Rogue Leader :
Power phases have nothing to do with temperatures. it has to do with providing the CPU the stable voltage it needs for an overclock. The more power phases you have (and yes you should keep them cool) the more voltage your CPU has access to. More voltage, means you can clock it higher (sometimes). Now that said there is always a chance of course that your CPU just can't take any more voltage, but with more power phases you will find that edge further up. You need to of course work under the assumption that your CPU will be able to handle it, but during the OC process of course test and tune to what its actually capable of.
There are no guarantees here, however more power phases can give you more of a chance of a higher OC.
Ok, that's really helpful.
So I understand how power phases are only on a potion of the time divided by the amount of them. 12 phase system = 1 phase is on 1/12 of the time and then down the line it goes to the next. Obviously when things heat up, efficiency drops. So as I understand it, more phases, less heat, better efficiency in the power phase. I'm guessing this is a factor for CPU's being not fully stable under load/stress, but i digress.
More Phases = Better or stable voltage efficiency even at hot loads = stable at higher loads/speeds due to lower temps accrued on each phase
So the latter will also be true...
Less Phases = Hotter phase components, less efficient, voltage degradation to a certain point = CPU fails and reboots due to lack of voltage to run at that load or heat level.
So to be stable (assuming the chip doesn't say nope), either you increase the voltage so it's
technically over-voltage at idle but when it's at 100% load, the phases are heated up, the phases go from 110% voltage to 100% voltage due to heat degrading the efficiency of the components
or
You run more phases or more cooling (assuming the in-case temp is hot, not the actual internal components in the phase) to drop the ambient temperature around the phase components and have them run cooler to guard against the heat degradation on the phase components.
ASSUMING I'm kind of right here, I'm guessing all of these factors will weigh on whether a chip is stable or not, and this is another factor (other than getting a good chip) of why some people can stabilize at 1.26 and others have to go to 1.31 voltage.... right?...
I Know there are no guarantees and I'm not going to blame anyone if I break something, but I love these conversations.
P.S. - I understand the whole "getting a good chip" thing, it's a lottery, I know. But I think these other factors may affect CPU stability, maybe only to a fraction, but it makes sense to me. I'm no electric engineer though.