Small update... after the server running for weeks, I randomly checked the performance chart for the host and found ~35% CPU usage with VMs doing nothing. It wasn't locked to one core in particular but distributed across all 4/8 (core/threads) to equal that ~35%. It appears there is another IMPI issue. Have not yet moved to ESXi 6 yet. Disabling IPMI in ESXi again for now. It did shut down properly during this condition, but I'd rather not waste that much power/heat for no real reason, considering I log in directly to the motherboard vs. relying on VMware to report on hardware health.
The only steps I took when setting up the bios were to disable all unused ports, make sure VT/d was enabled, disable option rom for nics & cards, set boot order, fan speeds, etc. Is there any specific page or settings you were unsure of? I can check them.