Let me see if I understand this:
1. User has combined an FX-8150, one of the most power hungry CPUs money can buy, with a M5A78L, a motherboard with one of the smallest VRM configurations possible for an AM3+ system.
2. A change is made to the CPU HSF. This change has a significant impact on the air-flow characteristics over the VRMs and region surrounding the CPU socket (where lots of current flows).
3. User reports sudden change in CPU behavior, Now throttling.
4. Recommendations to turn off power saving features which have nothing to do with the problem are made.
5. Recommendations to solve the CPU scheduling issue in windows are made, even though this has nothing to do with the problem.
6. The "best answer" is selected, the software "fix" that solves a totally different problem is now the "best answer" in this thread.
7. Meanwhile, the problem of how to deal with the throttling caused by VRM temps has gone unrecognized, and unsolved.
Kudos to Cemerian for answering correctly. We've seen this issue hundreds of times over at OverClock.net. It's a common problem when users switch to tower style HSFs, and even worse, when they switch to AIO CLC. The solutions are simple: Upgrade the motherboard or make cooling mods to the existing motherboard.
----------------
DazKins,
I'd be very surprised to learn that your problem was actually solved by the scheduler hotfixes. Sounds to me more like a coincidence of events. Did you happen to disable cool-n-quiet at some point along the way? Perhaps the same time you installed those hot fixes? Many boards actually tap into the low power states made available by cool-n-quiet as part of a self preservation measure when running too hot. Disabling cool-n-quiet *may* have effectively disabled the boards access to low power states, thus preventing the self preservation function. In any case, I suspect the problem is more or less still there, whether you are experiencing any current effects of it or not. An overclocked FX-8150 can a lot of power and potentially damage your motherboard if you aren't careful.
The socket temp of your motherboard is effected by the VRM temps, this is because the VRMs are close to the socket and are cooled by the board itself. When the socket temp readings hit ~70C or so on some boards, the board will automatically start forcing low power states to prevent the board from over-heating. This method of self preservation has the undesirable side effect of causing major performance glitches because it just immediately drops the CPU to the lowest power state (typically 1.4ghz and ~0.9V for BD/PD) for several seconds in order to "recover" thermal dissipation headroom. I can see that your motherboard has the common IT8728 SuperIO chip on it, which has inputs for 3 temp sensors. With any luck, hopefully your socket temps sensor is wired to the SuperIO chip, because then you can see the temps as well. The "K10temp" core temp sensor is not the sensor you should be interested in to figure out this problem (that's usually the one associated with "core" or "package" temps on this platform, a very precise but very inaccurate reading). One of the TMPIN's on your SuperIO chip will hopefully have an answer for you. HWmonitor and other hardware monitoring tools should allow you to probe and watch the temp of that sensor if it is there. See if the throttling isn't happening exactly when that socket temp hits a particular temp.