Ah ok the fans dont works properly and you dont use it daily, then i will wait for this but i can try it for fun, 
 
i also clic auto notify for the evga hybrid cooler with a 240mm rad but i not sure if i better wait for a real waterblock from EK, coz the real block wont cost muvh higher then this hybrid and will have better temp.
 
I belive temp on the hybrid will be around 55c and i think its not enough for 2100mhz to 2200mhz.
I think with the hybrid the card will clock at 2050mhz, what you think ?

It is an error in 3DMark. I run 5.1 Ghz adaptive max 1.35 V at the best cores and the rest at 4.9 GHz at steady ~1.22-1.23V. I have a slightly above average CPU silicon quality, nothing special. 
For the Strix, yes I think it is good silicon, for sure better than average but no golden chip at all. My goal was gaming at 2100 Mhz which it does without any issues and the 2220 Mhz on bench is just a bonus.
 
The card keeps 2220 Mhz up until 41-42 degrees then drop to 2190 Mhz or so up until 46-47 degrees then it drops to 2160 MHz. When I game it stays around 2160 Mhz sometimes 2145 Mhz but mostly 2160 Mhz. 
 
I will try a bench later in December at CL 16 or CL15 even (if I can make it stable) and +1300 Mhz on memory and open the windows to get an ambient to 18 or so. That should give me at least 100 or 200 points more.

zogge
It is an error in 3DMark. I run 5.1 Ghz adaptive max 1.35 V at the best cores and the rest at 4.9 GHz at steady ~1.22-1.23V. I have a slightly over average CPU silicon quality, nothing special. 
For the Strix, yes I think it is good silicon, for sure better than average but no golden chip at all. 
 
The card keeps 2220 Mhz up until 41-42 degrees then drop to 2190 Mhz or so up until 46-47 degrees then it drops to 2160 MHz. When I game it stays around 2160 Mhz sometimes 2145 Mhz but mostly 2160 Mhz. 

This is exactly what i was thinking, so the hybrid cooler is not enough coz you will gain just a bit coz the card will downclock too much coz temp.

I think someone said 15Mhz per 5 or 6 deg C. So if you just count backwards, if I have 2220 at 42 degrees, I could in theory get 2235 at 36, 2250 at 30 etc all the way down to the - 100 C or so which is like (100+40 / 6 = ~23 steps x 15 Mhz = 345 Mhz). Adding that to 2220 it should do 2565 Mhz at that temp which is what we see on the top ones in the HoF. 
 
The hybrid cooler is no where near enough for Kingpin! I run an Airplex Gigant 3360x140 + a 1080x120 with 33 fans at 900 rpm (Noctua 3000 rpms running low speed) and 130 l/h. With that I maintain close to 0.5 delta T or so ambient vs loop. So good luck with the hybrid 360 cooler and expect overclocking much.
 

Dear Customer, 

Thank you for reaching out EVGA support team.
Please use the Precision X1 to manual adjust the power to the maximun on TARGET, followed by screenshot the Percision, and send to us.
Regarding to your description, the 500W BIOS on the forum is the maximun Watts. However, according to EVGA forum, "EVGA does not guarantee any performance increase or overclock while using this BIOS update." The performance is already reached to our promised performance, it is related to the quality of the card itself if it could reach to the maximun 500W.

Thank you for your understanding.

Regards, 
EVGA
 
EVGA pretty much just told me the quality of their cards are trash. Max 430w on the 500w Bios on Port Royal. 
So much for this amazing EVGA customer service i heard about can't wait to have this crap out of my PC. What a flat out lie. 

arestavo

MaxTheOwl
I said "I think I have NOT a bad card". Only the average PCI-E draw worries me a bit.

Interesting way of wording it.

But that's overclocked. Set it to default speeds and test again - you'll likely see under 75W average. With overclocking you shouldn't be surprised to see things go out of spec - it's when they are out of spec when stock where that becomes the buyer's choice to accept it or not.

After reading your post, I wanted to ensure my refund for my EVGA 3090 Ultra had been fully processed back to my credit card before I responded.
 
With all respect this statement is seemingly inaccurate, "With overclocking you shouldn't be surprised to see things go out of spec". Without getting into my overclocking experience and different rigs, the current being pulled from the PCIe slot should not exceed 55 watts on a stock 3090 card (no shunts ect.), no matter how overclocked.
 
