Given the page that is corrupted is on the zero page list, this isn't a memory range that would be referenced to a running program or driver (you'd get a very different bugcheck if any driver or other system code tried to write to one of these pages directly). That really does only leave hardware failure, either the RAM or the motherboard (or CPU) memory controller:
Code:Probably caused by : hardware_ram ( ZEROED_PAGE_CORRUPTED )
Understood, but a corrupt page on the zero page list could very easily indicate bit flipping by the CPU. As far as I can tell, memory testing software only tests whether or not the machine's RAM can read and write memory without error, and doesn't really tax the memory controller itself to verify it's not at fault. It's never totally out of the question for a driver to cause it, but on the zero list it doesn't seem remotely likely (this list contains memory that has been freed and it's contents zero'ed, then put on the zero page list for further use - it should all be zeros, and none of the pages should be associated with anything at all, so a bugcheck saying one is corrupted is almost always hardware, as no software should be touching these!).