ASUS Maximus Z790 Hero BSOD, WHEA Errors and Application Crashes

[UserX]

Greetings everyone,

I've recently undergone a system upgrade, the specs of which are listed below. However, I've run into some troublesome issues including WHEA Errors, BSODs, and application crashes. The system initially ran smoothly for about three months but gradually began to degrade in performance. I've been diligently troubleshooting these problems and I believe I've pinpointed the issue to the CPU. Before I take my troubleshooting results to the manufacturer for a warranty claim, I wanted to share my findings here and gather your opinions and confirmations regarding my assessment.

Here are the system specifications:

• Operating System: Windows 11 Pro 22H2
• CPU: Intel 13900KS
• Motherboard: ASUS Maximus Z790 Hero
• RAM: 64GB Kingston FURY Beast RGB DDR5 6000MT/s M-Die
• GPU: ASUS TUF Gaming GeForce RTX™ 4090 OC Edition 24GB GDDR6X
• Audio Interface: Focusrite Scarlett 2i2 Gen3
• Monitors: Dell Alienware AW3423DW 34" OLED, LG 38WK95C-W HDR WQHD 36" IPS, ASUS ROG Swift PG278Q 27" TN
• Storage: m.2_1> 1X 2TB Samsung 990 PRO on Intel RST/VMD, m.2_2/3> 2X 2TB Samsung 980 PRO on Intel RST/VMD
• Power Supply: Corsair HX1500i 1500Watt
• Cooling: CPU/GPU Cooling with HydroLux PRO Custom Liquid Cooling Loop with HardLine Tubing, 14X Corsair QL Series (RGB Fans), 3X Corsair Commander PRO Management Control Board
• Keyboard: Logitech G815
• Mouse: Logitech G502 with Logitech Powerplay

BIOS Setting during testing.

• Asus Multicore Enhancement: "Disabled – Enforce All Limits”
• Turned off XMP
• CPU Load-line calibration: “Level 6”
• Long Duration Package Power Limit: “150”
• Short Duration Package Power Limit: “253”

Additional Software

• HWInfo
• AIDA64
• WhoCrashed
• WInRAR
• 7-ZIP
• ROG-GPUZ
• Intel Processor Diagnostic Tool
• Intel XTU
• Prime95
• Cinebench R23
• Sysinternals toolset
• And a few other NirSoft tools for diagnostics that were never needed.

Here's a summary of the testing I've conducted:

• Reset CMOS and loaded Defaults
• Base install of Windows and its updates with no preview or insider updates
• Fresh Windows installation multiple times with different driver combinations.
• MemTest on each stick of DDR5 in each RAM slot (all passed).
• Heavy GPU stress testing (no issues).
• BIOS update to version 1501 and CMOS reset (problems persisted).
• SMART and load testing on all drives (all passed).
• Fresh Windows install on a new Samsung 990 Pro NVMe (problems persisted).
• Fresh Windows install on a single SATA SSD (problems persisted).
• CPU load testing and Intel Diagnostic Tool (all passed).
• Experimented with different driver combinations (problems persisted).

After extensive testing, I noticed consistent 7-Zip CRC errors when installing Nvidia drivers. I decided to use WinRAR and the Task Mangers Affinity settings to test the cores. I used WinRAR to extract the driver files after setting the affinity to a P-Core and Hyper Thread pair. I then ran the test on each P-Core and its corresponding Hyper Thread one at a time. I ran into CRC errors on 2 specific CPU cores (P-Core 4 (8/9) and 5 (10/11).  Both cores were unstable and produced CRC errors on multiple runs. All other P-Cores and Hyper Threads had no issues. All E-Cores passed as well.

To stabilize the CPU, I adjusted the SVID Behavior to "Intel Fail Safe" and set the "Specific Core Ratio" to 56 on P-Cores 4 and 5, which are the favorite cores with a default core ratio of 60. Notably, not setting the SVID Behavior to "Intel Fail Safe" caused system instability.

For the past 24 hours, I have not encountered any BSODs or WHEA errors, with only one application crash, which might be attributed to running on a Windows installation initially used for CPU stability testing.

