2020/12/26 13:52:32
chester_306
it does 5.2 but it isn't completely stable. 
 
gonna load it up and compare the scores. 
2020/12/26 14:05:58
JonnyVee
pat182
i can confirm a 850watt psu ios not enough when you wanna oc the vram , getting hard shut down pass +1000 on mem :( 
 
strix on air with KPE bios : 14 862
 
 
NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i9-9900K Processor,ASUSTeK COMPUTER INC. PRIME Z390-A (3dmark.com)
 



I have Seasonic Focus Gold Plus 850w. I was able to hit 1200 on the memory without issue
2020/12/26 14:07:29
wwxww
@chester_306
It doesn't really need to be that stable for PR. Feel free to take it to 1.4xxv for a little run or so.


Yes you would gain some pts upgrading.
2020/12/26 14:22:18
Solar Trans
I have my 10900k at 1.4v LLC4 on a Z490 godlike for daily use. 1.4v is nothing at idle for 14nm Intel chips if you allow enough VDroop—by the time a heavy load hits, you’ll be under 1.2v. Remember—high droop and high idle voltage is safer then low droop at lower idle voltage because you won’t see 1.5-1.6v spikes that may be undetectable but will quickly degrade your chip (a flat load line curve is rarely a good idea, especially for daily use!). Also, Max out your VRM switching frequency so the VRM can respond quicker to changes in load.

Buildzoid has used an oscilloscope to measure transients on a number of different boards across brands to find sweet spots for LLC values (balance of spike/droop, load/idle voltages, and stability), worth watching some to get an idea of these trends.
2020/12/26 14:32:47
SoldierRBT
kooncey
 
That is what I had a feeling it what was going on when I was doing my testing anyways. Trying to find the perfect combination for your card with those tweak settings. We could compare numbers all day but each card chip is different and someone running the same settings as the other person could get a completely different score.



Yeah, the KPE just needs the proper combination to get good scores. PR runs with 2200MHz avg but getting below 14.6k is because they haven't applied the tweaks needed. Every voltage point in the curve uses different NVVDD/MSSVD combinations and sometimes too much/low won't work. 
2020/12/26 14:45:24
kooncey
SoldierRBT
kooncey
 
That is what I had a feeling it what was going on when I was doing my testing anyways. Trying to find the perfect combination for your card with those tweak settings. We could compare numbers all day but each card chip is different and someone running the same settings as the other person could get a completely different score.



Yeah, the KPE just needs the proper combination to get good scores. PR runs with 2200MHz avg but getting below 14.6k is because they haven't applied the tweaks needed. Every voltage point in the curve uses different NVVDD/MSSVD combinations and sometimes too much/low won't work. 


I just found my profile for the classy tool I used to get a score of 15756. I used 1.23750 on NVVDD it's most likely what I used to get 15802 nevertheless. However checking out different profile settings I was all over with different high numbers and low from NVVDD, numbers still got me at 15600 and up.
2020/12/26 14:54:12
Sajin
Voltages definitely help when they are right, but having a good chip is how you get the good scores.
2020/12/26 15:24:57
kooncey
Sajin
Voltages definitely help when they are right, but having a good chip is how you get the good scores.

very true, that is what I was trying to say to you when it comes with temps in a different post. I felt like having a lower temp to play with on my card and the possibility of having a good chip helped with those voltages. If so, I thought maybe yours could have been the same if temps were lower, but who knows you could be right about your card. I for one know for a fact if I plug those numbers in and have my pc in my bedroom for the test with no help and all side panels on with my room temp around 68F, it will crash before it even loads port royal. (that is the 15756-score classy tool setting) This is also why I've been doing all sorts of different types of testing. I could care less about my high score means nothing to me when playing games, I'll never reach those marks. I'm just glad I know it's possible for me to hit 17000 and up on liquid nitrogen with this chip I have.
2020/12/26 15:40:25
Sajin
Just need a chip like this... https://www.3dmark.com/pr/653720
2020/12/26 15:43:44
emmett
at pure stock on the kingpin I am getting 250-300 points LESS just by using my ryzen 5950 instead of my intel 9900k
is there any way to regain the points on the ryzen? or thats just how it is? If so i may have to bail on porte royal test
as I am getting tired of moving the card around, and would rather use the ryzen system. 

Use My Existing Forum Account

Use My Social Media Account