As you may readily ascertain by reviewing my Port Royal score on a 3090 Strix, I never exceed 50 watts on my PCIe slot. However, I will pull a total of 485+ watts (not spikes) off my PCIe 8 pins.
 
For example, on Port Royal ("too*slow" and ranked 45 = 15079 score) my card provides the following details:
 
Clock frequency: 2,190 MHz (1,395 MHz)
Average clock frequency: 2,136 MHz
 
The point, the correlation you sought to derive between overclocking and PCIe power draw is completely misplaced and provides bad information to people having issues with their cards. In short, these cards are defective. Perhaps not defective, but perhaps intentionally hobbled to prevent the Ultra from competing with the Kingpin.
 
Whatever the issue, it appears EVGA has made a marketing decision that appears to have exploited customer loyalty for the profitability of the Kingpin?
 
Just my two-cents
 
Note: I am not permitted to have links in my posts!

EvgaUser2711201
" / Twitter
 
 
 

Installed it. Did nothing for pushing my card up to 500W power limit.

I have about a week before I have to make a decision to return the card. Kind of ridiculous it's been 5 weeks since the beta BIOS has been out and no official confirmation that this is even a problem and a promise to fix it.

I paid extra for the FTW3 Ultra on the premise that is has the highest power limit of all the RTX3090s out there, and if that turns out not to be true, then I'm returning it and buying a Strix.

AzN1337c0d3r

EvgaUser2711201
" / Twitter
 
 

Installed it. Did nothing for pushing my card up to 500W power limit.

I have about a week before I have to make a decision to return the card. Kind of ridiculous it's been 5 weeks since the beta BIOS has been out and no official confirmation that this is even a problem and a promise to fix it.

I paid extra for the FTW3 Ultra on the premise that is has the highest power limit of all the RTX3090s out there, and if that turns out not to be true, then I'm returning it and buying a Strix.

I literally "upgraded" from the 3090 Eagle because i thought the EVGA cards would perform the best on their bios.... i was wrong. 

professordumbdumb
I'm in a similar boat.  I've asked for clarification from EVGA - as the card I have will not reach even 400w stock on the  bios.  Pulls 75w from the pci-e slot, and 110/110/ ~ 90w from the pci-e connectors.  If I push the power limiter to 107 on stock bios - it will sometimes pull up to 420w, but never higher despite varied workloads with different OC applied.
 
The XOC bios made no difference to performance, it will never reach above 420w.
20141 serial number model.

It would be nice to just get some type of reply. If i had to send my card in to to get some stuff fixed i would be happy too if it performed as well as the Strix. 

 I just do a run of time spy extreme and not score much but maybe not many try that test so iam #62 in hall of fame hehehe :
 
https://www.3dmark.com/spy/15655596

changboy
 I just do a run of time spy extreme and not score much but maybe not many try that test so iam #62 in hall of fame hehehe :
 
https://www.3dmark.com/spy/15655596

Got one of the good ones sir, congrats!

 The thing about that is my room temp is 25.5c actually, and my card still on air lol.
I'am waiting someone annouce a waterblock for this card, i not talk about optimus wich is twice the normal price.

drunknfoo

manatane

Zgapzy

professordumbdumb
I'm in a similar boat.  I've asked for clarification from EVGA - as the card I have will not reach even 400w stock on the  bios.  Pulls 75w from the pci-e slot, and 110/110/ ~ 90w from the pci-e connectors.  If I push the power limiter to 107 on stock bios - it will sometimes pull up to 420w, but never higher despite varied workloads with different OC applied.
 
The XOC bios made no difference to performance, it will never reach above 420w.
20141 serial number model.

It would be nice to just get some type of reply. If i had to send my card in to to get some stuff fixed i would be happy too if it performed as well as the Strix. 

Unfortunately they already answered you that being able to get 500W is now to be considered silicon lottery, which is basically what happens when sourcing components targeting mid-highend market as they did with the voltage regulator on those cards for example (just one of the concrete examples of stuff where they cheaped out). However, it is true it doesn't prevent the card from running at their advertised specs either ...
 
