Is this a GPU issue? Black screen and fans at 100%

[svalloric]

Hello all

 

I have a question.  I bought a custom built PC a the end of June.  I added a second NVMe and a SSD to it from my old PC.

 

Around mid-August... I had an issue for a few days where the PC would shut off the displays... and the fans in the system would go to full bore.  Especially on the GPU.  I would then have to shut down the system with the power button and restart. The GPU is a 3060Ti.

 

I did some things like using DDU to get rid of the display driver... and I updated to the newest at the time.  Still had the issue so I switched PCIe slots.  Mid-august.   I played many games... not many graphically intensive... some MMOs, Diablo 4... etc. 

 

No more problems... until yesterday.

 

After some googling... I have read it is either:  RAM, PSU, GPU

 

I did memtest last night with no errors.

 

This morning... I had a small monitor plugged into the mobo to use (trying to test if it was the GPU)... and this morning. and as it was just sitting there loading Discord ... I heard the fans on the GPU go sky high again.  I didn't lose video signal since I was using the onboard graphics... I could use the computer normally.  I shut it down before coming in to work today.  So I am looking for some guidance as to what you all may think the problem is now. 

 

I think I have the RAM ruled out?

 

AI mean it COULD be the PSU... but I would assume I would have an issue with computer after the fans on the GPU went to 100%... even though I wasn't using the graphics card.  That makes me think it is some sort of driver/windows/GPU issue.  I did a clean install of GPU drivers last night as well... but had the problem.

 

Any guidance with that info?

 

 

[RevGAM]

Take the GC out and see what happens. 

[191x7]

11 minutes ago, svalloric said:

Hello all

 

I have a question.  I bought a custom built PC a the end of June.  I added a second NVMe and a SSD to it from my old PC.

 

Around mid-August... I had an issue for a few days where the PC would shut off the displays... and the fans in the system would go to full bore.  Especially on the GPU.  I would then have to shut down the system with the power button and restart. The GPU is a 3060Ti.

 

I did some things like using DDU to get rid of the display driver... and I updated to the newest at the time.  Still had the issue so I switched PCIe slots.  Mid-august.   I played many games... not many graphically intensive... some MMOs, Diablo 4... etc. 

 

No more problems... until yesterday.

 

After some googling... I have read it is either:  RAM, PSU, GPU

 

I did memtest last night with no errors.

 

This morning... I had a small monitor plugged into the mobo to use (trying to test if it was the GPU)... and this morning. and as it was just sitting there loading Discord ... I heard the fans on the GPU go sky high again.  I didn't lose video signal since I was using the onboard graphics... I could use the computer normally.  I shut it down before coming in to work today.  So I am looking for some guidance as to what you all may think the problem is now. 

 

I think I have the RAM ruled out?

 

AI mean it COULD be the PSU... but I would assume I would have an issue with computer after the fans on the GPU went to 100%... even though I wasn't using the graphics card.  That makes me think it is some sort of driver/windows/GPU issue.  I did a clean install of GPU drivers last night as well... but had the problem.

 

Any guidance with that info?

 

 

Which CPU/motherboard?

[svalloric]

21 minutes ago, RevGAM said:

Take the GC out and see what happens. 

Yeah going to try that this evening when I get home... was afraid to leave the computer unattended if the fans went crazy again.

 

I have two questions then:

1.  should I delete the nVidia display drivers? since I will be using onboard mobo.

 

2.  Does the fast that the computer still works when this happens (and hoked up to mobo video out) mean the PSU is ok?

 

[svalloric]

18 minutes ago, 191x7 said:

Which CPU/motherboard?

 Intel® Core™ Processor i5-13600K 6P/12 + 8E 3.50GHz [Turbo 5.1GHz] 24MB Cache LGA1700

ASUS PRIME Z790-P WIFI D5 DDR5 ATX w/ Wi-Fi, 2.5GbT LAN, (4)PCIe x16,(1)PCIe x1, (3)M.2, (4)SATA

 

Jus in case you need it... I am not at home to see what it actually is... but my invoice shows a PSU as 

800 Watts - Standard 80 Plus Gold Certified Power Supply

[ItTakes2ToMango]

I'd say it's unlikely to be your PSU. I think we just need to wait for you to get home and try a few things.

 

My first guess is on the GPU or CPU overheating. I doubt it's RAM, but at this moment it's up in the air until you try a few more things

 

Can you get us temps on the machine during the crazyness?

[svalloric]

25 minutes ago, ItTakes2ToMango said:

I'd say it's unlikely to be your PSU. I think we just need to wait for you to get home and try a few things.

 

My first guess is on the GPU or CPU overheating. I doubt it's RAM, but at this moment it's up in the air until you try a few more things

 

Can you get us temps on the machine during the crazyness?

To be quite honest I am not 100% sure. I think in August when this first happened and I got it back to where it seemed like it was ok... I think I ran Cinebench for 30 min and Heaven.... and I don't think it ever got about 80C?  Maybe 72C was the top (running HWMonitor)

 

I think I ran HWMonitor all day long one day while I was at work and it only was like 70-71C for the CPU not sure on the GPU... but never saw any red warnings for the GPU.

 

I just know I never saw any red warnings in the CPU temps spot.

 

I am not sure what the temps were when it happened last night... but the bios (when I going in to boot from the USB for memtest) was 37C when it started and dropped to 30-31C almost immediately.

