I was tinkering around with the power limit to lower temps while I'm waiting for the hybrid kit to come out (tower 900 was not designed to be used for air-cooled cards... but a few industrial Noctua fans fixed that, 70c max power limit 118% achieved) and I noticed that if you use only 90% power target your score jumps around 700 points. Temps (63c) is the same as 100% and 85% power target runs but the score is much higher. It must be due to the boosting tables but I thought it was interesting that 90% PT got a better score than 100% PT at the same temperature.
 
https://www.3dmark.com/pr/419306 90% PT
https://www.3dmark.com/pr/419260 100% PT

Crazy how 3080s can get to the same score and 3090 XC3 Ultra gets when overclocked on its stock bios.

I think there are some other things going on there for you...
10,400 is what i score on my 2080ti

That said ill mess around with it and see if i can replicate better scores with lower power

My one Port Royal score ever (not even trying) with an overclocked 2080 Ti Hybrid is 10,258....
 
https://www.3dmark.com/pr/307699
 
The 3080 FTW3 should be at 380W @ 100% and 342w @ 90%, PR score should be over 11K with stock settings.  There is something kinda wrong with your system or card.

Mylez_
I'm sure it's my i7-8700 non-k, my friend is giving me his i7-8700k in a few weeks so it might be an improvement. But it's most likely that.

Everybody says port royal is not CPU dependent, but you might want to look at your CPU utilization per core to see if any cores are at 100% and bottlenecking.

Mylez_
I do monitor my CPU usage when benchmarking, no single core went over 60% utilization. 

Yeah then probably not a CPU issue then, but someone with better direct Port Royal experience will have to weigh in.

kevinc313

Mylez_
I'm sure it's my i7-8700 non-k, my friend is giving me his i7-8700k in a few weeks so it might be an improvement. But it's most likely that.

Everybody says port royal is not CPU dependent, but you might want to look at your CPU utilization per core to see if any cores are at 100% and bottlenecking.

Theres a youtube video, pairing a 3090 with FX 8350.
His benchmark on his 8350 vs 9900K was only 5 points different in port royal.
Port royal really doesnt care about CPU compared to games and benchmarks like time spy.

ehabash1
Did you enable XMP? Port royal seems to scale with memory frequency. What's your ram set at
Did you exit out of precision x1? Not minimize literally exit out of it
How is the power setting set in your Nvidia control panel? There are too many variables

Did you not set the fans to 100% for the benchmark? You mentioned some temps but they sound high. I would like to see you pass 12k

100% fan speed, prefer max performance,  XMP enabled 3200 MHz cl 16, power plan max performance (windows) px1 was saved applied saved then task manager killed, no other programs running. 750 watt PSU total sys power draw according to kilowatt was 620 watts. G-sync was enabled but I doubt that was causing it.

With max XOC power limit total system draw from the wall was 620 watts. I have 130 watts of headroom. I could get a new psu but it wouldn't do much as I have room for more power.

I totally agree, but a system will never push all core + full gpu while running a game, most of the time it's like 2-3 cores pushing 60-80% usage boosting high for speed. I do agree I need a bigger one eventually and I have a buddy who might trade me straight across for his 1000w evga psu for my 750w evga so we shall see. Also the 620 watt is from the wall through my kill-o-watt meeter not an on system measurement. so technically I have that 10% inefficiency to account for when figuring total power draw. I will most likely get a bigger psu at some point but for the time being it works.

My testing methodology is kill-o-watt from the wall on a surge protector 20 amp circuit. I can read down to .5 watts and see spikes for the past 5 minutes.

Cinebench r20 with igpu enabled pulled 220 watts avg no gpu in the system. Same benchmark with a 730 running in low power mode 190 watts avg. This was full system power draw at the wall with oc. The 8700 has a nifty little firmware bug in the early batches that allows you to overclock a locked chip it just gets really hot, so I don't mess with the core ratio alot just 4.6 instead of 4.3-4.4 GHz at 67c all core turbo. My thinking was if my total system power draw with all fans maxed, pump going, 360mm rad fans going max, cpu stressed and diskmark going on in the background was 200 watts overclocked I should have enough headroom for 450 watts peak 470 with 50ish left over. I've put alot of thought into it before I bought it because I was concerned about power limitations. But I could be wrong and I will most likely get that psu this week to see if I am limited.

I appreciate all your responses and I will keep you updated if I do get that bigger psu.

Mylez_
I totally agree, but a system will never push all core + full gpu while running a game, most of the time it's like 2-3 cores pushing 60-80% usage boosting high for speed. I do agree I need a bigger one eventually and I have a buddy who might trade me straight across for his 1000w evga psu for my 750w evga so we shall see. Also the 620 watt is from the wall through my kill-o-watt meeter not an on system measurement. so technically I have that 10% inefficiency to account for when figuring total power draw. I will most likely get a bigger psu at some point but for the time being it works.

My testing methodology is kill-o-watt from the wall on a surge protector 20 amp circuit. I can read down to .5 watts and see spikes for the past 5 minutes.

Cinebench r20 with igpu enabled pulled 220 watts avg no gpu in the system. Same benchmark with a 730 running in low power mode 190 watts avg. This was full system power draw at the wall with oc. The 8700 has a nifty little firmware bug in the early batches that allows you to overclock a locked chip it just gets really hot, so I don't mess with the core ratio alot just 4.6 instead of 4.3-4.4 GHz at 67c all core turbo. My thinking was if my total system power draw with all fans maxed, pump going, 360mm rad fans going max, cpu stressed and diskmark going on in the background was 200 watts overclocked I should have enough headroom for 450 watts peak 470 with 50ish left over. I've put alot of thought into it before I bought it because I was concerned about power limitations. But I could be wrong and I will most likely get that psu this week to see if I am limited.

I appreciate all your responses and I will keep you updated if I do get that bigger psu.

I upgraded from a 750W to 1000W power supply this week and my stability at higher overclocks immediately improved. Just because you're not maxing out your PSU doesn't mean it isn't a limiting factor. Voltage fluctuations kill overclocks.

Swapped my PSU with a buddy's 1300w. Now my max score is 12,208 with XOC bios and I'm actually power limited by the card!  but I'm hitting 83c with a box fan blowing over it. Time to watercool! I cant wait for the hydrocoppers to come out!
https://www.3dmark.com/3dm/52252557?