Do you agree with my diagnosis? Is this something I can confidently present to the manufacturer for further investigation?

Thank you for your insights and assistance.

12/06/2023 Update - 
Adjusting the SVID and Ratio settings has had a positive impact on system stability, but I'm still experiencing occasional crashes, although they are occurring less frequently. It appears that the problem is becoming more persistent so I'm currently in the process of pursuing an RMA for the 13900KS CPU.

12/10/2023 Update - 
The system stability has worsened. I changed the Specific Core Ratio to 54 on all cores to stabilize it.

I am currently in the process of handling the RMA for the 13900KS. Fortunately, I successfully convinced the manufacturer to acknowledge and agree with my diagnosis of the issue. However, it's worth noting that they no longer have the 13900KS in stock. Instead, they have proposed the 14900K as a replacement option, but with an additional $60.00 charge. They have also asked me to sign an agreement releasing them from any liability in case of any damage to the socket.

Considering these circumstances, I am leaning towards not engaging in future business transactions with this manufacturer and plan to resume building my own systems. I intend to provide a comprehensive review of this experience on another social platform at a later date.

12/16/2023 Update -
I purchased a 14900K which I installed a few days ago. My system is now 100% stable. The 13900KS RMA is still in process.

[UserX]

DELETED and moved into the original post.

[Silent_Scone]

Hi @UserX ,

1. Clear the CMOS via the button on the rear IO panel

2. Disable MCE by enforcing all limits within the UEFI. (Leave SVID behaviour in Auto)

3. Launch Windows, download HWiNFO and load the sensor panel

4. Download and run several loops MAXON Cinebench R20.0

5. Inspect temperatures and verify if the stress test passes. Run for 3-4 loops.

[UserX]

I do have MCE disabled. I'll add it to the troubleshooting steps. 

I will try this. For the record, I can only run the test if the SVID behavior is set to Intel Failsafe and the Specific Core Ratio for P-Cores 5 and 6 are set to 56. Without those settings, I get Random App crashes, BSOD's --A wide range of them--, hanging, freezing, etc.

Here are some additional settings that I need to add into my troubleshooting steps. I completely forgot about these changes until your post.

Asus Multicore Enhancement: “Disabled – Enforce All Limits”
CPU Load-line calibration: “Level 6”
Long Duration Package Power Limit: “150”
Short Duration Package Power Limit: “253”

[Silent_Scone · Accepted solution]

I’d consider RMA’ing the CPU.

[UserX]

Thank you. I'm heading that direction. I wanted to make sure that I covered all the troubleshooting steps before requesting an RMA.

[Jevelous]

Greetings. Over time, this problem appeared on my 13700kf after updating the BIOS (blue screens, crc errors, insta crash in Cyberpunk on game start). For some reason, the Asus AI Overclock stopped working stably and began to increase the core multiplier (x60, instead of x58). On stock and via Intel XTU everything works stably.

[UserX]

Interesting. I have not noticed this behavior, but I will be on the lookout.

[Shinchan0125]

my system was working fine for months.  Until one day, it crash, bsod...etc   but it turns out iCue 5.8 was the issue.  Do you have iCue install?

[UserX]

Originally, when the issues started, I had iCUE 5.8. I never installed it on the fresh installs of Windows for troubleshooting purposes. I should have mentioned this in the troubleshooting steps. I'll add it in.

I had a base install of Windows and its updates with no preview or insider updates. I added the following software to aid in troubleshooting.

• HWInfo
• AIDA64
• WhoCrashed
• WInRAR
• 7-ZIP
• ROG-GPUZ
• Intel Processor Diagnostic Tool
• Intel XTU
• Prime95
• Cinebench R23
• Sysinternals toolset
• And a few other NirSoft tools for diagnostics that were never needed.

After I changed the SVID behavior, Specific Core Ratio and running stable for more than 24hours, I installed Fury CTRL, iCUE and the Logitech GHUB to extend the testing. 

I should uninstall the add-on software and continue testing the base system. Thank you for the advice.