06-23-2023 05:53 PM
- 7900x, Kingston renegade 6000 c32 2x16GB (kf5600c32rwk2-32)
default setting, XPM1,2 setting all bsod while idle load in windows, errors are random memory controller, IRQ less... etc
I even testing with one module and same issue, bio is newest v1616
I know that memory is not listed in asus support but it should work defo on default settings at least, cpu, movo and ram are new, wuite hard to find where is the issue
memory test passed for 2h while using XPM 1 setting (strange) but windows crash mostly at idle load
Solved! Go to Solution.
06-27-2023 08:18 AM
After days of testng locking CPU at 3600Mhz it worked for 24h stable, then presumed its CPU quality issue and I left all stock + memory at XPM1 and I adjusted curve optimization to overvolt the CPU to +10, 12h and testing still ongoing
So yeah defo AMD has binning and quality testing of their CPU, I am not sure how to RMA as I refuse to stay without CPU for days and the CPU is bought in other state, will send a email to AMD as I expect they resolve it for me, as this is not acceptable we have many DAO or unstable CPUs at all!
06-23-2023 09:20 PM
Hello,
It sounds like the CPU isn’t stable at stock. Have you contacted AMD support?
06-24-2023 02:22 AM
I have not, I never seen bad CPU, also i can not be sure its not mainboard issue, after loading bios default it stay stable longer 5-6h, and while editing some basic settings as qfan and uefi/csm i had bsod in few minutes
I could test if I had spare parts for AM5 but i do not
06-24-2023 08:41 AM
I tested CPU at 48x and it still BSOD after few hours
Now I am testing cpu at default + XPM1 on memory + MEMORY CONTEXT RESTORE -> disable, 3h no bsod
also I left CPU fan at default as it only keep VRM under 45C on high rpm as I always tweak it on each build to be silent, but I have no idea why would mobo with good VRM at ~60C be a issue even at idle load
Also last bsod i had I monitored WHEA and I had CPU cache L0 errors while the cpu was at 4800Mhz up limit lock
06-24-2023 11:47 AM - edited 06-24-2023 11:48 AM
Likely less to do with the [VRM] temperature and more to do with the CPU being unstable at the applied setting. Report back if still having trouble 👍
06-24-2023 12:04 PM
Its harder to test all setting as more I test bsod happen less, at first 5-10min now its 4-6 hours, still need to test at some mediocre CPU setting as limiting CPU to 3600Mhz and even try to overvolt it to be sure if the cpu is guilty one
I still bet its mostly mobo fault as shop where I bought has 5% fail rate for it
06-24-2023 10:52 PM
You'd need to troubleshoot or take to a system integrator. Looking at various Reddit cases, replacing the CPU resolved the issue.
06-27-2023 08:18 AM
After days of testng locking CPU at 3600Mhz it worked for 24h stable, then presumed its CPU quality issue and I left all stock + memory at XPM1 and I adjusted curve optimization to overvolt the CPU to +10, 12h and testing still ongoing
So yeah defo AMD has binning and quality testing of their CPU, I am not sure how to RMA as I refuse to stay without CPU for days and the CPU is bought in other state, will send a email to AMD as I expect they resolve it for me, as this is not acceptable we have many DAO or unstable CPUs at all!
07-27-2023 08:12 AM
Hi. Another user had a lot of problem with 1616.... https://rog-forum.asus.com/t5/amd-600-series/tuf-gaming-b650-wifi-latest-bios-1616-makes-windows-cra...
08-09-2023 02:43 AM
I'm testing the last BIOS version , 1636 NOT beta, anch with a little overvolt on DDR (1.4V) it works with EXPO I @6000