Post your overclocking settings, results, and scores on the new Z390 Dark! Could be fun and helpful for the folks who recently upgraded to and built around this new beast of a motherboard.
 
First thing I tested was to see if my G.Skill Trident Z could run at the spec speed of 4,133Mhz. Enabled XMP, ram running at specified speeds and timing of 17-17-17 swimmingly. Going to try and see if I can bring those timings down.
 
After that I ran Prime95 for maybe a half hour to get a quick gauge of my 9900k's temps at stock but with the ram's XMP. Temps were high 50's to 60 max with the vcore kinda fluctuating as the CPU multiplier was still set to auto. Was ranging from 4.7ghz to 5ghz.
 
Currently using the BIOS Stress Tester and have the CPU multiplier set to 50 across all cores, VCore set to Override with a target voltage of 1.300. I first tested with VCore Vdroop DISABLED, CPU Voltage was jumping to 1.355. Too high, but I let the test ran for about 20 minutes just to see if temps would rapidly climb. Ranged from 67-71c.
 
Then I tried setting VDroop to -75% which appears to be "minimum droop". Vcore jumped to about 1.333, temps staying in high 60s.
 
As I type I'm running the stress tester with VDroop set to Auto and it's adjusting to 1.272v, temps staying at 63 to 65 and even occasionally dropping to 62. Going to let this run a bit longer then might load up Prime95.
 
My 9900k was from Silicon Lottery, delidded and binned at 5.0ghz at 1.300Vcore. I don't recall what motherboard they tested with, but it'll be interesting to see if the Dark mobo can run the CPU at this speed in comparison to their testing equipment.
 
What are people's thoughts on the BIOS Stress Tester vs Prime95? Unless folks have different opinions, I'm kinda thinking of the Stress Tester as a good baseline just to see if your CPU can even run a certain configuration, and that if it fails a test in just the BIOS without Windows and whatnot, then you know ASAP that you need to make adjustments.

G.Skill F4-3200C14D-16GTZSK @ 4400 CL17 with tight timings



 
 
Figure I'd use my old account to post pictures:
 
9900k Chip #1 poop - vdroop at disabled
 
5ghz prime

 
9900k Chip #2 - vdroop at -25% which is no vdroop at all
 
1.81 5.0

1.23 5.1

1.29 5.2

1.37 5.3

 
 
5ghz prime

 
5.2ghz prime

 
 
will order more soon in order to find a decent cpu.

Cinebench Passed with a score of 2183
5.1 GHz across all cores
1.3 VCore but with vdroop set to Auto goes to 1.275v
4,133 MHz on RAM, timings of 17-17-17-37, at 1.4v
 
 
 
Going to run Prime95 for awhile now to further test

Heads up, last night I was running the latest version of Prime95 that has the Disable AVX checkboxes so you don't need to edit the local.txt file. Even if checking disable, I'm pretty sure it was still running the AVX test as my frequency was dropping to what it would be based on my AVX offset. At first I thought my OC was unstable so I tried a few lesser OCs but it was still dropping. Tried an older version of Prime and edited the local.txt file to turn off AVX and it worked perfectly.
 
Final OC for last night...
5.2 GHz
1.340 VCore
Vdroop at -25, which brought the voltage down to 1.337 but fluctuated between that and 1.340. Pretty happy that the vdroop wasn't too much with that setting.
RAM settings: 4,133MHz, Timings of 17-17-17-37, 1.4v. 
 
Ran Prime95 Blend for about 4 hours, temps hit high 70s to 82 ONLY when Prime was running the more intensive tests. Aside from those super intensive ones, temps ranged from mid 50s to high 60s based on the given test. Based on this, I might stick with this for awhile. If I remember correctly, Cinebench score was around 2260 if I remember correctly.
 
Going to try and lower ram timings later today to 15-15-15.

So jelly, Im at 5.1 @ 1.33v. Nothin left. Hope to be able to flip this chip soon n put the money towards a real good one

DDR4500 CL17
 

wut? 
 
everything but eleet shows just the VID of the cpu, you see the real vcore only in eleet and in bios/show debug LED vcore setting.

it's only with the evga boards afaik that they show only the VID.

mattman657
E-LEET shows VCore in the Monitoring tab.
 
From what I can tell, the VDroop settings in the BIOS, both negative and positive options, provide a voltage based on the voltage that would be set via Intel's standard VDroop. For example. I have my VCore set to 1.340v. With Intel's standard drop it would have brought it down to maybe 1.300v. But with setting it to the -25%, it reduces LESS of an amount based on that standard 1.300v drop to result in 1.337v. If I set it to -50%, the droop would be even less, which actually caused it to go above 1.340v. I'm guessing the positive settings would cause MORE of a droop, causing the VCore to adjust to something LOWER than 1.300v such as 1.250v.
 
Again the above is just speculation based on my observation and testing the negative values. Someone PLEASE correct me if I'm wrong, don't want to spread false info.

I will throw this out here, not sure if it will help you -
 
 

Tin_EE will have to chime in on that

I'm going to extrapolate some things on the load line calibration since I've had some more time to spend with the board.  And of course the engineers of the board will have to verify my findings.  
 
Here's a good video to watch on load line calibration.    https://www.youtube.com/watch?v=NMIh8dTdJwI
 
the board is set up this way:
 
 will add voltage under load
-75% -  min vdroop     Adds moderate voltage to idle vcore for increase voltages under load 
-50% - small vdroop   Adds a small amount of voltage to idle vcore for increase voltages under load
-25% - less vdroop     Maintains idle vcore and adds minimal vcore voltage under load (probably to maintain idle and a small amount of voltage ripple effect)
default
these will drop voltage under load
+25%  more vdroop   
+50%  Large vdroop    
+75%  High vdroop    
 +90    Max vroop        
 
Negative percentages add vcore under load.   positive percentages allow drop in vcore under load to allow for voltage spike when recovering from load removal. 
 
