issues with asrock 990fx killer

In that case trying the FX-4300 in your rig seems worth the trouble. I would also suggest running MemTest86, as I would expect it to produce errors and the error log may help pinpoint the source of those errors.

Unfortunately I don't know any more of that error message than you do. :/ I share your belief that the RAM modules should be fine, but the fact that you're using four of them puts extra stress on you CPU-NB. This is why stock voltages may not be enough and bumping the voltage up could help. I suggested bumping the RAM voltage up first only because that's what helped the other guy and it doesn't make the CPU run any hotter. Increasing the CPU-NB does, so you need to monitor the temperatures when or if you decide to do it.

Okay, that's just due to us having different motherboards and thus different heat sensors. The names don't matter as long as you know what each temperature means.

Those temps do seem a bit high. For gaming 140F should be okay, but it's still over the often recommended 130F maximum temperature under stress testing. Even though an FX CPU may be able to take higher temperatures than older Phenom II's, your system might still not be fully stable under Prime95. Also, at idle I would've expected something around 85F, especially since there should be no issues with the case cooling. Maybe the TIM between the CPU and the heatsink is getting old or has not spread properly...

Btw, the motherboard temperature and socket temperature are two different things. The socket temperature is the CPUTIN temperature in HWMonitor and for that the often recommended maximum temperature is 158F, but again the FX CPU might be able to stay stable even at 185F. See here, for example:

http://www.overclockers.com/forums/showthread.php/728489-FX-its-tougher-than-you-think

While your temperatures are a bit higher than what I would have expected, I don't think they're what causing the crashes. At idle even the VRM should be running cool enough. If you know of a way to make the system crash (like running Aida's stress test), you could try bumping the CPU-NB voltage slightly and see if that has an effect on stability. Also bumping the RAM voltage slightly might be worth a try, it seemed to help for someone experiencing similar issues (it might be worth reading the whole thread):

http://www.xtremesystems.org/forums/showthread.php?284441-strange-stability-issues-with-FX-8350

I'd probably start with bumping the RAM to 1.55V and then 1.6V if that doesn't help. If that's not enough, then I'd try 1.2V for the CPU-NB and 1.55V for RAM, and then again increasing RAM voltage to 1.6V if it still doesn't work.

That may be a good way to isolate the issue. I'm not surprised you had stability issues running the CPU-NB at those speeds on stock voltages, but I've no idea if and how that could actually damage the IMC. Increasing the clock speed without changing the voltage doesn't increase the head load that much and since the CPU-NB is located on the die, your CPU cooler should have been able to handle it.

Have I understood correctly that your system occasionally crashes even when you're not doing anything on it? You're not running Seti@home, Folding@home or anything similar? If the system crashes when it's basically idle, that's not exactly a good sign. Your temperatures and voltages are still a question mark. Have you checked both using HWMonitor? How high are the CPU and socket temps at idle and under load? What sort of case cooling do you have?