In a situation like that I'd try to configure the memory more slowly, if the BIOS Setup allows it; or if you have four modules temporary remove two. That might not be a solution, but if it helps then you at least know what causes the trouble.
@kernellogger I have doubts it's really the memory tbh, I see from prior logs it has happened in the past (these mce reports) but it didn't segfault things like it has now...
I am seeing sudden temperature spikes when compiling with all 32 cores (i9-14900k) but no such error when stress-ng'ing.
BTW I have 192 GB of RAM (don't ask me why I can't answer) which is the max the mobo can support so possibly a cause of weirdness... :>)
@kernellogger so if I built with fewer jobs permitted, everything's fine. So this surely is not the memory that's the problem?
Seeing thermals go crazy.
Smells like mobo to me, I wonder if recent kernel (I am on 6.8.2) has made it more possible to hit this somehow? I know the i9-14900k has had some issues
Less jobs, less memory pressure. I thus would not be so sure.
But yes, you raise valid points why it might not be the memory. But from my experience in a hardware lab (which was many years ago, so things definitely have changed!) I'd still give it a try, as limiting the memory speed or ripping out a module is something that can be done in a few minutes.
Add comment