Uuuuuuuuuuh yeah that's low indeed.
 
Do you have a link to your run so I can take a look at it?
 
The PSU isn't a problem, that GPU temp is definitely too high on the first run, that's for sure - if you could post a picture of your case setup that might also be helpful.
 
For reference, this is the score I got with the GPU at full stock, with me running my daily-use OC on my CPU/RAM (5GHz CPU, RAM at stock 3600MHz CL16): https://www.3dmark.com/pr/818304 (13790)

What drivers are you using?

I just added the link to my test results in the OP. I haven't tried LN2 bios. I tried OC bios and got even lower 12.5k score. Power also never exceeded 418W. I'm using the latest Nvidia drivers.

12.5 is a OC'ed 3080 score...something is amiss

That's a tiny case :D You don't use a PCIe extension cable to make your GPU vertical etc do you? I tried one that I knew was too long and I was getting 12.5k too o.O

You can upload the file to EVGA if you click "Pop out full veresion" and the file is under 500k, just drag it into the box and it will upload, or you can upload it to imgur etc and link to it.

The thing about these threads that makes me really interested is the fact that in the original 3090 release the FTW3 Ultra 3090 was ticking 14,900 and sometimes up to 15,000 in port royal on air.  Granted, that was some cold air on those runs...but we are also talking about a 450w limit on the BIOS too.  Without crazy cold air, we saw 14,400 and 14,500 across many cards.  
 
So how is it that a card with better power delivery, a factory AIO, and a 520w BIOS is so damn far off that mark?  Yes, the silicone lottery plays a part, but we are really seeing 3080's outscoring a lot of these 3090's.  Hell, even my 3080 is hitting 12,712 on air with an average 44c temp across a run and my current 3090 undergoing RMA was clocking 12,200.
 
Another thread I was watching has been tracking if the cards are coming from China or Taiwan.  My 12,700 scoring 3080 is a 2012 serial number from China and my 12,200 scoring 3090 is a 2114 from Taiwan.  

Dyezak
The thing about these threads that makes me really interested is the fact that in the original 3090 release the FTW3 Ultra 3090 was ticking 14,900 and sometimes up to 15,000 in port royal on air.  Granted, that was some cold air on those runs...but we are also talking about a 450w limit on the BIOS too.  Without crazy cold air, we saw 14,400 and 14,500 across many cards.  
 
So how is it that a card with better power delivery, a factory AIO, and a 520w BIOS is so damn far off that mark?  Yes, the silicone lottery plays a part, but we are really seeing 3080's outscoring a lot of these 3090's.  Hell, even my 3080 is hitting 12,712 on air with an average 44c temp across a run and my current 3090 undergoing RMA was clocking 12,200.
 
Another thread I was watching has been tracking if the cards are coming from China or Taiwan.  My 12,700 scoring 3080 is a 2012 serial number from China and my 12,200 scoring 3090 is a 2114 from Taiwan.  

Kingpins can hit 16k+ if the conditions are right, 15k really isn't that hard for them if your system is setup for it either.  I hit 15.4k for my personal best, and that was at a benchmark stable (as in it could run multiple runs and have no crashes) +150/+1250, with ambient being at room temperature so nothing special.  At stock, any 3090 is going to get around the scores we're talking about, at around those clocks (1965-1995MHz core clock) - with some tweaks you can see slightly higher or slightly lower, depending on how your system is behaving.  I've gotten as high as 13933 completely stock on the GPU, with no special air temps.
 
I also had the problem of not being able to hit 15k at those very clocks (+150/+1250) then after a few driver changes, randomly hit 15.4k, repeatedly, then haven't been able to hit it since.  So there is entirely a real thing of having 400+ point swings for completely random reasons.  I'm on the same drivers I was for a 15410 run and I could run with the exact same conditions right now and probably barely get 15k.
I don't know the reasons "why" or how to correct them/make it happen again - if I ever figure that out, I'll be sure to post about it.
 

EVGA1225

Dabadger84
 
 
That's definitely much better, and pretty crazy that Precision was leeching your score that much.
 
Your core is still boosting lower than it should, you have a 2010MHz starting boost vs my card was hitting 2055 at the start, and your average was 1979 vs my 1998 - I'm unsure if that small of a clock difference could actually be a 250pt difference, but at least now you're close to where you should be.  Your average temp is 10C higher than mine was on a similarly-stock run, that's probably an explanation as to why the boost clock is lower.  Fact that it's only going down ~20MHz for a 10C increase in temps is actually pretty good.
 
