Is there a change in the BSOD errors after a new build?
Is there a change in the BSOD errors after a new build?
Built this system on June 1st and been trying to diagnose BSODs ever since. Started soon after the 1st boot. Initially the BSOD was related to Hypervisor errors and IRQL not less or equal. After a little research I disabled SVM in the bios and it seemed to help a little, but the BSODs continued randomly or when trying to game. I don't game too often and my gaming is only a few games like Madden 25, Forza 5, and Worms WMD (10 year old game). It's crashed on every game at some point.
Anyhow, everything I've researched pointed to drivers, software, or ram. The bios, chipset, drivers, and software has all been updated directly from manufactures websites. I've ran OCCT on the CPU, GPU, and memory and it passed the tests. I ran MemTest 86 for like 6 hours and it passed. I was recommended on the MSI forums to run TestMem5 on 1usmus_v3 profile. I've ran the test with both DIMMs and single DIMMs in slots A2 and B2. Ran the tests at JEDEC, EXPO1, and EXPO2. TestMem5 failed in 5-10 minutes in all scenarios.. Sometimes it would throw error 10 or 20. Sometimes it would just BSOD and reboot. Then I was recommended to try a different brand of memory, so I ordered a G.Skill kit and I got basically the same result on that kit when I ran TestMem5. So I'm going to assume that it's not a memory problem.
I've reloaded Windows 11 Pro/Home 5-6 times while paying special attention to the drivers and order of installation. Everything is updated to current releases. So I'm kind of ruling out the software end for now.
So far over the past 5 weeks the BSOD errors I've gotten are Hypervisor errors, IRQL not less or equal, Kernel, Ntoskrnl, System service exception, Page fault in nonpaged area, Critical process died, and Watchdog violation. There may be more that I'm not remembering.
I've remounted the CPU and cooler twice. Inspected the socket pins and DIMM slot pins and they all look fine as far as I can tell. I even disassembled the system and rebuilt it, but yet they system won't get past 5-10 minutes in MemTest5. While waiting for the 2nd kit of ram, for entertainment purposes I pulled out my previous i5 10400 system and ran the same tests on it and there were no issues.
Where to go from here, I'm at a dead end? At this point the 5070 is removed and I'm running on the iGPU with minimal software on a fresh Windows install. I'm thinking it's a CPU or motherboard issue. All advice or recommendations to help pinpoint the issues are greatly appreciated
Specs:
Ryzen9600x
MSI B850 Tomahawk Max WiFi
Samsung 990 EVO NVMe 2tb
Nvidia 5070 FE
Teamgroup T-Create Expert 2x32 DDR5 6000 / 34-44-44-84 QVL
(Replacement memory: G.Skill Flare X5 2x32 DDR5 6000 / 36/36/36/96 QVL)
Corsair RM750e ATX 3.1 (2025)
Corsair Frame 4000 D
Windows 11 Pro/Home
The BIOS, chipset, drivers, and software have all been updated directly from the manufacturers' websites. For relevance, what BIOS version are you using for your motherboard? I've reloaded Windows 11 Pro/Home multiple times while carefully checking the drivers and installation order. Everything is up to date, so I’m not focusing on the software side just yet.
Do you mean you recreated a bootable USB installer, reinstalled the OS in offline mode, then installed all necessary drivers with the latest versions using an elevated command while offline? Or did you reset the operating system?
Samsung 990 EVO NVMe 2tb
Is this the sole storage drive in your system?
Following the early BSODs, I upgraded from the default BIOS to the 7E62v2A30 released on 5/27. Next, I upgraded to a version they released on 6/11, which they later removed from their site for some reason. Then they issued a new release of 7E62v2A40 on 7/7, which I installed right away.
I built a bootable USB using Rufus, reinstalled Windows 11 without internet, turned off the iGPU in the BIOS, removed AMD graphics drivers via DDU, and added chipset and LAN drivers directly from manufacturer sites. After that, I installed Nvidia drivers with NVCleanstall. For troubleshooting, during several Windows reinstalls I stripped out the 5070 and relied solely on the iGPU.
Yes, the only drive remains the Samsung one. I ran the Samsung Magician diagnostic scan and CrystalDiskMark64, confirming the drive is functioning properly. Additionally, yesterday I reran OCCT on CPU & Memory, Linpack, and it finished without errors. I’m unsure why MemTest86 passes 6+ hours but fails TestMem5 within ten minutes.
Yeah, that's something I've thought about too. Either that or the board issue. It seems to work well most of the time, except sometimes when the fans get really loud even when idling. But it's still running stock BIOS fans. When I set up a custom profile in FanControl, it helped a bit. I've heard that AMD can get a little hot sometimes, but OCCT tests show the temperatures were 70°C or lower.
Regarding the BIOS, I'm using the latest version released on 7/7. I've kept updating every time they release something new. I believe I've updated three times: from the stock BIOS to the 5/27 release, then to the 6/11 release (which they later pulled), and finally to the 7/7 release.
For the sake of it, I tried another set of RAM and got the same outcomes. It seems at this stage I can confidently eliminate the memory entirely. I've come across some information suggesting that C-State control or memory context restore might lead to problems on certain systems. All BIOS settings are in their default state except for the MSI center download I turned off. What do you think?
Have you reset the CMOS after the BIOS updates? This is recommended following a BIOS update. If the problem persists with the default BIOS configuration, then I would promptly initiate an RMA for the CPU. If the upgraded CPU still encounters the same issues, I would proceed with an RMA for the motherboard.
Yes, I attempted that following the latest BIOS update. I also re-ran all the stress tests once more. OCCT (CPU+Mem, Mem, Linpack), AIDA, and MemTest86 were executed; all full tests completed without errors. Despite this, I still experienced BSODs and reboots after 7 to 10 minutes in TestMem 5. Since I'm leaving town next week for a month, I plan to return from vacation around the end of August to handle the RMA for the 9600x. I believe the CPU and memory have a 36-month warranty, so problems shouldn't arise there. This is the first build in 25 years that required an RMA, probably because the odds came together this time. LOL...