My 2080 Ti FTW3 can't seem to hold any voltage above 1.050V. I get brief spikes but it always settles back to 1.050-1.044V. The card is watercooled running at 47-50C. This doesn't seem right... 

Higher voltage = lots more performance on my card .

I would have to disagree .
I can only get good numbers on benchmarks by wappping the core and memory voltage up .
Unless you are talking about different voltages ?

Why don't you try a different bios ?
Then take the voltages up and see how you go .
The other thing, my clocks start dropping at 30C and over so you need to get card colder .
It will start dropping at your temps .

Take your pick .
there are quite a few different ones that will work on a FTW3 ( my last card so I know)
Have to do some leg work and look for yourself .
I have put a guide here somewhere on how to flash the bios .
Again, a bit of surfing and searching .

Search...
TechPowerUp and Xdevs kingpin.
You find the bios roms there.
You are using afterburner ?
Why don't You try precision x

Only one way to find out .

HeavyHemi

Graphite8five
My 2080 Ti FTW3 can't seem to hold any voltage above 1.050V. I get brief spikes but it always settles back to 1.050-1.044V. The card is watercooled running at 47-50C. This doesn't seem right... 

Why would you desire your GPU to use a higher voltage than what it needs? Are you thinking higher voltage equals higher performance? It does not. What would be better, is if you told us what your boost clocks are and what offset if any you are using. At default settings, the GPU using GPU Boost, will typically use between 1.037 and 1.050 volts unless you manually change the settings to either force a higher voltage or undervolt.

coolmistry

spit051261
Higher voltage = lots more performance on my card .

There no difference to force high voltage to performance and the voltages is automatic monitor due temp and depend on some games how much draw power!!!

+1
 

spit051261
Higher voltage = lots more performance on my card .

spit051261
Search...
TechPowerUp and Xdevs kingpin.
You find the bios roms there.
You are using afterburner ?
Why don't You try precision x

That's partially incorrect.  More voltage doesn't necessarily mean better.  In some instances yes but not always.  It's what the GPU/brand can handle.  Think of it in a silicon lottery sense.  Not everyone will be the same.

Also, you're comparing your experience with a Kingpin that you own to OP's FTW3.  Two different beasts.  Kingpin's custom bios's are designed for it's VRM, not FTW3s.

GTXJackBauer

HeavyHemi

Graphite8five
My 2080 Ti FTW3 can't seem to hold any voltage above 1.050V. I get brief spikes but it always settles back to 1.050-1.044V. The card is watercooled running at 47-50C. This doesn't seem right... 

Why would you desire your GPU to use a higher voltage than what it needs? Are you thinking higher voltage equals higher performance? It does not. What would be better, is if you told us what your boost clocks are and what offset if any you are using. At default settings, the GPU using GPU Boost, will typically use between 1.037 and 1.050 volts unless you manually change the settings to either force a higher voltage or undervolt.

coolmistry

spit051261
Higher voltage = lots more performance on my card .

There no difference to force high voltage to performance and the voltages is automatic monitor due temp and depend on some games how much draw power!!!

+1
 

spit051261
Higher voltage = lots more performance on my card .

spit051261
Search...
TechPowerUp and Xdevs kingpin.
You find the bios roms there.
You are using afterburner ?
Why don't You try precision x

That's partially incorrect. 
NO ITS NOT

Also, you're comparing your experience with a Kingpin that you own to OP's FTW3.  Two different beasts.  Kingpin's custom bios's are designed for it's VRM, not FTW3s.

MY Last card was an FTW3 and I used XOC , LN2 and a few other higher powered ones from TechPowerUp.
They worked fine .
Why is everyone so afraid of trying another bios ?
So I am comparing my experience with the same card as the OP.

Sajin

Cool GTX
2) You can't force voltage - Nvidia Boost 4.0 Has the Last word
 

Sure you can...
 

 
On air the above would be a different story. heh.

I did not say you can't allow more voltage  ..yes, the slider works if the load requires more voltage
 
unlike 900 series & earlier Cards .. you can't Force more Voltage ..
 
Nv Boost will lower MHz or Voltage .. to control total load & Temps

@Sajin
I was not talking about locking the Voltage - guess that was not clear
 
Not sure I'd Voltage lock for daily driving
 
My reply is based on the OP, who wanted More out of the card - Post #1 & 7
 
 

Graphite8five
 
My 2080 Ti FTW3 can't seem to hold any voltage above 1.050V. I get brief spikes but it always settles back to 1.050-1.044V. The card is watercooled running at 47-50C. This doesn't seem right... 

 

Graphite8five
If you force the card to use maximum voltage that is what you'd expect the card to do, mine is not. Voltage + power sliders to max and +115 offset = crash. The card cannot do 2100MHz stable. MSI OC Scanner gives me an average offset of +75 which is poor. This results in 2085 for a short moment before settling at 2055-2070Mhz. The card is watercooled at 47-50C. I can't possibly have a card this bad. My previous FTW3 ran 2100 at 60-64C air cooled.
 
