VRAM is broken

[baugeange14f702cf]

Hello, I'm writing my problem here because I couldn't find how to open a ticket elsewhere (can someone direct me to a proper support).

My GPU is a GeForce RTX 3080 GAMING Z TRIO 10G LHR from MSI still under warranty (less than 2 years) order brand new on amazon from "Amazon EU S.a.r.L.".
Here is my problem, my whole computer crashes without any blue screen and without any information in the event history report. The crash occurs in less than 5 minutes when running an expensive program and in a few hours with an inexpensive program thus leading me to test the VRAM (I already checked the RAM and it didn't trigger any crashes). After some experimentation, I was able to reproduce the crash accurately by running a vram test program that produces the specific behavior:

The GPU temperature normally rises 60-70°C but the VRAM temperature rises to 98°C before crashing (one thing to note is that the VRAM temperature rise does not cause a significant increase in the RPM of GPU ventilation (always at 1,000 rpm) versus 3,000 at maximum(if I ask the fan to run at 100%). So the rise in temperature of the VRAM does not trigger the rise in RPM of the GPU. The VRAM test wasn't even close to completing its routine, like 1/5 of the test planned before crashing the computer.

here is a link to the VRAM test launched:

Releases · GpuZelenograd/memtest_vulkan

Vulkan compute tool for testing video memory stability - GpuZelenograd/memtest_vulkan

github.com

In order to monitor the GPU temperature, I was using GPU-Z and monitoring the "memory temperature". (I also doubled checked it using CPUID HWMonitor which display it as memory).

While looking for the origin of the problem I have already updated all possible drivers as well as the motherboard bios.

What should I do? Should I request an RMA from you or from Amazon? Does this appear to be faulty or could it be something else?
(i'm not attaching any file because i don't have much more information i can provide)

 

[Alan J T]

Hello, I'm writing my problem here because I couldn't find how to open a ticket elsewhere (can someone direct me to a proper support).

See https://www.msi.com/support for options

 

[Svet]

Have you tested the GPU in another PC?
What PSU you are using, have you tried a different one?
If you manually force the GPU fans to 100%, does this helps to delay the crash?

 

[baugeange14f702cf]

Have you tested the GPU in another PC?
What PSU you are using, have you tried a different one?
If you manually force the GPU fans to 100%, does this helps to delay the crash?

I can't really test either my GPU or my power supply on different computers.
Yes, the crash still seems to happen in this case (the GPU temperature is kept at an acceptable temperature but the VRAM temperature is still the same).

 

[Svet]

Then not much to do, but bring the PC to computer repair shop to figure what's the issue.
Or you can try to RMA the GPU, but if the problem is not there will be wasting of time.

 

[darkhawk]

Are you using the included GPU riser to keep the GPU from bending when mounted?
I think it would still be possible the VRAM is rising in temp because it's not making good contact with the heatsink, most likely due to the GPU sagging when mounted and not supported.

 

[baugeange14f702cf]

Are you using the included GPU riser to keep the GPU from bending when mounted?
I think it would still be possible the VRAM is rising in temp because it's not making good contact with the heatsink, most likely due to the GPU sagging when mounted and not supported.

i'm using it