Given I’m quite a fan of undervolting (which can be seen here) I have the plan of undervolting my new 5900X as well, once the BIOS and AGESA have been properly updated.
Recently I stumbled over the video below:
In it the author claims that undervolting did not work properly before Ryzen 5000 and now, thanks to PBO we can do something like that. But this clearly is false, as shown in my thread linked above. (But maybe I misunderstood the author.)
So, my question now is if there is any advantage of using PBO’s voltage curve for undervolting? Theoretically, I could still apply a negative offset and that would be it again. So, why should I use the voltage curve?
You can tune the negative voltage per core with curve optimizer so you should be able to get overall less voltage than an offset. Example you have 1 bad core that cant handle more than -5 but the others might be able to do -15.
Basicaly you can fine tune it down to the limit of your cpu to get lower temps and boost longer/higher and get more performance.
I’m not sure how it works under the hood. But i guess curve optimizer could be more intelligent/dynamic in the way it affects the voltage curve compared to a flat negative offset.
It’s not a static value as an offset. It’s a dynamic offset that’s bigger at low frequencies and smaller at high. This makes it so that you get a bigger power budget for low threaded workloads to boost higher.
