This also happened with my previous DZ68DB, and still happens with my DZ77RE-75K to a certain extent (except the latter does end up turning off after a while of the weird behavior).
I'd agree that it seems to be an Intel thing. I wonder if the "flash update sleep delay" setting has anything to do with it...
EDIT: Nope, it doesn't, but I'm going to keep trying things in the POWER tab.
EDIT 2: Still nothing. I do want to add that when I first switched my motherboard from the DZ68DB to the DZ77RE, it DID shut down correctly, but only ONCE, and it was on the "maiden run" of the Z77. So I do think it's entirely possible to fix this issue, and now I'm wondering if it has something to do with Sandy Bridge. Is anyone running an Ivy Bridge CPU in either a 6 or 7 series intel board and having this trouble?
EDIT 3: I GOT IT TO SHUT DOWN CORRECTLY! I noticed that, when I needed to use the rBoot CD, it would actually work! So now I've narrowed it down to either using PCIRootUID=1 or the MacPro3,1 smbios. I'll edit in my results once I get the chance to figure out what's behind this.
EDIT 4: It seems that it's related to PCIRootUID. Setting it to 1 breaks the Integrated Graphics, as well as causes (at least in my case) the graphicsenabler to load the generic RadeonFrameBuffer. However, it shuts down properly... Sort of (if a screen is plugged into the integrated graphics, it seems like the shutdown turns off the discrete card first, has a delay, then turns off the integrated GPU while simultaneously shutting down). Setting PCIRootUID to 0 fixes the graphics issues, but causes a 30 second delay between "GPU off" and "CPU off."
For people having this issue, at the moment I'd say try booting from rBoot and seeing if it shuts down, then proceeding from there. I'm going to keep looking into this, though.