that high. However, I don't believe that that is the issue because it always freezes up when there is no load on the GPU or CPU. To be completely honest, I'm not super surprised by that temp because I'm using a Hyper 212 to cool a 9700k inside the hot-box H500.

 

I thought that that PSU would be enough. On the Asus website is says that the recommended PSU wattage is 550. Am I wrong?

[Frizz]

28 minutes ago, drdrewnatic said:

I cleared the CMOS. No issues yet.

 

GPU Idle: 28 °C

GPU Stress (Furmark): 68 °C

CPU Idle: 27 °C

CPU Stress (Prime95): 89 °C

 

Last time I tested the temps I don't remember them being that high. However, I don't believe that that is the issue because it always freezes up when there is no load on the GPU or CPU. To be completely honest, I'm not super surprised by that temp because I'm using a Hyper 212 to cool a 9700k inside the hot-box H500.

 

I thought that that PSU would be enough. On the Asus website is says that the recommended PSU wattage is 550. Am I wrong?

Hmm guess I looked up the wrong info, maybe the ASUS card is more efficient then the rtx reference one. 
 

Oh also I forgot to ask, does event viewer -> windows logs-> system show when that happens? Are there any error or critical logs?

[drdrewnatic]

Just now, Frizz said:

Hmm guess I looked up the wrong info, maybe the ASUS card is more efficient then the rtx reference one. 
 

Oh also I forgot to ask, does event viewer -> windows logs-> system show when that happens? Are there any error or critical logs?

No errors or critical logs, but I do get this warning when the third issue happens.

 

I included it in the Google Drive link.

[drdrewnatic]

5 minutes ago, Frizz said:

Hmm guess I looked up the wrong info, maybe the ASUS card is more efficient then the rtx reference one. 
 

Oh also I forgot to ask, does event viewer -> windows logs-> system show when that happens? Are there any error or critical logs?

Also forgot to mention that I get this error in control panel when the first issue occurs:

[Frizz]

24 minutes ago, drdrewnatic said:

No errors or critical logs, but I do get this warning when the third issue happens.

 

I included it in the Google Drive link.

Ah sorry I use my phone which can't open those links. Yeah definitely a GPU based error. But since you did DDU I am led to believe you have a faulty PSU if the CMOS clear did not work as you already RMAd the GPU.
 

The only other thing I can think to do is make sure you tried one RAM stick again but chose another one then what you used. To make sure that one is not bad. Best of wishes

[drdrewnatic]

24 minutes ago, Frizz said:

Ah sorry I use my phone which can't open those links. Yeah definitely a GPU based error. But since you did DDU I am led to believe you have a faulty PSU if the CMOS clear did not work as you already RMAd the GPU.
 

The only other thing I can think to do is make sure you tried one RAM stick again but chose another one then what you used. To make sure that one is not bad. Best of wishes

Thanks. It's difficult to find out if what you've done has solved the issue with this as it is very difficult to re-create. I've managed to find a way to trigger it but sometimes it takes more than four hours for anything to happen. Wouldn't that be funny if I spent all that money in shipping to RMA the card just to find out that I could have solved the problem by resetting the CMOS .

 

Will try the RAM if the CMOS reset doesn't work although I'm 99% sure that's not it as I ran MemTest86 and I'd think if there was an issue with any of the sticks, it would have caught it.

 

Anyway, Thank you for your help and I'll keep you posted.

[drdrewnatic]

On 2/27/2022 at 1:20 PM, Frizz said:

Ah sorry I use my phone which can't open those links. Yeah definitely a GPU based error. But since you did DDU I am led to believe you have a faulty PSU if the CMOS clear did not work as you already RMAd the GPU.
 

The only other thing I can think to do is make sure you tried one RAM stick again but chose another one then what you used. To make sure that one is not bad. Best of wishes

Well, 24 hours and still nothing. I'll still leave the thread unsolved as I can't 100% confirm that it solved the problem.

 

Why did turning XMP off work (for now)? I mean yes it's technically overclocking your memory but if the memory is rated for that speed and the processor is more than capable of handling it. How could that effect the GPU? (Again, still not totally sure but theoretically)  

[Frizz]

11 hours ago, drdrewnatic said:

Well, 24 hours and still nothing. I'll still leave the thread unsolved as I can't 100% confirm that it solved the problem.

 

Why did turning XMP off work (for now)? I mean yes it's technically overclocking your memory but if the memory is rated for that speed and the processor is more than capable of handling it. How could that effect the GPU? (Again, still not totally sure but theoretically)  

Only guessing here, certain drivers /services are always there and can crash if unstable for RAM/ enhancements you had on CPU (don't quote me on this but I am almost certain GPU driver is loaded in RAM not VRAM. )

That or it was pushing PSU just a little too hard if it's degraded some. There be monsters when it comes to tech

[drdrewnatic]

On 3/1/2022 at 3:22 AM, Frizz said:

Only guessing here, certain drivers /services are always there and can crash if unstable for RAM/ enhancements you had on CPU (don't quote me on this but I am almost certain GPU driver is loaded in RAM not VRAM. )

That or it was pushing PSU just a little too hard if it's degraded some. There be monsters when it comes to tech

Well I'm pretty certain that resetting the CMOS and turning XMP off hasn't solved the problem. When I got back to my system and woke up my displays, both displays speakers no longer showed up in the list of audio devices. No logs in Event Viewer but all my other audio devices were showing up. Unless this a separate issue that is coincidentally connected to the GPU somehow, I believe this is enough evidence to show that it hasn't fixed the problem. Still no crashes though.

 

I do remember now checking both sticks of memory individually when I first tested them so their not the issue.

 

I could maybe see the power supply causing the issue for the crashes but I can't see how that could disconnect audio devices. The voltages seem normal and it's power draw is normal so how could it affect it?

[ThomasAsmussen]

@drdrewnatic Sorry to post in such an old thread, but im facing something rather similiar with my new ish gpu, did you ever get to the bottom of all this?

[drdrewnatic]

2 hours ago, ThomasAsmussen said:

@drdrewnatic Sorry to post in such an old thread, but im facing something rather similiar with my new ish gpu, did you ever get to the bottom of all this?

As long as you don't manage to summon the mods, you should be ok.

 

I spoke with an EVGA rep and we determined that it was unlikely that my issue would be caused by the power supply. This basically left two components. The CPU and the motherboard. Knowing that the CPU is probably the least likely thing to cause any issues in a system, I turned to the motherboard. Long story short, I was unable to get an RMA for my motherboard which forced me to buy a new CPU + motherboard which I suppose I was due for an upgrade anyway. Four months have gone by and I have yet to encounter any issues though I am not actively looking for it as I am a bit afraid to find it. I am still encountering the 4th issue (Monitor speakers) but I am using my headphones more and more now and it is not affecting me as much now as it was before. I suspect the monitor cables are at fault.

 

I'd be more than happy to assist you in your issue in DM's as I'm not sure what the repercussions are for resurrecting a 6 month old topic.