Irregular crashes, highly volatile machine (13900K + RTX 4090)
Irregular crashes, highly volatile machine (13900K + RTX 4090)
Hey there, welcome to the community. This is my debut here, so please be kind – I’m really struggling to come up with solutions. Over the last few months, my computer has been acting up a lot. It’s become increasingly unstable, and things have only gotten worse lately. I’m experiencing random BSODs without any clear cause. Sometimes it freezes during games or when under heavy use, other times it just shuts down completely. It can run smoothly for hours before suddenly crashing. The BSOD messages I’ve seen so far include KMODE_EXCEPTION_NOT_HANDLED and PAGE_FAULT_IN_NONPAGED_AREA, plus a few others I missed earlier.
My setup looks like this: Intel Core i9 13900K processor, MSI RTX 4090 graphics card, ASUS ROG STRIX Z790-F motherboard, gaming PC with Wi-Fi, G.Skill Trident Z5 RAM (2x16GB), Dark Rock Pro 4 cooler, be quiet! Dark Power 13 1000W PSU, and storage devices Samsung 980 Pro 2TB plus 990 Pro 2TB M.2 drives.
I’m aware that the 13th and 14th generation Intel CPUs have had stability problems before. I did encounter some issues early on, but after updating the UEFI firmware, things improved temporarily. Now, however, the crashes are back and more severe than before. Yesterday it happened again – the system BSOD appeared, and Windows won’t boot properly. I only see a black screen with a loading spinner. Even after reinstalling Windows, the installer BSOD occurred at one point. Eventually I managed to reinstall, but it crashed again while idle, leaving me back on that black screen.
So far, I’ve done things like running Memtest86, checked my SSDs with Samsung Magician, applied several UEFI updates, and tried a clean Windows install multiple times. But I can’t share any diagnostic files right now because the system won’t boot at all – I can’t even reach the desktop.
Right now, I’m unsure whether the problem is hardware-related (like the motherboard, CPU, or memory controller) or something else. The unpredictability of these crashes makes it tough to figure out what’s wrong. Do any of you have faced similar issues? Any tips on troubleshooting, replacing parts, or testing would be really helpful. I’m open to suggestions at this stage. Thanks a lot!
It seems like the system might be struggling with its processor. Did you install the latest BIOS right after purchasing the computer? How long did you operate it without the most recent microcode updates? It could also be that you received a pre-worn CPU from a previous buyer. Some sellers who picked up a used, worn-out chip might avoid the hassle of an RMA with Intel and instead try to resell it as brand new. If the unit came in almost original packaging or was placed in a ready-made machine, they’ll likely attempt a fresh sale. You can help by asking someone with a compatible CPU to check it out. Test both your PC and theirs to see if the issue lies with your own hardware. If you confirm it’s the CPU, consider using the warranty. Alternatively, you might upgrade to a more recent model.
Hey, thank you for getting back so quickly! I’m beginning to think the problem might be with the CPU itself—those unpredictable BSODs, the growing instability, and the fact that reinstalling Windows isn’t fixing it. It seems like something more serious is at play. I bought this CPU in February 2023, and I recall the BIOS/UEFI updates for 13th generation issues only became available in September or October. I installed them right away when they were released, and I’ve kept the BIOS updated ever since. Back then, there wasn’t much awareness about long-term degradation risks for 13th/14th gen CPUs. For months, I ran the system using the older, possibly unstable BIOS settings without realizing it. Looking back, that could have caused damage. I also purchased the CPU from a reputable store, and the box looked brand new—hopefully it wasn’t used before. You’re right; sometimes components get repackaged without people knowing. Right now, I don’t have another compatible CPU to test. Do you know if it’s still possible to file an RMA directly with Intel as a regular user?
Reach out to the seller first. If that’s not possible, contact Intel directly. Either way, you own an ASUS board, and even a lower-end model from them can be problematic. These newer high-tier products push performance limits more strictly and often receive slower BIOS updates due to reduced sales volume. The likelihood your CPU is affected is almost certain. Positive news: if the RMA fails or any other issue blocks you from getting a replacement, i5 chips are still acceptable, and for gaming the difference is minimal.
Thanks for your thoughtful response. I’m glad you found the information useful! I plan to contact the retailer first, using the invoice as proof. If that doesn’t work, I’ll follow Intel’s return policy. I also have an active support case with ASUS in mind, just in case the issue lies with the board itself.
I didn’t realize the Z790-F was seen as a less stable option despite being part of ASUS’s premium lineup. I expected better reliability given its reputation.
For future reference, if you’re considering similar boards, safer choices for stability tend to be those with proven long-term support and consistent performance. Your setup is well thought out for both gaming and development tasks—great pick for the 13900K!
Let me know what I can do to help further.
ASUS ROG products often push more chips out of the box than usual, which isn't typical. However, due to a manufacturing defect, the ROG lineup was impacted more severely compared to other high-end models. Limited production boards usually experience minor delays or serve as early testing grounds for updates that are just standard. The board you received is of good quality; the real issue lies with Intel's oversight. It seems the boardmakers weren't significantly responsible.
Memory testers struggle with DDR4 and newer types. They often overlook faulty RAM, making them unreliable (no false alarms). I favor running the PC normally with one stick at a time, without XMP settings. CDI performs self-checks on drives (SMART), but this feature has been significantly reduced for NVMe devices. There are no dependable tools for testing NVMe drives. The only useful indicator is Critical Warning. Among the many faulty NVMe drives I've seen, fewer than five displayed any signs of failure in SMART. Check C:\Windows\Minidump for any minidump files. If present, return to the Windows folder and move the Minidump folder to the Downloads directory (use your desktop if you don’t have OneDrive syncing). Compress the folder and attach it to a post. Please follow instructions carefully—Windows doesn’t like changes in this location. The main issue here is the CPU. As a 13th generation chip, it likely operated well for a long time without the voltage correction, possibly causing permanent damage. When failures occur from this, it often appears as memory-related problems.
I successfully reinstalled Windows again. It started up properly but stopped after roughly five minutes due to a PAGE_FAULT_IN_NONPAGED_AREA error. I followed your guidance and managed to capture the minidump file – I’ve compressed it and included it here. Also, do you have any knowledge of DDR5-related utilities or more dependable methods for testing memory on newer systems? It would help to confirm the issue is fully resolved. EDIT: Experienced additional crashes while away from my computer. Please update the attached ZIP file. Minidump.zip
All standard memory testers have been updated for DDR4 and DDR5 support, but they still struggle with them. Run the PC as usual using one stick at a time. Performance varies based on how often crashes occur—until it stops working or you notice it should fail. The issue seems to stem from memory found in the dump files. While memory can refer to storage, it’s usually the main concern here. Windows moves low-priority RAM data into the page file and retrieves it when needed, making storage appear as memory. The memory controller is built into the CPU; failure there will cause a memory-like behavior. Additionally, the Intel problem adds complexity. Three dump files aren’t definitive proof of storage failure, especially since one involved storage drivers, but it likely isn’t the case.