You shouldn't have to mess with voltage at all for 4.5ghz, I think. If you don't hit it, though, try 1.28v and then 1.3v. I've heard that 4.7ghz generally takes around 1.35v, but that might wear the chip out too quickly for anyone's taste.
Geez, I envy everyone's i5-2500k cpus. Mine is not 45x stable (I actually can't achieve it no matter what I do). I didn't get the luck of the draw...
The most I can push out of mine is 43x with an average vcore of 1.293v. Stable using P95 Blend for about 15 hours now with no errors. But notice vcore did spike as high as 1.536v for a very brief moment. I'm going to see if this is rock solid 24 hours stable with P95 Blend (or at least see how long it remains stable).
Mind you, I won't have mine on 24/7. Maybe at most 48-72 hours straight, but that's about it.
Overall, vcore for your 43x is looking really good. Very close to stock voltage. As for me, I'm not that lucky...
It should never spike that high, whatever vcore you set will at most change +/- .1, even with llc on extreme. It could be an incorrect reading, it happens, but since you can't get to 43, there may be other issues. More info could help.
With these settings I'm 4.3Ghz rock solid stable with P95 Blend for 24+ hours. My previous 4.3Ghz OC using the previous BIOS only lasted ~10 hours before system froze and required a hard shut down.
P95 Blend still running after 24 hours:
P95 Blend stopped after 24 hours:
I'm a novice OC'er and I did try pushing it to 4.5Ghz, but every attempt I made always resulted in P95 rounding error or BSOD code 124. Increasing vcore didn't help with stability and even playing with DIMM voltage didn't help either (ranging from 1.55 to 1.60625v - rated RAM is 1.6v). Always got an error or BSOD. Memtest86+ came back no errors at rated timings and lower 1.55v (ran over night 8+ hours). Then I heard about this 'wall' which is what I think is happening here with 4.5Ghz. I could be wrong. Anything you can suggest?