Is the PC experiencing a crash due to a BSOD?
Is the PC experiencing a crash due to a BSOD?
I don't know what to do anymore.
System details:
- Graphics card: ASUS TUF GeForce RTX 4070 OC
- Processor: AMD Ryzen 7 5800x
- Motherboard: Gigabtye B550 AORUS ELITE
- RAM: G.Skill Trident Z RGB 16GB 4000Mhz
- Storage: Gigabtye Aorus 1TB, Samsung 870 Evo 1TB
- Cooler: Arctic Liquid Freezer II 240mm
- Powersupply: Corsair RM650x
Everything is functioning properly, temperatures are normal, and stress tests were conducted on all components. I’d like to mention that based on the notes I’ve recorded, I haven’t encountered a consistent pattern of crashes. Each situation seemed unique.
Before proceeding, I want to briefly review past problems. About a year ago, after buying the PC (which was pre-owned at that time, with a Gigabyte eagle 3070), I noticed my RAM wasn’t operating at the advertised 4000Mhz—it was stuck in stock mode with XMP disabled. I adjusted the BIOS settings and turned XMP back on, but the instability persisted. At first it was a Blue screen of death, not a full crash.
Over time, the system became more stable under load, but crashes still occurred. I spent hours troubleshooting: reseating RAM, resetting BIOS (including removing the battery), reinstalling Windows multiple times (even tried Windows 10, but I can’t recall the version), changing BIOS settings, swapping RAM and graphics card. Nothing resolved the issue. Eventually, after a few months, Diablo 4 was released. I played it briefly before a sudden black screen appeared. Later, the crashes became random—sometimes lasting hours, other times just minutes. Despite retrying all troubleshooting steps and updating BIOS and graphics drivers, the problem persisted.
After a while, I changed my graphics card to an older GTX model and reduced RAM to 1x8GB. This eliminated crashes, so I suspected the graphics card might be faulty. I then replaced it with a new RTX 4070, which performed well and resolved the issue. However, after several days of troubleshooting, the problem disappeared on its own. I then re-enabled XMP and set it to 3600Mhz, and it worked again.
Now I’m left wondering if the issue lies with my power supply (which seems to be functioning) or a faulty main drive. I’ve tested the drive and confirmed it runs at advertised speed, with no errors. Some potential fixes I’m considering include reinstalling Windows 10 on a different drive (to rule out a hardware bug) or changing the main SSD if there was a crash there.
I’m seeking advice and hoping to find a solution.
Identify precisely which DIMM slots house your memory modules, beginning at the CPU socket and moving toward the motherboard edge. Listing them in order from closest to the edge—1, 2, 3, 4. Provide the exact model of your memory kit. If uncertain, run the following command in an administrator command prompt to retrieve details: wmic MEMORYCHIP get BankLabel, DeviceLocator, MemoryType, TypeDetail, Capacity, Speed. Additionally, confirm the exact BIOS version running on your motherboard, not just a general description.
My first thought is that a 650w PSU might not be enough for either of the GPUs you've used, especially considering their power needs based on age and peak demand. The 4000mghz RAM wouldn't let the CPU's memory controller match the memory clock, infinity fabric clock, and controller clocks unless you have a special chip like the 5800x. Faster kits with XMP settings can run between 1.45v and 1.55v, which might cause overheating and instability. I would set the RAM to default JDEC speeds by removing it from a DOCP/XMP profile, which should likely resolve the problem.
Are you using a PCIe riser cable?
Do you have the GPU in the top x16 slot, and if yes, is it configured for PCIe 4.0 in the BIOS?
Have you attempted to uninstall the Nvidia driver using DDU and install the latest version manually without an internet connection?
I don't recall seeing that reference anywhere. And while you mentioned doing it, as "I" already stated, they had attempted that too and still faced problems. If they continue to encounter issues with the default JEDEC setup, it's probably not the main concern. More likely, the issue lies elsewhere—such as incorrect population data, an incompatible kit, possibly a misaligned CPU cooler, bent pins, or even an outdated BIOS version.
Not disputing your point, just noting it wasn't explicitly mentioned before.
I’m about to address your questions directly, starting with the recent updates I’ve made. As I mentioned earlier, I plan to replace the main drive and Windows 11 by installing Windows 10 on my secondary storage, hoping it will trigger a crash (though I haven’t done that yet). Both drives have been cleaned, the same edition of Windows 10 installed, and a game has been added. Currently, the PC fails whenever heavy tasks are launched, especially after using the Gigabyte Nvme SSD. It crashed repeatedly when testing CrystalDiskMark, even after changing the slot to rule out a faulty one. In contrast, my Samsung SATA SSD performed flawlessly under all conditions, except when attempting to run a game from the Nvme drive. This suggests the main drive might be failing, which could explain the erratic crashes and eventual failure after some time. The only oddity is that read/write operations work normally, except for the post-CrystalDiskMark crash. The SMART data from the Nvme controller also supports this theory—it appears to be malfunctioning, which is unusual given the PC’s previous condition. This information could help clarify what’s happening. One additional point is that my Samsung drive had a longer lifespan than the Nvme one, which might be relevant.
First, I would upgrade to F17b. It has been over two months since that version was made available, so if there were any issues with it being a beta BIOS, it should have been removed by now. I’m surprised it hasn’t been reissued with a stable BIOS version already.
Next, I’d perform a hard reset of the BIOS. After that, adjust any custom configurations you need—specifically enabling XMP—but only after activating XMP and restarting, then immediately return to the BIOS and set the memory frequency to 3600MT/s. This is because attempting to run it at 4000MT/s would incur significant penalties due to the infinity fabric. After saving your changes and exiting, observe the performance.
From a compatibility standpoint, my memory kit isn’t listed as supported on the G.Skill website or in the QVL directory for my motherboard. This could be the root cause, not because it’s the correct type of memory (like DDR4 or DDR5), nor because of the frequency itself, especially if you manually set it to 3600MT/s. It’s more likely related to the specific board, CPU pairing, or timing settings that don’t align.
For clarity, memory is typically described in MT/s, not MHz, and the board’s configuration plays a bigger role than the memory package alone.
Kit isn't officially compatible. It doesn't appear in the G.Skill memory configurator's compatibility list, which is a significant issue. Even though there are similar models available, a minor difference in the memory type can completely prevent successful installation on the board. And even with identical models, compatibility can still be affected if manufacturers alter the kit's components over time. Sometimes, a memory kit that worked previously may not function after changes in the ICs or other modifications made by the manufacturer. This is why it's important to verify compatibility before purchase.
Thank you for your assistance and the information you provided. I will proceed with a bios update and follow all your advice. Before doing that, could you confirm whether you believe the drive is faulty based on the most recent testing and SMART controller data? Should I also repeat the same test with both SSDs after updating the bios? Thanks again.