Crash under gpu load only on M2.

[MykeMadeline]

Sorry if you don't understand me very well, my english isn't the best but I'll try my best.

After a while my pc begin to crash, blue screen appears always with code "WHEA UNCORRECTABLE ERROR", this only happends when I play heavy games or stress GPU in furmark.

I thought It was the GPU or PSU, I replaced both but the blue screen it still happens under same behaviour (When GPU stress so much). Recently I tried Furmark stress in another HDD and surprisingly system didn't crash, so I think m2 is damaged but I'm not sure, I tried reinstall windows in m2 again but It didn't solve it.

This is my PC:
CPU: 5600x.
GPU: RX 6800xt.
ADATA SX6000PNP 256,0 GB M2 (Where windows is installed)
WDC WD10EZEX-22MFCA0 1000,2 GB HDD
ST9250410AS 250,0 GB HDD
XPG PYLON 750W
2x8gb XPG SPECTRIX D41 DDR4 RGB
2x8gb DDR4 3200 PATRIOT

Any idea what it could be?

 

[AMF]

sounds like ur m2 may be getting hot.. do you have a heatsink on it? check smart data ... u can findin dashboard for m2 or in aida64

 

[MykeMadeline]

sounds like ur m2 may be getting hot.. do you have a heatsink on it? check smart data ... u can findin dashboard for m2 or in aida64

Yeah I have it, but honestly after use it I didn't notice any difference in temperature. I have CrystalDiskInfo and it's common to have around 50 °C on m2.

 

[AMF]

Yeah I have it, but honestly after use it I didn't notice any difference in temperature. I have CrystalDiskInfo and it's common to have around 50 °C on m2.

check smart data for corruption ....post the event viewer error and smart data please... do u have cpu overclocked at all? or memory? your psu seems under wattage for your hardware to me ... maybe transient spikes

 

[Mussels]

Check HWinfo64, it may show if you've got issues with the SSD

Example from my OS drive:

You have mismatched ram in your system, could you try with everything setup the way that triggers the fault, but remove two of the sticks and stay with just a matched pair?

Mismatched or incorrectly setup RAM on ryzen can lead to WHEA errors quite easily, some components inside the CPU can affect others in seemingly odd ways (RAM and PCI-E errors can cause USB ports to reset and disconnect, for example)

you arent running a PCI-E riser card to your GPU are you?

 

[AMF]

Check HWinfo64, it may show if you've got issues with the SSD

Example from my OS drive:
View attachment 253574

does adata have a rep for bad drives?

does adata have a rep for bad drives?

good catch

 

[Mussels]

does adata have a rep for bad drives?

That's irrelevant, what matters more is individual products and how they're used - that drive has a high low endurace rating, but HWinfo will let us see the temperatures it's running at, as well as how many reads and writes its done over its lifespan

Edit: when i googled the product code you put in the first post, it showed the pro model with 1200TBW life span.
Further searching showed the non-pro 256GB only has a 150TBW lifespan, which is low - but normal usage shouldnt have used that up yet. HWinfo will tell us for sure.

 

[kapone32]

Sorry if you don't understand me very well, my english isn't the best but I'll try my best.

After a while my pc begin to crash, blue screen appears always with code "WHEA UNCORRECTABLE ERROR", this only happends when I play heavy games or stress GPU in furmark.

I thought It was the GPU or PSU, I replaced both but the blue screen it still happens under same behaviour (When GPU stress so much). Recently I tried Furmark stress in another HDD and surprisingly system didn't crash, so I think m2 is damaged but I'm not sure, I tried reinstall windows in m2 again but It didn't solve it.

This is my PC:
CPU: 5600x.
GPU: RX 6800xt.
ADATA SX6000PNP 256,0 GB M2 (Where windows is installed)
WDC WD10EZEX-22MFCA0 1000,2 GB HDD
ST9250410AS 250,0 GB HDD
XPG PYLON 750W
2x8gb XPG SPECTRIX D41 DDR4 RGB
2x8gb DDR4 3200 PATRIOT

Any idea what it could be?

You are using 2 different sets of RAM. For some reason Windows updates can cause that issue. I was using the same XPG with some Gskill units and after a about 3 months I started getting the exact same issue. Try to only use 2 of the same like the XPG or Patriot and try it again. Another thought process is sometimes updates make you have to separate the power cables so instead of using the 8 pin attached to one cable instead use 2 separate 8 pin cables.

 

[Mussels]

@AMF I have a bad habit of editing my posts after writing them, so you might want to double check i didn't add things in after you saw the posts last.


My guess is you've just got general instability from the mismatched RAM not being configured correctly, and the NVME drive is fast enough to expose it quicker than a HDD.
If you can remove two of the sticks and run just two matched ones with XMP enabled in the BIOS, we can find out for sure.