 

Just weird it happened like 2 days in August... and once I tinkered with a few things... it was fine until I got home yesterday.

[svalloric]

I did notice yesterday the FIRST time the thing went nuts (I was not home when it did)... there was an even in the event viewer for the display driver crashing (which is also weird because I would assume my monitors were asleep/off when I was not home on the computer).

 

But the 2nd, 3rd, and 4th times it happened... no such error showed up.

 

[GuiltySpark_]

1 hour ago, svalloric said:

I had an issue for a few days where the PC would shut off the displays... and the fans in the system would go to full bore.  Especially on the GPU. 

Continue troubleshooting but this is a classic symptom of a dead/dying GPU. Every single time, same exact symptom. Prepare for that eventuality. 

[RevGAM]

1 hour ago, svalloric said:

Yeah going to try that this evening when I get home... was afraid to leave the computer unattended if the fans went crazy again.

 

I have two questions then:

1.  should I delete the nVidia display drivers? since I will be using onboard mobo.

 

2.  Does the fast that the computer still works when this happens (and hoked up to mobo video out) mean the PSU is ok?

 

1. Can't hurt. As @ItTakes2ToMangosaid, it's likely to be the GC, but could be something else. 

2. Probably but not definitely. 

 

When you get home, before you remove the GC, run a 3DMark benchmark like Port Royal or Time Spy. You may need to have HWInfo recording temps as well. Post results.

 

If that doesn't cause the problem, do whatever normally causes it. 

 

Then remove the GC, remove the drivers,  and do the benchmark and other stuff, same way, and post those results. 

[svalloric]

1 hour ago, RevGAM said:

1. Can't hurt. As @ItTakes2ToMangosaid, it's likely to be the GC, but could be something else. 

2. Probably but not definitely. 

 

When you get home, before you remove the GC, run a 3DMark benchmark like Port Royal or Time Spy. You may need to have HWInfo recording temps as well. Post results.

 

If that doesn't cause the problem, do whatever normally causes it. 

 

Then remove the GC, remove the drivers,  and do the benchmark and other stuff, same way, and post those results. 

Will do.  Although this morning I was just opening Discord when the thing went bananas... and I wasn't even outputting via the GPU... I was hooked up to the mobo's output.

[ItTakes2ToMango]

19 minutes ago, svalloric said:

Will do.  Although this morning I was just opening Discord when the thing went bananas... and I wasn't even outputting via the GPU... I was hooked up to the mobo's output.

This convinces me even more that it's your GPU dying

 

IIRC Hardware acceleration still works with just the PCIe connection. Hopefully someone else can correct me if I'm wrong

 

[svalloric]

14 minutes ago, ItTakes2ToMango said:

This convinces me even more that it's your GPU dying

 

IIRC Hardware acceleration still works with just the PCIe connection. Hopefully someone else can correct me if I'm wrong

 

Same.... I mean.. the computer still worked... just the fans on it went bananas. 

[RevGAM]

41 minutes ago, svalloric said:

Same.... I mean.. the computer still worked... just the fans on it went bananas. 

Yeah, is your GC under warranty?

[svalloric]

I would assume.  Bought it in a system in June.

[svalloric]

Home now... going to do some testing.

 

Here if HWInfo sensor stuff  going in

 

 

[RevGAM]

10 minutes ago, svalloric said:

Home now... going to do some testing.

 

Here if HWInfo sensor stuff  going in

 

 

Are those idle or load temps?

[GuiltySpark_]

1 minute ago, RevGAM said:

Are those idle or load temps?

Core load at 15% so, take that as you will. Not quite idle but certainly not load. Also 210Mhz.

[svalloric]

Could my RAM voltages and or frequencies be jacked up?

 

I didn't do anything but enable my RAM to 6000Hzand enable XMP.  I think.

[RevGAM]

2 minutes ago, GuiltySpark_ said:

Core load at 15% so, take that as you will. Not quite idle but certainly not load. Also 210Mhz.

Thanks. Math is my weakness. Seems a bit high, don't you think?

[svalloric]

6 minutes ago, RevGAM said:

Are those idle or load temps?

That was prior to loggin into the MMO... Star Wars:  The Old Republic for about 15 min.

 

Max temp during that play ing for about 15 min was 58.6... hottest spot was 71.0.

 

Was waiting for 3D Mark to download.

 

About to run the tests now.

[svalloric]

Crapped the bed into the first pass of Port Royale.

 

What do I need to upload here?

[svalloric]

Here is the csv HWInfo Saved.

gputest3.CSV

[GuiltySpark_]

1 minute ago, svalloric said:

Crapped the bed into the first pass of Port Royale.

 

What do I need to upload here?

Nothing, really. I'll make this quick. 

 

The fastest possible diagnosis comes from just sticking the GPU in another system and trying to use it. If the card is dying, it'll happen again. You have your culprit, on to replacement. 

 

If you don't have that option, the best you can do is DDU. Still happens? Have another GPU to swap into this PC? Still happening? 

 

Unfortunately the sure fire ways to narrowing down a problem like this is part swapping and following the problem. Given your symptoms, and the dozens of identical situations i've seen with identical symptoms, the card is simply faulty and requires replacement. There really isn't anything in software you're going to be able to show here that is of any use. Overheating isn't going to do what your system is doing, the card will simply downclock heavily. 

I'd get that RMA page fired up. 

[svalloric]

Saw this in event viewer.... this seems to pop up a lot when this happens.