Do you have the GPU radiator setup as intake or exhaust, how many fans are setup as intake (again a picture of the actual system setup would be great)?  Try running again with the side of your case off & see how it goes.
 

I have the 3 SL-120 setup on pull and exhausting out of the case. It averaged 43deg during the benchmark with open case and around 57deg with closed case. During intense gaming, it gets about 63deg. It is comletely stock with no o/c at all. I'm wondering if the stock thermal paste is not applied correctly which is common but I don't want to disassemble it since I don't know which thermal pads to get for the memory. Can you tell me what average temps you got at stock and if you were running open case?

I haven't run it at stock for an extended period of time, but for a Port Royal Benchmark run I think the highest "average" temp I've had listed on the results page is like 49-50C, and that was with a pretty heavy OC. 
 
https://www.3dmark.com/pr/818304 13790 @ stock, 44C average temp, 1998MHz average clock
https://www.3dmark.com/pr/819440 14975 (before I magically fixed things), 50C average temp, 2179 average clock
https://www.3dmark.com/pr/820304 15037 (still before magic fix), 49C average temp, 2177 average clock
Then there's the magic fix scores, with ambient of 66-68F:
https://www.3dmark.com/pr/826610 15409, 46C average temp, 2178MHz average clock
https://www.3dmark.com/pr/826695 15461, 48C average temp, 2192MHz average clock
 
I think the hottest temp I've seen on the GPU core thus far was 57C back when I still had the stock/included fans on the radiator, when I forgot to prop up the top of my case to let the exhaust air flow (dust filtration hinders exhaust flow & causes some recirculation of hot air which in turn raises temperatures) and my ambient was around 74F during a Cyberpunk session with my room heater on.  Other than that I don't think it's ever went past 55C during gaming/actual usage.
 
If I recall correctly, the O11 Dynamic has the option for bottom intake, do you have fans as intake on the bottom?  If not, try that, it should help with case temps.
What's the current fan setup?  Having something like this with the CPU radiator (if you have one) as the other intake, and the GPU radiator up top as exhaust, would be preferred:
 

Dyezak
The thing about these threads that makes me really interested is the fact that in the original 3090 release the FTW3 Ultra 3090 was ticking 14,900 and sometimes up to 15,000 in port royal on air.  Granted, that was some cold air on those runs...but we are also talking about a 450w limit on the BIOS too.  Without crazy cold air, we saw 14,400 and 14,500 across many cards.  
 
So how is it that a card with better power delivery, a factory AIO, and a 520w BIOS is so damn far off that mark?  Yes, the silicone lottery plays a part, but we are really seeing 3080's outscoring a lot of these 3090's.  Hell, even my 3080 is hitting 12,712 on air with an average 44c temp across a run and my current 3090 undergoing RMA was clocking 12,200.
 
Another thread I was watching has been tracking if the cards are coming from China or Taiwan.  My 12,700 scoring 3080 is a 2012 serial number from China and my 12,200 scoring 3090 is a 2114 from Taiwan.  

High scores in PR and pretty much every benchmark comes down to 50% silicon lottery 50% optimized setup/overclocking skills. People can’t expect to hit 15K+ in a system with stock CPU/RAM speed and 10+ apps running in the background.

With the right settings and temps, every 3090 KPE should be able to hit 15K no problem. My 3090 KPE can score 15K below 420W ambient temps and that’s powering 4 fans in total + an AIO pump. When your system is optimized and you have the proper OC skills, the rest it’s just silicon lottery.

KPE waterblock hasn’t been released but I’d expect people getting 16K+ on water with these cards. My max is 15.7K with 52C max temp.

I also have a 3080 FTW3 2014 serial number can do 13K under 390W and 13,4K 450W on air 51C avg.

its too weak, try to switch to the LN2 BIOS

I have a similar issue but mine are actually lower, around 11K.  I did a clean install of the drivers and have plenty of cooling.  The only issue I see is that the 3rd 8 pin power is always lower than the other 2 by around 10w.  I seem to be getting fair Clocks but low scores.
 
Here is a link to my PR results
www.3dmark.com/pr/993738 Don't have enough posts to link directly
 
 
Here are the results from GPUZ