The second option is that the NVME drive has had an unusual amount of writes, and after seeing first hand someone using a WD green SSD to download torrents and killed it under a week, that sometimes people do things they shouldn't and kill SSD's really fast.

 

[MykeMadeline]

AMF said:
check smart data for corruption ....post the event viewer error and smart data please... do u have cpu overclocked at all? or memory? your psu seems under wattage for your hardware to me ... maybe transient spikes

Sorry, I'm not sure if you mean this:

Event Viewer:

I forgot to mention that when blue screen happens, it doesn't generate any minidumb, It stuck at 0%.

CPU isn't overclocked, RAM either. In fact, I tried furmark with cpu clock boost disabled (Only works between 3600-3700mhz) and xmp disabled too (ram worked at 2400mhz) but doesn't work.

Check HWinfo64, it may show if you've got issues with the SSD

Example from my OS drive:
View attachment 253574

You have mismatched ram in your system, could you try with everything setup the way that triggers the fault, but remove two of the sticks and stay with just a matched pair?

Mismatched or incorrectly setup RAM on ryzen can lead to WHEA errors quite easily, some components inside the CPU can affect others in seemingly odd ways (RAM and PCI-E errors can cause USB ports to reset and disconnect, for example)

you arent running a PCI-E riser card to your GPU are you?

This is the info:

Edit: Now drive temperature is in 39°C, recently PC crashed and I remember that it was around 46.



Yeah, I tried only use one kit of ram but it doesn't work. I don't have a PCI riser.

Edit: Sometimes when blue screen happens, M2 doesn't show in BIOS and I have to boot from it manually.

 

[kapone32]

Sorry, I'm not sure if you mean this:
View attachment 253579

Event Viewer:
View attachment 253581

I forgot to mention that when blue screen happens, it doesn't generate any minidumb, It stuck at 0%.

CPU isn't overclocked, RAM either. In fact, I tried furmark with cpu clock boost disabled (Only works between 3600-3700mhz) and xmp disabled too (ram worked at 2400mhz) but doesn't work.

This is the info:
View attachment 253583

Edit: Now drive temperature is in 39°C, recently PC crashed and I remember that it was around 46.



Yeah, I tried only use one kit of ram but it doesn't work. I don't have a PCI riser.

Edit: Sometimes when blue screen happens, M2 doesn't show in BIOS and I have to boot from it manually.

How do you have your GPU plugged in to the PSU?

 

[eidairaman1]

Check your event viewer.

Also get a screenshot of sensors in hwinfo64 for the power supply

 

[ir_cow]

@Mussels has the right idea here. My first suspicion is the memory as well. Yank one kit and see if your problem goes away.

 

[eidairaman1]

@Mussels has the right idea here. My first suspicion is the memory as well. Yank one kit and see if your problem goes away.

Mixed ram for the majority of the time confuses the bios in determining if spd/xmp can be used. In that instance configuring the modules to the slower clocks and looser timings and higher volts is necessary.

Another reason why to just buy a quad kit and then sell the dual kit

 

[ir_cow]

Another reason why to just buy a quad kit and then sell the dual kit

Well if both kits have the same ICs, you could flash one of the SPDs to match the other. Basically a Quad kit. But 4 DIMMs is usually the problem in the first place.

 

[eidairaman1]

Well if both kits have the same ICs, you could flash one of the SPDs to match the other. Basically a Quad kit. But 4 DIMMs is usually the problem in the first place.

Just means lower speed, mixed quad is the biggest issue

 

[Mussels]

Sorry, I'm not sure if you mean this:
View attachment 253579

Event Viewer:
View attachment 253581

I forgot to mention that when blue screen happens, it doesn't generate any minidumb, It stuck at 0%.

CPU isn't overclocked, RAM either. In fact, I tried furmark with cpu clock boost disabled (Only works between 3600-3700mhz) and xmp disabled too (ram worked at 2400mhz) but doesn't work.

This is the info:
View attachment 253583

Edit: Now drive temperature is in 39°C, recently PC crashed and I remember that it was around 46.



Yeah, I tried only use one kit of ram but it doesn't work. I don't have a PCI riser.

Edit: Sometimes when blue screen happens, M2 doesn't show in BIOS and I have to boot from it manually.

That kernel-power message is just saying the PC shutdown unexpectedly, that message doesnt tell us why.


at 85% the drive is showing wear, but nothing that should affect stability.
Losing 15% of it's lifespan from just 2.5TB writes tho, is terrible. That implies it'll only manage ~40TB before it fails.
Could you run and screenshot AS SSD next?
What i'm mostly interested in is if any speed results are way way slower than they should be, and if those bits of green text in the top left are still green on yours - not being 4K aligned would explain some of this

Next up is removing the second set of ram sticks and going from there. When you take the RAM out, take *clear* photos of the PC so we can see how things are wired up - as someones brought up, we need to see how the GPU is powered, what slots things are in, and various other random bits we havent thought to ask yet could all be revealed