Right so my rig CAN be stable at 3.2Ghz, and even with around 1.10v Core, but the VTT has to be something like 1.29v for the proc to be stable. After reading both Anand's post on high VTT and reading the Q9450's electrical specs, it seems that I am operating outside of what Intel calls "functional operation condition limits" with my VTT voltage, that is that my VTT is above 1.15v, and in an area where the lifetime may be degraded.
I know, I know, welcome to overclocking, may I answer your FAQ's?
Nevertheless, I ask what size grain of salt should Intel's own proc specifications be taken with. I plan to upgrade to Nehalem around spring or summer next year, but current parts are going to a family member and I hate to simply throw them a system on crutches.
The NB voltage is 1.29v as well, the PLL is at 1.508v, SB is 1.508v and 1.101v. Load-line is enabled, otherwise my load voltage, OC'd or not, rises to ~1.22v which I believe isn't what should happen.
As a side note, my mobo has an LCD Poster I use for monitoring voltages, I also use CPUID's HWMonitor and CPU-Z. All of them are typically within .01v of each other, the Poster is always lowest, but at times CPU-Z reports a Vcore of 2.040v! Temps don't rise per HWMonitor, nor does anything explode or my power consumption rise per my Kill-A-Watt. Is this a known problem with CPU-Z or am I slaughtering my CPU? The jump is sporadic and occurs seemingly at random times, at load or at idle.