n0ns3ns3 :
watts are not that important if the temps are fine.
what are the cores temp under load ?
Yeah, that's the thing. The temps are ridiculously high and failing some kind of improper BIOS setting I have a chip with TIM on the die that is so bad it actually would qualify for an RMA as it cannot be run at stock settings with the stock cooler.
So I'm chasing down why the total watts being used is so high. Right now it's seeming like the LLC controller (or however that is set up) is allowing/setting the amperage too high under load -> hence the very high watts at relatively lower frequency and volts
😉
For specifics, hottest core under load (IBT) hits 84C. 1.18 Vcore (hits 1.2 under load) @ 4.6GHz.
Cooling is done with a new "Old style" H100i (the square block that plugs into a SATA and not a fan header) with 4 SP120 fans in push-pull, mounted at the top of a 400r.
In addition to the two SP120 fans at front I've mounted an extra one on the back of the drive cage and at the bottom of the case for intake.
2 140mm Cougar Vortex fans on the side of the case round out the intake fans. There is also an SP120 at top rear for exhaust.
I think the case is well ventilated and my CPU cooling is pretty adequate
😀
I have tried two different H100i's in different configs with different TIM (stock, mx2, mx4, gelid gc-extreme) best has been gelid with the new H100i in push-pull + lower volts and lower watts (achieved by lowering LLC from "Extreme" to "Turbo"). Used to be at 94C (hottest core) but was also at 1.224 Vcore, 1.9VRIN and LLC "Extreme" and the watts were reported at +/- 180W. Just those voltage changes are worth 10C (and 40W!!!) so the cooling made no difference but the wattage did.
Anyways, I think I've narrowed this thing down to BIOS setting that will need to be adjusted - if possible. If that doesn't give me the results I'm after I can always consider a real WC loop and delidding :lol: