Time Spy
Normal OC bios
https://www.3dmark.com/spy/3108911
Extreme OC bios
https://www.3dmark.com/spy/3114850
 
2126/1575 (+800)
 
I hate to be the one bringing it to you and it must be weird coming from someone that got a pretty fast and good performing Kingpin, but the Kingpin isn't faster clock for clock is all I can tell you. I try to break it down for you.
 
I thought the special extreme bios would be the key, but it's not. Between the normal OC bios and the Extreme bios is exactly 25 points in Time Spy overall score - margin of error. Most of it because of the higher CPU score. Graphics score wise it's only 11 points difference.
 
I tested 4 new FTW3 Elite 12Gbps editions, 3 Kingpins and they have similar speeds clock for clock, it's only the new FTW3 Elite has much faster GDDR5X. So it wins all the time. I will prove this later on when I test my FTW3 the same way I bench the Kingpin now.
 
It's still beyond me why they didn't put the faster chips on the Kingpin. The memory chips on the Elite are the same as on the TitanXp, so not that new. Would have been easy to use them on the Kingpin, that would have been truly insane.
 
What matters more for graphics score is what plattform you are on. I lost 5% graphics score going from old Devils Canyon to SkylakeX. I gained 2% graphics score going from the X299 Micro to the X299 Rampage VI Apex. The reason for Sajins great score lies between his plattform (the 5690X produces good graphics scores whereas SkylakeX lacks behind for example, just wins overall. Look up graphics score with an 6950X, these are usually much better than SkylakeX) and a good motherboard and Sajin is using a very good record breaking Gigabyte X99 SOC Champion, one of the best overclocking boards for X99 and very hard to get your hands on.
 
Buy the kingpin for the toy it is, more voltages and knobs to play with. But there is no secret sauce. The 980Ti Kingpin was actually faster clock for clock, because it used different (Samsung) memory chips with better timings than any other 980Ti.

My Kingpin kicked the bucket after 3 weeks with no OC at all

Thank you Snipes7 for your excellent analysis. So after all it's the platform difference that affects Graphics Score and there's nothing inherently faster about Kingpin clock-to-clock.
 

Snipes7
It's even more surprising that the Kingpin still gets ~7°C hotter.

Yes, thermal paste difference couldn't cause such a high variance?
 
Most likely the Kingpin GPU is using more power than the FTW3, even though it reports lower vcore. 
 
One thing I've noticed is that there's a difference in how power leaky a particular 1080ti GPU is.
 
For example, I've tested 4 identical 1080ti cards, put the exact same frequency & voltage & fan settings but the best card uses 101% power while the worst card consumes 109%, causing 7C higher temps.

https://www.3dmark.com/spy/3133274

11393

2240 Core

Pretty crappy score.

Lost the lottery on my 8700K (bad... real bad), ran out of DICE (20lbs, 4 hours), not enough time to push Vmem and mem (new Apex Bios w/new ME). After only 2 subzero sessions DVI, HDMI, and 2 miniDP ports gone, literally down to 1 IO and monitor can’t run 144hz with miniDP. HDMI out was pure ice at the end of 4 hours, need to fix. :) still enjoying my time with this card, thought it died completely the first time.