I think there's still a v/f points bug in this bios. I was able to confirm that if you set several points at different settings, it only takes the highest frequency one you set. For example, if I set 5100 and 5400 at -35 and 5800 at -10, it applies -10 at all frequencies. HWiNFO64 was actively showing the offset concurrently with the frequency, and even looking at the min/max columns showed that it only ever did -10 even if you hit it with a heavy workload. The lower frequency settings never seemed to take if you give it an offset for higher frequencies.
I can see if I set 5400 and 5700 at -30 but leave all the higher ones at 0, the offset will trigger at the correct frequencies. It does not, at least as far as HWiNFO64 shows, do a linear curve-fit between different points or outside those points. It just does a global offset between the two. I tried to check this by setting one at -25 and the other at -35. I believe it did -35 or no offset with that setting (only applying the offset from the highest altered v/f point).