The problem seems to be with the voltage. A +100 offset should result in 2100MHz at 1.080-94V (according to the voltage curve) but the card keeps falling back to 1.050V. 

BTW   Nice Temp on your KP
 
Even in a 3 C room my air cooled 2080Ti run a little warmer @ 45 - 46 C
 
 
 

I don't like how much of a "get gud" vibe I keep getting in this forum. People always seem to prioritize user error or incompetence over hardware or software fault. It's also ridiculous imo to so casually suggest lowering temps to sub 40c for better performance. Even on water cooling gaming at below 40c is not at all easy. Neither is getting the same under real loads.

I've become of the mind now that all these benchmark apps are terribly useless. Just as an example; in furmark I can stress the GPU all day long and stay in the low 40s, and I can pump up the volts and clocks to obtain higher bench scores to the point where the card just won't take anymore juice. The same thing for superposition. Temps stay in the 40s, and the card devours the volts. What's even worse in superposition, is that it lies. The numbers it feeds me are way above what the card is actually doing. Both of them will tell me the card is being utilized at 98-99%, but the card doesn't sweat it. Gaming on the other hand is a different story. I never get the same results while gaming. Last night benchmarks were letting me max out volts and set the clock up to 150+ to the point where adding more didn't raise the clock while benching. So I loaded up Metro Exodus, sat on the title screen with 2115. Loaded my game, and crashed seconds later. Tried it again and again and again, each time lowering the clock more and more, until eventually it was stable again at just 90+, giving me the same 2050ish that I had been getting before when I only used what the scanner gave me.

There is something going on with how power is being used. Afterburner's scan will normally give me a score of 90-99 if I don't touch the power bar, but if I set it to (supposedly) allow 124% the scanner will appear to be getting good numbers until it ends and gives me an average of 75. Clearly there is a stability issue when allowing the card to take more power.

Graphite8five

HeavyHemi

Graphite8five
My 2080 Ti FTW3 can't seem to hold any voltage above 1.050V. I get brief spikes but it always settles back to 1.050-1.044V. The card is watercooled running at 47-50C. This doesn't seem right... 

Why would you desire your GPU to use a higher voltage than what it needs? Are you thinking higher voltage equals higher performance? It does not. What would be better, is if you told us what your boost clocks are and what offset if any you are using. At default settings, the GPU using GPU Boost, will typically use between 1.037 and 1.050 volts unless you manually change the settings to either force a higher voltage or undervolt.

If you force the card to use maximum voltage that is what you'd expect the card to do, mine is not. Voltage + power sliders to max and +115 offset = crash. The card cannot do 2100MHz stable. MSI OC Scanner gives me an average offset of +75 which is poor. This results in 2085 for a short moment before settling at 2055-2070Mhz. The card is watercooled at 47-50C. I can't possibly have a card this bad. My previous FTW3 ran 2100 at 60-64C air cooled.
 
The problem seems to be with the voltage. A +100 offset should result in 2100MHz at 1.080-94V (according to the voltage curve) but the card keeps falling back to 1.050V. 

Sorry that is not how it works.  You are not FORCING anything. All the voltage slider does it ALLOW the GPU to a higher voltage IF GPU Boost determines it is usable. I've never ever seen it in practice and only seen those voltages using the voltage curve to push it higher. Personally I UNDERVOLT to 1.025v at 2037mhz.

Putting aside the issue of benches vs games stability, there is still the issue of voltage when it comes to said stability. I would have been ready to accept that the game it's was responsible for the lose of stability as apposed to the benchmark's supposed enhanced stability, if it were not for the Afterburner scanning that's supposed to be scanner different combinations of clocks and volts, without all the bells and whistles of game's programing and graphics. So it seems to me that voltage has more effect on stability if I'm getting more errors (and thus a lower average scan score) with higher allowed voltage, and less errors when voltage is not allowed to go over. Which then leads me to ask at what point can we say that the hardware is faulty if and when that hardware is designed to allow for 24% extra power, but then that extra power causes stability issues?

The issue I'm talking about stands apart from the temp:clock ratio. I've used many different benches, for long periods of time (not just the short scoring tests), and I've also had the issue present itself in mere seconds after a game has been loaded. While using the benches for long periods of time I've seen temps stay low no matter how much the clock is raised and\or volts are supplied. Even though the benches report using the card fully, this is clearly not true. If the card were being truly taxed at those frequencies then the temps would shoot up and thus lower the clock. On the other hand, I've used those same clocks and volts in games only to have them fail almost immediately after the card is truly taxed, before the card even had a chance to heat up and lower the clock. The only apparent culprit is the voltage management for maintaining frequency.