Does anyone know if this is normal? My time spy extreme and time spy scores are just horrible running these 2 kingpins in SLI. When running the gpuz rendering test the bottom card stays about 80 watts but the top card goes to over 300 watts. When running like furmark both cards jump to pretty much equal wattage as well and they also do in time spy extreme but yet my graphics score is just pitiful. In port royal they both boost to around 700 watts as well. In a nutshell port royal single card, port royal sli, time spy extreme single all perform as they should but time spy extreme sucks to all get out and i can't figure out why. In this gpuz shot the pcie slot power on the one card stays at zero all the time even during benchmarking. This photo is of the cards at idle. 

Threadripper isn't the issue... https://www.3dmark.com/spy/13065779

What happens when you run the cards at stock?

On my godlike Z490 , there is a seperate power connector just for PCEi.
Could it be a power issue on your
MOBO?
Saying that, my KPE sli scores were pants .

This is the first board I have used that had a separate, dedicated PCEi power connector.
Maybe his board is not up to it ?
It is an AMD after all .

Yeah i have both cpu 8 pin power connections plugged in, the cpu 6 pin power connection plugged in, and the 4 pin molex at the bottom of the board plugged in for extra power to the pcie slots for multiple gpu's. 

OK so with the kingpins back in my intel set-up the pcie power usage still shows the same thing with 0 usage so that seems to be a kinda normal thing i guess because my scores are way back up now. 
 
 
 
 
 
 

spit051261
What was the problem?
Don't tell me it was AMD, as if I didn't know lmao .

Don't know yet, once i put the cards back into the AMD setup i will have to try and narrow it down from there now that i know its not the kingpins or the nvlink bridge, don't know if its a bios issue for the AMD motherboard, might try a different bios and i'll probably reinstall the operating system as well.  Its weird, on the AMD setup my 3970X cpu score is almost 5000 points higher then the intel 9980XE, but on the AMD setup my GPU score is 3000 points lower then the Intel board with the same 2 gpu's. Challenging my knowledge for sure. 

spit051261
I always thought GPU benefited from faster CPU clock speeds?

Depends on the resolution. Back when the 980 Ti's were new, my 980 Ti SLI at 1080P setup was hugely CPU bound - it wasn't with a single 980 Ti. 4K moved the bottleneck to the GPU(s) where it has stayed so far even with newer & faster GPUs.

spit051261
Even with KPE in SLI ?
Wouldn't it be back on the CPU then .

At 4K? Only for games that don't need GPU horsepower to begin with.

Nice scores!  SMT was the culprit?  Still couldn't the bottleneck be on PCIE and not the KPE's  I am still not convinced that their are not AMD PCIE gremlins at work.

Glad you figured it out. 👍

sparetimepc

Sajin
Glad you figured it out. 👍

Amazing what one setting in bios can do !!!!

It is always the little things that trip you up
 
 
Congrats on the killer Score        https://www.3dmark.com/spy/13100025

I get 21-24K when i run my 2 RTX 2080Ti and a 9900 processor.  Scores seem low for 2 Kingpins!

Nazarus
I get 21-24K when i run my 2 RTX 2080Ti and a 9900 processor.  Scores seem low for 2 Kingpins!

You are comparing Time Spy to Time Spy Extreme.