Question Is my MB broken?

[Cerox_]

Hi there,
I have been having issues with my PC randomly restarting (no bluescreen, just black and instant restart) for some while now. It occurs randomly in idle, light use or gaming and realiably in Stresstests (Furmark2 + CinebenchR23 all core).

System:

• AMD Ryzen 7 5800x3D
• XFX Speedster MERC 319 Radeon RX 6950 XT Black Gaming
• ASUS ROG Strix B550-F Gaming (WIFI)
• G.Skill Trident Z RGB DIMM 16GB, DDR4-3600, CL16-16-16-36 (2*8 GB)
• Boot Drive: WD_Black SN770 1 TB
• other Drive: Seagate BarraCuda Compute 2TB, 3,5"
• Corsair RM1000x Shift
• Corsair iCUE H115i Elite Capellix
• Reseat GPU within the Top Slot

CPU and GPU run nice and cool, under stress max CPU: 80 °C, GPU: 70 °C, GPU Hotspot: 80 °C

Unsuccessfull things I did:

• Change of Bootdrive (SATA SSD to NVME SSD)
• Upgrade PSU: Corsair RM850e to RM1000x shift
• test RAM modules individually in diferent Slots
• Upgrade CPU: Ryzen 7 3700x to 5800x3D
• Windows reinstall
• DDU
• CMOS reset (everything Stock)
• BIOS Update to Version 3404

Successfull:
Today I tried (i don't know why I didn't do this earlier. I could cry, because I spent 180 € on a PSU which didn't fix anything, but the PSU is so nice I don't want to return it) putting the GPU in the Bottom Slot (Top: PCIe 4.0 16x (CPU), Bottom: PCIe 3.0 16x (Chipset)) and I haven't encountered any crashes so far (not even under the stress test), but the random crashes sometimes don't occur for days.
Since I changed the CPU in the meantime I would assume the PCIe Slot on the MB is broken, but since RMAing takes a while, I want to be 100 % sure, before sending it in.

I'd love to hear your opinions, thx in advance

 

[Roland Of Gilead]

CMOS reset (everything Stock)

Hey there,

When you say reset, what exactly did you do? Short 2 pins?

 

[Cerox_]

Hey there,

When you say reset, what exactly did you do? Short 2 pins?

Yes.

 

[Roland Of Gilead]

Yes.

Given what you have tried and the result with secondary PCIe slot, then it would appear the primary PCIe 4 slot is damaged. Have you cleaned it out with compressed air/ISO? You might just try that before RMA'ing. But, I would just send it back now.

Just for info, PCie 3 16x still can't be fully saturated by most high end GPU's. So you could do just fine as is. Still, RMA is the way to go I think. The issue would just always niggle in the back of my mind.

 

[Cerox_]

Given what you have tried and the result with secondary PCIe slot, then it would appear the primary PCIe 4 slot is damaged. Have you cleaned it out with compressed air/ISO? You might just try that before RMA'ing. But, I would just send it back now.

Just for info, PCie 3 16x still can't be fully saturated by most high end GPU's. So you could do just fine as is. Still, RMA is the way to go I think. The issue would just always niggle in the back of my mind.

Ok imma try the cleaning, also my waranty is running out in the next month or so, so I will have to do it soon anyway.

EDIT: the slot was crispy clean and cleaning it didn't do anything

 

[SkyNetRising]

putting the GPU in the Bottom Slot (Top: PCIe 4.0 16x (CPU), Bottom: PCIe 3.0 16x (Chipset)) and I haven't encountered any crashes so far (not even under the stress test), but the random crashes sometimes don't occur for days.
Since I changed the CPU in the meantime I would assume the PCIe Slot on the MB is broken, but since RMAing takes a while, I want to be 100 % sure, before sending it in.

Only first PCIE x16 slot works in x16 mode.
Second slot PCIE x16 slot is electrically only x4.

Try forcing first PCIE x16 slot into PCIE 3.0 mode. This can be done in BIOS.

BTW - are you using PCIE riser cable?

 

[Cerox_]

Only first PCIE x16 slot works in x16 mode.
Second slot PCIE x16 slot is electrically only x4.

Try forcing first PCIE x16 slot into PCIE 3.0 mode. This can be done in BIOS.

BTW - are you using PCIE riser cable?

No Riser, my GPU is rawdogging the PCIe Slots
If the first slot would work in PCIe 3.0 mode, what would that tell me? Even if if worked, that would still mean the slot is broken, right?

having a temporary fix is nice for me, but I am aiming for a replacement within the warranty, because that would mean a new MB with another 3 years of warranty and that pretty sweet. Not having a PC for 2 Weeks sucks, but I'll manage.

thank you for the suggestion.

 

[SkyNetRising]

If the first slot would work in PCIe 3.0 mode, what would that tell me?
Even if if worked, that would still mean the slot is broken, right?

Not necessary.
Might be a problem with your graphics card. Not handling PCIE 4.0 mode well.

