Semi-random 0x124 BSODs, more frequent under load

[friqgvfkhg]

Ever since I got my PC - in October '13 - I have been experiencing semi-random BSODs.
The majority of them have STOP code 0x00000124 (uncorrectable hardware error), and are caused by either ntoskrnl or hal.dll (although there are exceptions to both of these observations).

Here's a link to an overview of the minidumps, as seen in BlueScreenViewer: http://i.imgur.com/mCcflQu.png

So far, the main two factors affecting the frequency of crashes have been uptime and load. For example, running a minecraft server 24/7 - with loads hovering around 40-60% most of the time - would result in a BSOD every couple days. Shutting down the server for the night, however, would let me run without problems for perhaps a week or so.
Running prime95, with the CPU load therefore being at a constant 100%, usually results in a crash after 10 hours, although that can vary quite a bit.

In contrast, when my PC is idle or used for web browsing, without any significant stress, I can usually go 2-3 weeks without BSODing.

I did run the SF diagnostics utility thing, here's the zip file with all the complete minidumps, as well as the sys info dumps and stuff: View attachment 331399

My personal, perhaps very incorrect, opinion is that since the majority of the BSODs return the 0x124 hardware error code, and since the crash is usually cause by either hal.dll or ntoskrnl.exe, two very low-level processes, it is highly likely that something in my hardware is indeed faulty.

It may be worth mentioning that the CPU itself is very unlikely to be the cause of the problem; prime95 never reported any errors during torture testing, and my temperatures even after multiple hours at constant 100% load usually hover around 50*C. I will run memtest86 during the night and will post the results in the morning to rule out the possibility of my RAM being faulty.

 

[SGT Oddball]

0x124's are a pretty generic hardware fault that is pretty hard to pin down, try working through the trouble shooting stuff here - http://www.sevenforums.com/crash-lockup-debug-how/274565-stop-0x124-troubleshooting-new-method.html

 

[friqgvfkhg]

Well, from what I gather, since the first argument is 0x0, I'm getting a generic CPU machine check exception that doesn't really give any info about what exactly failed.

Am I right in assuming that if it was a PCI(e) problem, the first argument would necessarily be 0x4? Or is it possible in some cases for a PCIe-caused problem to still give the generic 0x0 argument?

Seems like it's back to stress testing for me, and maybe eventually replacing my mobo. :/