Seemingly Random GPU Crashing

[Fizzik]

Hey guys. Been all over several forums and have tried resolving this issue but have had no luck, was hoping I might have better luck here.

 

Long story short, at random times anywhere from right after logging in to while I'm playing a game, my computer will start lagging until I get a full crash.

 

Even on just a blank windows desktop, it will sometimes start lagging and then black screen. Then the lights on my keyboard will shut off. The fans stay running though.

 

If I look at the POST monitor LED, it shows a VGA error. This led me to believe that it is the GPU. However, I have replaced the GPU with a similar model (same card, different brand) and the issue still happens.

 

It will eventually resolve itself, but takes a few reboots. I have tried all manners of fixes, from complete driver installs, windows reinstalls (from scratch, 3 times), clearing cmos, checking temps, changing bios settings to default, etc. I am at a loss here.

 

My specs are:

Ryzen 5800X3D 

Sapphire Nitro+

RX 6700 XT 

32GB G.Skill TridentZ-3600 

1TB Micron NVMe

240GB Samsung NVMe

500GB Samsung SATA

Corsair SF-600w SFX PSU 

B550I MSI Gaming Edge

 

I just spent $120 total (round trip) shipping my original card to California for an RMA just for that to not be the issue, so I'm sure you can imagine my frustration.

 

Thanks for reading through all this!

[ItsDestini]

Yeah ngl the psu is kinda sus here because just looked up the recommended PSU for the 6700 xt and it says the optimal amount should be 750W so yeah plus you've got a pretty high end and maybe unstable cpu because of the model. I suggest upgrading the PSU and maybe it will fix it

[ItsDestini]

This might confirm the previous anwser. Hope this helps

 

[DuckzWasTaken]

So from my knowledge, it could possibly be a PSU issue if you have the time i would recommend taking each part out and making a test bench your PC could just have a lack of power, And another thing if you haven't done it yet try plugging it into a seprate outlet

[Fizzik]

15 minutes ago, ItsDestini said:

Yeah ngl the psu is kinda sus here because just looked up the recommended PSU for the 6700 xt and it says the optimal amount should be 750W so yeah plus you've got a pretty high end and maybe unstable cpu because of the model. I suggest upgrading the PSU and maybe it will fix it

 

13 minutes ago, DuckzWasTaken said:

So from my knowledge, it could possibly be a PSU issue if you have the time i would recommend taking each part out and making a test bench your PC could just have a lack of power, And another thing if you haven't done it yet try plugging it into a seprate outlet

I know the minimum listed is 750 but I ran a check on PC Part Picker and it seemed like the wattage was all good. Maybe not the case I guess?

[ItsDestini]

3 minutes ago, Fizzik said:

 

I know the minimum listed is 750 but I ran a check on PC Part Picker and it seemed like the wattage was all good. Maybe not the case I guess?

Yeah I would recommend swapping the PSU and seeing whether that fixes it because yeah youve got a pretty high end system regarding other components and the PSU just seems low end hope this helps 

[191x7]

43 minutes ago, Fizzik said:

Hey guys. Been all over several forums and have tried resolving this issue but have had no luck, was hoping I might have better luck here.

 

Long story short, at random times anywhere from right after logging in to while I'm playing a game, my computer will start lagging until I get a full crash.

 

Even on just a blank windows desktop, it will sometimes start lagging and then black screen. Then the lights on my keyboard will shut off. The fans stay running though.

 

If I look at the POST monitor LED, it shows a VGA error. This led me to believe that it is the GPU. However, I have replaced the GPU with a similar model (same card, different brand) and the issue still happens.

 

It will eventually resolve itself, but takes a few reboots. I have tried all manners of fixes, from complete driver installs, windows reinstalls (from scratch, 3 times), clearing cmos, checking temps, changing bios settings to default, etc. I am at a loss here.

