Did you go into BIOS and check the voltages at all? To my knowledge the core scheduler is always in the CPU. Not the Northbridge if that info was sent over the FSB that would be kinda slow. Anywho i would have checked voltages and such first on the northbridge and CPU, could be old BIOS too.
EDIT: And my board says its rated for 1866mhz memory, oh well its a free board i got from a junk computer which a PSU blew up in so that may be why it craps out anything over 1333mhz
Voltages are "fine"; the oddity being that the VCore can't be changed (supposed to be able to change it on the 990FXA-UD7), and the BIOS simply cannot be updated (system crashes on boot and instantly reverts to F10 BIOS version). All signs point to it being a north bridge/memory controller voltage issue but we'll see if Gigabyte gives me a detailed report whenever the hell they're done.
Here's the thing: were it the CPU, I would've been able to get another CPU and the problem would have vanished. However, that's not the case, as I've tried
four different FX-8350s, and the same problem occurs. My memory tests absolutely fine in all stress testing programs, too. The easiest way to recreate the problem is doing a CPU test with Prime95: with one core selected, the test runs flawlessly; with two cores, it runs for a few seconds and then freezes; with more cores than that, it freezes the second the additional cores initialise.
I'm hoping that in due time, this problem will be nothing but a memory.