Originally Posted by chew*
Measure directly behind socket or pad behind socket if there is one.
Caps were far higher....which is where build zoid measured...
You really only need one person to measure correctly...
Unless a board is defective all boards of that model should measure within .002 of each other..@ identical vcore/pwm/llc settings.
I didn't go through his video till now and saw he did measure at caps. I'll have to look at my board more and see, gotta take the side panel off and move gear around
Originally Posted by shadowxaero
So do you think with the latest bios updates, LLC has gotten "better" I guess and therefore software readings would be more accurate?
I know auto with the Taichi board defaults to LLC5. I had 1.36v set in bios with LLC2 as Hwinfo reported a constant 1.36v.
with LLC5 HWinfo reports 1.305v under load which was strange to me but hey vdroop is a thing. But I still maintained stability at 3.9Ghz (passed IBT on extreme with 32GB of Ram).
The fact that I was still stable makes me question what voltage my chip is actually at.
I just seriously doubt my chip is stable at 3.9GHz at 1.305v under load.
Billzoid believes LLC5 on the Taichi board was the only one that actually has the VRMs deliver close to the voltage set in the bios.
If I've learned anything from AMD boards in the past, LLC is just weird with AMD, I normally just run Auto LLC, I got some weird results with higher LLC settings on the Biostar board, work in progress though with that.
I'm roughly at 1.320v @ 3.9Ghz and I've been running like this for about 10 days now, system shutdowns, benchmarking, stressing, random work loads and no issues. Issue is finding a "good" DMM pointEdited by savagebunny - 4/17/17 at 12:32pm