Intermittent stability issues of brand new high end system

[jobrick3]

Hello everyone,

I recently built a new PC, but I am facing stability and booting issues.

Specs:

• CPU: AMD 7950X
• Motherboard: ASUS Crosshair X670E Hero
• GPU: Gigabyte 4090 Aorus Master
• RAM: Corsair Vengeance 96GB Kit (2x48GB), DDR5-6000, CL30-36-36-76, 1.4V (CMH96GX5M2B6000C30)
• Storage: Samsung 990 Pro 4TB
• Power Supply: BeQuiet Straight Power 12, ATX 3.0, 1200W
• Cooler: LianLi Galahad Trinity 360 Performance II
• Case: LianLi Lancool III
• Operating System: Windows 11 Pro, 22H2, Build 22621.2506

Settings:

• All components are running at standard settings. Notably, the RAM is running at DOCP I settings according to the manufacturer's specifications (timings, frequency). However, I had to manually set the UCLK/MCLK ratio to 1:1 in the BIOS, which is unusual since 6000MHz RAM modules are typically expected to run in 1:1 mode automatically.

Issues:

• Initial Stability: During my first Windows 11 installation, the computer unexpectedly restarted during the BIOS loading
• Booting Issues: Sometimes (not always), the PC freezes during boot-up right after the ASUS ROG logo and just before loading the Windows 11 login screen. The screen just shows black. Forced restart required.
• BSODs and Random Restarts, both during boot-up and operation in Win11 (e.g., TeamViewer sessions).
• Issues when waking up from Sleep Mode: Freezes and graphical glitches on Win11 login screen upon waking from sleep. Forced restart required.
  
  
• Black Screen: Once, after leaving the PC running for a while, I came back to a black screen. I couldn't wake up or restart the system. Only a CMOS reset helped
• System Responsiveness: Rarely, there are moments when Win 11 is laggy or unresponsive. Often it resolves on its own after a few seconds, but a restart was necessary once.

Issues with the PC are intermittent, often running smoothly for a week before problems resurface.

Problem Escalation today:

1. First Restart: When waking from sleep mode, the PC restarted on its own.
2. Second Restart: The system was laggy and there were graphical errors (white stripe). After a few seconds, it led to another restart.
   
   
3. Third Restart: The PC booted past the ASUS logo screen, but not until Win 11 login screen. Instead, there was just a black screen, followed by another restart.
4. Fourth Restart: PC froze during boot-up at the ASUS logo loading animation, followed by another restart.
5. Complete Failure: After that, the PC failed to boot to the BIOS loading screen, showing only a black screen.
6. (Temporary?) fix: I noticed that the motherboard code was always stuck at "Memory Training" (Code 15) and finally switched to "00" (= not used). A CMOS reset did not help. After turning the power supply off and on and another CMOS reset, a very long memory training process was carried out, and the PC started normally.

BSOD Dump file readouts:

1. BSOD 1
2. BSOD 2
3. BSOD 3
4. BSOD 4
5. BSOD 5
6. BSOD 6

Troubleshooting steps taken so far:

1. I've swapped out the RAM modules, but the problem persists.
2. I've ensured the motherboard firmware is up-to-date.
3. Double-checked the following BIOS settings: CSM = deactivated, Fast boot = activated
4. Turned off Windows fast startup in System Settings.
5. Changed BIOS Settings:
   • OS Type: Changed from [Other OS] to [Windows UEFI mode]
   • Secure Boot Mode: Changed from [Custom] to [Standard]
6. Reinstalled and updated the GPU and chipset drivers. Old drivers were removed using DDU and the "Microsoft Install and Uninstall Troubleshooter" in Safe Mode.
7. Made sure Win 11 is up-to-date
8. MemTest86 ran for two passes without any errors
   
   

Next steps:
I’m at a loss for what to try next. Any suggestions or similar experiences?

• Is it a software issue? Re-install Windows?
• Is it a hardware issue?
  • Reseat GPU?
  • Memory incompatibility?
  • RMA mainboard / GPU / memory / SSD?

 

[faye]

Did you swap the RAM for another identical kit or something different?

 

[jobrick3]

Did you swap the RAM for another identical kit or something different?

I had another Corsair kit before (2x 48GB 6400 MHz running @ DOCP I in 1:2 mode).
I had the same issues with that previous kit. I swapped it for the 6000 MHz version to run it in 1:1 mode straight out of the box.

I am considering to testing the RAM at basic Speed @ 4800 MHz. And I also heard entering the DOCP I settings manually increases stability vs. simply enabling DOCP I.

 

[unwind-protect]

My money is on a broken mainboard.

 

[wheresmycar]

have you tried removing the 4090 and running with the 7950X's radeon graphics? If not, give that a shot to iron out the 4090 being the culprit.

Alternatively set the BIOS to default and disable XMP (memory OC) and see if that eliminates the issue.

 

[jobrick3]

have you tried removing the 4090 and running with the 7950X's radeon graphics? If not, give that a shot to iron out the 4090 being the culprit.

Alternatively set the BIOS to default and disable XMP (memory OC) and see if that eliminates the issue.

Thank you! I will try that. Troubleshooting is just a bit annoying as I cant force the issues and it might take days or weeks until the next problem occurs.

 

[Bill_Bright]

Your next step needs to be swapping in a known good power supply since EVERYTHING inside the case depends on good, clean power.

What are your temps?

 

[Assimilator]

Considering that two-thirds of the BSODs are around DirectX, and the images of graphical corruption you've shown, my suspicion would be the 4090. As has already been suggested, try with the CPU's IGP - or another backup GPU.

 

[jobrick3]

Hi! Quick update:

Currently, the system is not booting any more. There is either a BSOD after Win login or when entering the BIOS everything freezes after a couple of seconds / minutes.
I haven't done a CMOS reset trying to fix it, as I thought I could use this state to test different hardware components.

Results:
- runing the RAM wit DOCP I = disabled @ 4800 MHz did not fix the issue
- removing 4090 and running everything via iGPU did not fix the issue

Components left to test:
- Mainboard
- RAM
- SSD
- PSU

 

[jobrick3]

Update:
When runing the RAM wit DOCP I = disabled @ 4800 MHz I got into a situation in which the computer got a BSOD immediately at win11 login or shortly after. In the BIOS I got constantly freezes requiring a forced restart.

I brought the PC to the local PC repair shop. After they enabled the DOCP I profile again, the PC started without any problems. They stress tested everything for 2 days non-stop and could not induce another BSOD.

As the issues are intermittent, they suggested me trying memory sticks from a different manufacturer to see whether it is a compatibility issue (even thought he Corsair sticks are on the QVL). They said they have often problems with Corsair RAM.

Would you suggest either:

1. Gskill Trident Z5 RGB DDR5-6400 (2x48GB, CL32-39-39-102, 1.35V): I want to set it manually to 6000 MHz in 1:1 mode. This kit is on the QVL at its native 6400 MHz frequency.
2. Kingston FURY RGB DDR5-6000 (2x 48GB, CL32-38-38, 1.35V): Not on the QVL.

Would you expect those kits making actually a difference to my stability issues, as they probably have all the same Hynix M dimms?

Alternatively, should I send back the Mainboard and SSD and get a replacement - just in case?