I was surprised because the spikes only started appearing when I upgraded to 1253. The vdroop I noticed with the old bios was very minimal. If I ignore the random spikes with the new bios, the vdroop is still about the same. (~0.02 or so)
That's why I'm tempted to think that the spikes had nothing to do with LLC.
I'm using offset mode so that it will lower voltages when it's idle. If I forced it to a particular voltage, it stays that high even at idle.
I'll try lowering the offset and turning on LLC to see what happens.
I'll repost here:
BIOS: 1253
vcore offset mode: +0.085
LLC: forced off (standard)
PLL overvoltage: on
Everything else left on defaults...
Idle vcore: 1.09v.
Load vcore: 1.32v - 1.34v
Random vcore spikes: 1.38v - 1.41v
Results with these settings:
4.2 GHz (24/7)
4.5 GHz (completely prime stable)
4.6 GHz (4 hours prime stable)
Those are my 24/7 settings. 4 hours prime-stable @ 4.6 GHz. Completely prime-stable @ 4.5 GHz. But I'll be doing software development on it - so I'm clocking it at 4.2 GHz to be safe.
For my temporary suicide run, I put the offset to +1.25 or so (I don't remember the exact number.)
With this I was able to get 4.9 GHz at least 10 min. prime-stable. 5.0 GHz was screenshot-able but not bootable. Load voltages were over 1.43v so I didn't want to keep it there long enough to get them stable.
For this suicide run, I left my window open to let the sub-zero ambient outside air into the room and the case. (ambient outside was -10F, inside the case, it was about 40 - 50F)
So temperatures were never a problem. I'm "sorta" on water. (see my siggy for exact specs)
Bookmarks