What evga need to realize tho is that whilst they can hold on that and therefore be proud they managed to save some money by not reworking those cards/having a mass replacement for those who got those for higher PL/perceived reputation of product quality / prevent harm to KP sales, it may (and WILL) backfire as burnt customers are unlikely to buy from them again, and unless they're betting on this pandemic being the end of the world, there will be many more GPUs to buy in the future ... from other companies

it isn't based on silicon, cause 500w target isn't reached with the beta bios alone but, 650w+ draws can be achieved with the stock 450w bios while shunted

was refering to this sentence he got back from evga support : 
 

 it is related to the quality of the card itself if it could reach to the maximun 500W

 
this clearly implies that getting any effect from that bios only comes from some sort of silicon lottery (which in this context wouldn't only apply to the GPU itself obviously ) as it implies that not all our cards (though all being FTW3's) are built equal, regarless of the GPU binning (or lack thereof) itself

manatane

drunknfoo

manatane

Zgapzy

professordumbdumb
I'm in a similar boat.  I've asked for clarification from EVGA - as the card I have will not reach even 400w stock on the  bios.  Pulls 75w from the pci-e slot, and 110/110/ ~ 90w from the pci-e connectors.  If I push the power limiter to 107 on stock bios - it will sometimes pull up to 420w, but never higher despite varied workloads with different OC applied.
 
The XOC bios made no difference to performance, it will never reach above 420w.
20141 serial number model.

It would be nice to just get some type of reply. If i had to send my card in to to get some stuff fixed i would be happy too if it performed as well as the Strix. 

Unfortunately they already answered you that being able to get 500W is now to be considered silicon lottery, which is basically what happens when sourcing components targeting mid-highend market as they did with the voltage regulator on those cards for example (just one of the concrete examples of stuff where they cheaped out). However, it is true it doesn't prevent the card from running at their advertised specs either ...
 
What evga need to realize tho is that whilst they can hold on that and therefore be proud they managed to save some money by not reworking those cards/having a mass replacement for those who got those for higher PL/perceived reputation of product quality / prevent harm to KP sales, it may (and WILL) backfire as burnt customers are unlikely to buy from them again, and unless they're betting on this pandemic being the end of the world, there will be many more GPUs to buy in the future ... from other companies

it isn't based on silicon, cause 500w target isn't reached with the beta bios alone but, 650w+ draws can be achieved with the stock 450w bios while shunted

was refering to this sentence he got back from evga support : 
 

 it is related to the quality of the card itself if it could reach to the maximun 500W

 
this clearly implies that getting any effect from that bios only comes from some sort of silicon lottery (which in this context wouldn't only apply to the GPU itself obviously ) as it implies that not all our cards (though all being FTW3's) are built equal, regarless of the GPU binning (or lack thereof) itself

 
It all seems a bit disingenuous to me if that is the answer.  I had another 2 pin card that ran 390w all day long, and my 3 pin ftw3 ultra wont run above that unless I use the power slider to 107% to hit maybe a sustained 415.  It will never hit 100% TDP, and the advertised spec of up to 450w is a fairy tale (bench scores reflect this lack of power draw, as does the power meter on my ax1600i).
 
I'm not opposed to troubleshooting this thing - and working with EVGA to find a solution - however calling this a silicon lottery - as quoted above is bogus (for shunting reasons).
 
 

professordumbdumb
 
 
It all seems a bit disingenuous to me if that is the answer.  I had another 2 pin card that ran 390w all day long, and my 3 pin ftw3 ultra wont run above that unless I use the power slider to 107% to hit maybe a sustained 415.  It will never hit 100% TDP, and the advertised spec of up to 450w is a fairy tale (bench scores reflect this lack of power draw, as does the power meter on my ax1600i).
 
I'm not opposed to troubleshooting this thing - and working with EVGA to find a solution - however calling this a silicon lottery - as quoted above is bogus (for shunting reasons). 

I was just "translating" what the only "answer" we seem to have been able to get from evga on that matter equates to, and yes, thanks for pointing out how nonsensical it is, that was part of the point ;)
 
Then again, considering some have been able to get 500W using this bios and others wouldn't still suggest there actually *is* some kind of lottery happening somewhere unfortunately, call it silicon lottery or not ...