 

My specs are:

Ryzen 5800X3D 

Sapphire Nitro+

RX 6700 XT 

32GB G.Skill TridentZ-3600 

1TB Micron NVMe

Corsair SF-600w SFX PSU 

B550I MSI Gaming Edge

 

I just spent $120 total (round trip) shipping my original card to California for an RMA just for that to not be the issue, so I'm sure you can imagine my frustration.

 

Thanks for reading through all this!

 

1st - do you use the latest available Bios for the motherboard? If not, flash it, this might solve the issues.

2nd - if you're running XMP (DOCP), try disabling it, if your RAM is unstable this might help.

3rd - make sure you are using the latest available SSD firmware

4th - I am not sure the PSU would be the culprit here, the system is not anywhere near 600W under load. But if it's faulty - trying with a different one might not be a bad idea.

5th - check your temperatures. You did not mention your CPU cooler, and the 5800X3D gets quite hot, especially if not undervolted (PBO curve offset negative 15-20).

 

My bet would be the Bios or temperatures. Just go in the mentioned order. Report back.

[Fizzik]

2 hours ago, 191x7 said:

 

1st - do you use the latest available Bios for the motherboard? If not, flash it, this might solve the issues.

2nd - if you're running XMP (DOCP), try disabling it, if your RAM is unstable this might help.

3rd - make sure you are using the latest available SSD firmware

4th - I am not sure the PSU would be the culprit here, the system is not anywhere near 600W under load. But if it's faulty - trying with a different one might not be a bad idea.

5th - check your temperatures. You did not mention your CPU cooler, and the 5800X3D gets quite hot, especially if not undervolted (PBO curve offset negative 15-20).

 

My bet would be the Bios or temperatures. Just go in the mentioned order. Report back.

Bios is up to date. I leave XMP disabled by default just because the benefit is marginal with the 3d cache.

 

The temps for the cpu do get a little hot, but they don't overheat. It's a kraken x52 240mm aio. If it were temps, it should just cause the pc to shut off though, not black screen with it running.

 

I can check the SSD firmware, didn't occur to me that there'd be a correlation there. Good thought!

[191x7]

1 hour ago, Fizzik said:

Bios is up to date. I leave XMP disabled by default just because the benefit is marginal with the 3d cache.

 

The temps for the cpu do get a little hot, but they don't overheat. It's a kraken x52 240mm aio. If it were temps, it should just cause the pc to shut off though, not black screen with it running.

 

I can check the SSD firmware, didn't occur to me that there'd be a correlation there. Good thought!

RAM speed effect with the X3D isn't that marginal. It makes 3000, 3200, 3333 or 3600 not that different to each other but there still is a huge difference between 2133/2400 and 3200/3600.

If it works, XMP should be on.

[Fizzik]

11 hours ago, 191x7 said:

RAM speed effect with the X3D isn't that marginal. It makes 3000, 3200, 3333 or 3600 not that different to each other but there still is a huge difference between 2133/2400 and 3200/3600.

If it works, XMP should be on.

Either way I'm not too worried about that right now haha, I want to solve the video issue first before enabling XMP.

[191x7]

18 minutes ago, Fizzik said:

Either way I'm not too worried about that right now haha, I want to solve the video issue first before enabling XMP.

Can you disable the Multiplane Overlay MPO and retest?

[Fizzik]

1 hour ago, 191x7 said:

Can you disable the Multiplane Overlay MPO and retest?

Never heard of that, will give it a shot. Thanks.

[Fizzik]

On 7/28/2023 at 11:43 AM, 191x7 said:

Can you disable the Multiplane Overlay MPO and retest?

Did not solve it. It crashed just now, but strangely enough it restarted on its own. Never had that happen before.

[Fizzik]

Another thing... Interesting revelation.

 

PC would not boot up completely, which usually requires me to open it up and reseat the gpu. This is a fairly common occurrence at this point. I'm wondering if reseating the gpu was unnecessary.

 

I just physically moved the case while the pc was "on" (stuff was spinning but no video out, like usual) and it immediately fired up. I'm wondering if it's shorting with the case somehow??

 

I'll test it a few more times when this issue happens again to know if it works 100% of the time.