 

[geofelt]

Is your ram a single kit of 2 x 8gb in the original package, or is it two identical separate part numbers?
Ram must be matched for proper operation.
Run memtest86+
It boots from a usb stick and does not use windows.
You can download it here:

Memtest86+ | The Open-Source Memory Testing Tool

Memtest86+ is an advanced, free, open-source, stand-alone memory tester for 32- and 64-bit computers (UEFI & BIOS supported)

memtest.org

If you can run a full pass with NO errors, your ram should be ok.

Running several more passes will sometimes uncover an issue, but it takes more time.
Probably not worth it unless you really suspect a ram issue.

Is your motherboard bios current?
ryzen motherboards constantly come out with fixes related to ram compatibility.
Do not update on speculation. Do so only if you see a potentially relevant fix.

Have you installed the relevant motherboard drivers, directly downloaded from ASUS?
Windows default may not be correct.

If you have installed the asus tweak software, try uninstalling it.

 

[Cerox_]

Not necessary.
Might be a problem with your graphics card. Not handling PCIE 4.0 mode well.

Ok I thought this was a lost cause but so far it works in PCIe 3 mode. I googled a bit and found out that multiple people had that kind of trouble and they were all using ASUS or ASRock Motherboards, but the GPUs were from NVIDIA andAMD and the CPUs were also Intel and AMD.

Are you sure it's the GPU or maybe the MB not handling PCIe 4.0 well?

 

[Cerox_]

Is your ram a single kit of 2 x 8gb in the original package, or is it two identical separate part numbers?
Ram must be matched for proper operation.
Run memtest86+
It boots from a usb stick and does not use windows.
You can download it here:

Memtest86+ | The Open-Source Memory Testing Tool

Memtest86+ is an advanced, free, open-source, stand-alone memory tester for 32- and 64-bit computers (UEFI & BIOS supported)

memtest.org

If you can run a full pass with NO errors, your ram should be ok.

Running several more passes will sometimes uncover an issue, but it takes more time.
Probably not worth it unless you really suspect a ram issue.

Is your motherboard bios current?
ryzen motherboards constantly come out with fixes related to ram compatibility.
Do not update on speculation. Do so only if you see a potentially relevant fix.

Have you installed the relevant motherboard drivers, directly downloaded from ASUS?
Windows default may not be correct.

If you have installed the asus tweak software, try uninstalling it.

As I replied above, the issue seems to be PCIe 4.0, but I will answer this nevertheless, in case it might be also relevant.

I am running a proper Kit (same package, same part numbers) G.Skill Trident Z RGB F4-3600C16D-16GTZR (CL16-16-16-36) They are high quality modules with memory Chiplets from Samsung and were way to expensive back then. I am using the DOCP profile.

I have previously tested the modules individually in different Slots to no effect, so I was pretty confident it wasn't the RAM because what are the odds both memory modules break down with the same issue around the same time.
Do you think it is still worth it or that RAM could be in any way responsible for the PCIe issue?

Also yes BIOS recent (end of October 23), newest BIOS came out beginning of April, but I think it only adresses a recent security problem, so I might need to update anyway.

I am using Armoury Crate (horrible software) to have all my drivers up to date and it works well for that.

No I am not using any ASUS tweak software only AMD Adrenaline, but the problems persisted even on a clean install.

Thank you for your contribution

 

[SkyNetRising]

Are you sure it's the GPU or maybe the MB not handling PCIe 4.0 well?

Could be GPU, Motherboard and also CPU.
Any of them or all of them in this particular configuration.

You're not loosing much with forced PCIE 3.0 mode.
May be ~2% - 3% performance. Probably less.

https://www.techpowerup.com/review/nvidia-geforce-rtx-4090-pci-express-scaling/28.html

 

[Cerox_]

Could be GPU, Motherboard and also CPU.
Any of them or all of them in this particular configuration.

You're not loosing much with forced PCIE 3.0 mode.
May be ~2% - 3% performance. Probably less.

https://www.techpowerup.com/review/nvidia-geforce-rtx-4090-pci-express-scaling/28.html

Ok so I was excited a little to early. While I still can't make it crash with furmark and Cinebench (which always got it in 4.0 mode) when the PCIe Slot is in 3.0, it just died on me playing LoL.

 

[Cerox_]

It is so super frustrating, that fixes often appear to work for a day and then it breaks down again. Followed that pattern after changing the CPU, changing the SSD, later wiping the SSD and now forcing PCIe 3.0.

 

[Richj444]

It is so super frustrating, that fixes often appear to work for a day and then it breaks down again. Followed that pattern after changing the CPU, changing the SSD, later wiping the SSD and now forcing PCIe 3.0.

I would really suggest swapping a different GPU into your system and / or your current GPU into a different system and see what happens with your issue. Does it follow your GPU? Does your GPU work fine in a different system? Or does the problem disappear entirely?

Note that to make it a valid test the GPUs and system components should be as similar as possible.

 

[Cerox_]

I am so furious. Again it worked fine for days in the bottom PCIe Slot, but now suddenly it decided not to work again and crash basically every other minute.
I have not been able to get another system or GPU to troubleshoot, so I will try the Memtest, since I don't have any other options right now.
I will also try to disconnect my HDD and maybe Boot of my old SSD again to check whether that makes any difference, even though the chances are extremely slim.