5900X undervolt curve

[therealdippy]

I while back I went through the undervolt process using Cinebench and tuned a set of PPT/EDC/TDC values I liked with what I thought were some pretty good undervolt values: -20 on the fastest 2 cores, -25 on the next fastest 2 cores and -30 on the rest. Some time later after a game crash I discovered WHEA errors which indicated that the 2nd fastest cores had caused it so I increased the undervolting for those 2 cores by 1 notch to -24. A month or 2 later after gaming with no problems, I thought that since I seemed to have found the limit for those 2nd fastest cores and the slowest cores were already at the lowest -30, maybe I should try to find the limit for the fastest cores. So I have been reducing them by a notch every so often. Now I have reached -24, the same as the 2nd fastest cores, and seemingly no problems.

My question is, since I have "proved" that the limit for the 2nd fastest cores is -24, should I assume that the fastest cores shouldn't be undervolted any further? It would seem strange to be able to undervolt the fastest cores by more than the 2nd fastest.

 

[Cyber-Mav]

Your not testing these undervolts properly.
Prime95 is a start along with cinebench cycling.
-20 is very aggressive id suspect -10 to be more inline

 

[humbug]

I while back I went through the undervolt process using Cinebench and tuned a set of PPT/EDC/TDC values I liked with what I thought were some pretty good undervolt values: -20 on the fastest 2 cores, -25 on the next fastest 2 cores and -30 on the rest. Some time later after a game crash I discovered WHEA errors which indicated that the 2nd fastest cores had caused it so I increased the undervolting for those 2 cores by 1 notch to -24. A month or 2 later after gaming with no problems, I thought that since I seemed to have found the limit for those 2nd fastest cores and the slowest cores were already at the lowest -30, maybe I should try to find the limit for the fastest cores. So I have been reducing them by a notch every so often. Now I have reached -24, the same as the 2nd fastest cores, and seemingly no problems.

My question is, since I have "proved" that the limit for the 2nd fastest cores is -24, should I assume that the fastest cores shouldn't be undervolted any further? It would seem strange to be able to undervolt the fastest cores by more than the 2nd fastest.

Its quite difficult to decipher what you are saying here.

As i understand it you have the second fastest cores at -24 and the fastest at -24 with the slowest at -30 and you're asking if you should try for below -24 on the fastest because being better they should go lower, but you have the slowest cores at -30.

I don't really know what to tell you, other than anything over -20 is very high, i'm surprised its stable at your settings. but if it is just leave it and be happy with what you have, don't over think it