 
 
positive values, voltage is subtracted to your vcore you set in the bios.  negative values, load on the CPU will show a larger vcore value from what you set in the bios indicating gain of voltage when placed under load.      +90% will be the largest vcore voltage drop under load, whereas -75% will be the largest addition to idle vcore under load
 
So the question is, what should I test my cpu at?    I think for the purposes of establishing a small overclock and a base line, run the auto program in the bios and see what it gives you on auto voltages and vdroop.  this will establish a baseline overclock in the lower spectrum of potential.    For my 9900K, an auto robot test gave me 5.1 ghz at 1.36 volts.   So I then went in the bios and disabled vdroop to see what the minimum absolute stable vcore was needed to maintain stability and found that 1.335 was minimal at no vdroop.    So I then set the droop to auto and tested 1.335 again and it failed.  I bumped vcore to 1.34 and it passed (under load, my vcore was 1.32 so this indicates that auto allowed some vdroop).  Evidently the auto setting allowed too much vdroop to maintain stability lower than 1.34 volts.   So.  Now I know that 1.34 is my baseline with vdroop on auto.  Auto appears to be variable based upon what value you place into vcore.  I suppose it adjusts to make sure the systme stays stabile with varying vdoop.
 
I also confirmed that the robot adds more vcore than what I actually needed.
 
So when do you need to increase or go in the upper direction of vdroop?  (approching negative values)  It will depend on how close you are to your base line stable vcore when placing load and type of loads.    If you have room to go 5.1 or 5.2 then use the +25% setting to allow some vdroop and test stability and temperatures.    If you have extreme means of cooling, like liquid nitrogen and want to push the upper limits for a limited session, then set vcore and push the higher negative values for vdroop.    
 
I hope that helps explain the bios values and until we have something more concrete from the engineers of the board, I'm betting this is correct.   If a person wants a stable, long term overclock on your CPU, you want VDROOP to allow for the over-voltage spike.  That spike over time degrades the CPU.  The video explains this and I think anywhere in the auto range or +25-50% will give you a little and obviously +95% will give you alot. Default will be intel spec.  Disable will be no VDROOP.    So, think about how long you want your CPU.   I don't know what the setting equates to as far as years go, only time will tell.   I don't know how less vdroop will degrade the CPU either.  time will tell.  I do think that an auto vdroop will probably get you the best stability, heat control and longevity. It seemed to allow some vdroop in my tests.     That's probably how they designed auto to work.  Although, I generally don't like auto settings but this is a new board and new time frame.   
 
scenario 1:  I like to maintain my CPU for years, want a modest overclock and need longevity.   Keep vdroop in the positive ranges to offset voltage spike that leads to degradation over time.  
scenario 2:  I like high overclocks 24/7 and don't care how long my cpu lasts, as I replace my gear fairly frequently.   Keep vdroop in the -25 to -50% ranges to add vcore  load. Expect to have a higher voltage spike when coming off load. 
Scenario 3:  I'm kingpin.   Then probably use -75% He may tell us. 
 
One question might be asked.  why not just disable vdroop and set my vcore to a high voltage?..  You'll cause a higher voltage spike, potentially killing your CPU pretty quick. 
 
 

Custom loop with 2 360mm radiators
9900k @ 5.2GHz (x52)
VCore: Override set to 1.340v w/ -75% VDroop
AVX Offset: 0
DIMM: 4133MHz 17-17-17-38 @ 1.45v
 
Cinebench: Passed w/ 2280 - vcore spiked to 1.4v w/ CPU temp ~80c
Prime95: All thread workers passed for over an hour (min/max FFT size set to 1344, run FFTs in place enabled) - vcore maxed at 1.391v w/ CPU temp 60-65c
 
When I tried to tighten up my DIMM timings to 16-16-16-36, Prime95 failed about 20min in. I'm going to try 1.5v on the DIMM to see if it will pass over an hour.
 
Do you think my voltage settings are too high? Overall, very pleased with this board.

Delirious
I don't think your vcore is too high.  so under load your vcore went from 1.34 to 1.4?   and you have -75% vdroop.  sure that's not positive value?
 
what is the recommended baseline voltage for your memory?
 

Yup, negative. My memory baseline is 4000MHz 18-19-19-39 @ 1.35v

Delirious
I don't think your vcore is too high.  so under load your vcore went from 1.34 to 1.4?   and you have -75% vdroop.  sure that's not positive value?
 

 
Wait if one is dialing in a negative number for vdroop, he is reducing vdroop right?
So to get more vdroop, would need to add a positive number?
Therefore a -75% vdroop he is only allowing a 25% margin for vdroop? So to get the voltage to decrease more then positive values should be dialed in.
 
Or am I totally befuddled lol ? Heck I think I may be confusing myself.
 

did some tests monitoring vcore at rest and then under load.  
 
setting            rest vcore            under load            net
-75%              1.352                  1.37-1.394             gain of   .018-.042  volts
-25%              1.346                  1.346-1.352           gain of   .000-.006  volts
+25%             1.337                  1.317-1.302           loss of    0.02 - .035  volts
 
the under load voltages wavered.   So it seems -75% gave a larger increase and -25% gave very minimal increase in vcore under load.    +25% gave decrease in vcore under load and caused me to fail the test.    Brought my vcore at 5.1 ghz down to 1.302 where it failed the test.   So maybe my write up is wrong.  don't know why my rest vcore is varying with the settings.  I've corrected the write up.