mattschlosser
Well, I am another EVGA 3080 Ti FTW3 disappointed buyer you can add to the list.  The card almost never breaks 400w even under stressful test conditions that pull other cards to that limit.  It is not "normal" behavior, even if it does sound nice (or sound true) to say that cards "don't need to stay operating at the very limit just because that is the stated limit". That might be true, but the behavior of these cards in the past is that they DO operate close to the power limit when your power limit slider is at 100%.  And when you have a bios that goes beyond 100%, the average power draw goes up!
 
Not so with these cards, and I would say it is a defect.  I even have voltage headroom (supposedly the voltage limit is 1.09v on these) at like 1.05v when hitting the 400W power limit.  So my card NEVER goes beyond 1.05 or so, even if I lock the voltage higher in Afterburner.  This is because its hitting the 400W power limited sooner than the voltage limit.
 
I wish EVGA would release a BIOS with this fixed.  But, since they are no longer in the GPU game, they probably will do nothing.  I consider it to be false advertising.  The reason I got this was because of the stated power limit of 450W lending to better overclocking headroom.  Well, based on what I've seen others go through when discussing this with EVGA, it seems like pretty terrible and apathetic customer service.

I don't have any problem at all hitting 450 watts. It's not a BIOS problem. How about some more details on the rest of your system and what you're using to load the system?
MSI Kombustor stress test should load the GPU to 450 watts using the default stress test. Make sure you have Vsync and any frame limiters disabled. You should be seeing around 140 FPS with the default settings.

HeavyHemi

mattschlosser
Well, I am another EVGA 3080 Ti FTW3 disappointed buyer you can add to the list.  The card almost never breaks 400w even under stressful test conditions that pull other cards to that limit.  It is not "normal" behavior, even if it does sound nice (or sound true) to say that cards "don't need to stay operating at the very limit just because that is the stated limit". That might be true, but the behavior of these cards in the past is that they DO operate close to the power limit when your power limit slider is at 100%.  And when you have a bios that goes beyond 100%, the average power draw goes up!
 
Not so with these cards, and I would say it is a defect.  I even have voltage headroom (supposedly the voltage limit is 1.09v on these) at like 1.05v when hitting the 400W power limit.  So my card NEVER goes beyond 1.05 or so, even if I lock the voltage higher in Afterburner.  This is because its hitting the 400W power limited sooner than the voltage limit.
 
I wish EVGA would release a BIOS with this fixed.  But, since they are no longer in the GPU game, they probably will do nothing.  I consider it to be false advertising.  The reason I got this was because of the stated power limit of 450W lending to better overclocking headroom.  Well, based on what I've seen others go through when discussing this with EVGA, it seems like pretty terrible and apathetic customer service.

I don't have any problem at all hitting 450 watts. It's not a BIOS problem. How about some more details on the rest of your system and what you're using to load the system?
MSI Kombustor stress test should load the GPU to 450 watts using the default stress test. Make sure you have Vsync and any frame limiters disabled. You should be seeing around 140 FPS with the default settings.

 You're right, it does appear that the program using to load the GPU is the difference.  I used Superposition and 3dMark and Furmark at various settings trying to get one that would load the GPU completely (basically setting resolution and graphics quality to ultra or extreme to draw more watts).  On previous cards (not 3080 ti), Superposition and 3Mark would push the GPU to its max, no so with this 3080 Ti.  Since it seemed to be capping out at 400w, I assumed I was having the same issue that others have stated, all else seeming to be exact or equal in terms of similar symptoms.
 
But when I used OCCT to load the GPU yesterday it pushed the power to 445w or so and held that for a while, so there is no real "limit" at 400w.  Maybe its just coincidence that the majority of stress test applications, benchmarks, and real-world gaming are only pushing this card to around 400w.  But I can confirm now that my card will reach 450w and hold it, with the right test -- I've only found OCCT to do that, but I will try Kombuster as you mentioned.

But in the game where it was limited to 400 watts, what was the "perfcap reason"? Was it "PWR"? If so, you can't blame vsync or computer setup or power supply or anything else other than the card itself. If vsync or framerate limiter was keeping the card from using more power, the perfcap reason wouldn't be "PWR".

Please verify the perfcap reason. It is the metric in this discussion which matters the most.

It isn't fair to use a power virus and then say that a power virus proves that the card is working as intended, when you game as intended with a perfcap reason of "PWR". The reason you bought the card was to game. It should balance and power limit well enough to reach its full advertised potential in the application intended. That is the expectation set forth.

Hemi likes to ignore the details when it comes to this matter. He loves proving that there is one situation where the card uses more power, even though that situation suits no one and the supposed explanation ignores all other details on the matter and is contradictory to what makes sense. Since you haven't verified to us the perfcap reason, I can't say for sure whether he is right or wrong in this instance, but as a general rule this is the same nonsense argument he makes.

ty_ger07
But in the game where it was limited to 400 watts, what was the "perfcap reason"? Was it "PWR"? If so, you can't blame vsync or computer setup or power supply or anything else other than the card itself. If vsync or framerate limiter was keeping the card from using more power, the perfcap reason wouldn't be "PWR".

Please verify the perfcap reason. It is the metric in this discussion which matters the most.

It isn't fair to use a power virus and then say that a power virus proves that the card is working as intended, when you game as intended with a perfcap reason of "PWR". The reason you bought the card was to game. It should balance and power limit well enough to reach its full advertised potential in the application intended. That is the expectation set forth.

Hemi likes to ignore the details when it comes to this matter. He loves proving that there is one situation where the card uses more power, even though that situation suits no one and the explanation ignores all other details on the matter.

I will need to look and see if there even was any perfcap reason.  I do not know.
 
But using OCCT to get the card to pull 450w does prove that its possible to hit 450w and sustain it, and that there isn't a hard power limit at 400w.  How would it allow a power virus to pull 450w but know to limit a game to 400w (presuming that game SHOULD be pulling more than 400w)?

mattschlosser

ty_ger07
But in the game where it was limited to 400 watts, what was the "perfcap reason"? Was it "PWR"? If so, you can't blame vsync or computer setup or power supply or anything else other than the card itself. If vsync or framerate limiter was keeping the card from using more power, the perfcap reason wouldn't be "PWR".

Please verify the perfcap reason. It is the metric in this discussion which matters the most.

It isn't fair to use a power virus and then say that a power virus proves that the card is working as intended, when you game as intended with a perfcap reason of "PWR". The reason you bought the card was to game. It should balance and power limit well enough to reach its full advertised potential in the application intended. That is the expectation set forth.

Hemi likes to ignore the details when it comes to this matter. He loves proving that there is one situation where the card uses more power, even though that situation suits no one and the explanation ignores all other details on the matter.

I will need to look and see if there even was any perfcap reason.  I do not know.
 
But using OCCT to get the card to pull 450w does prove that its possible to hit 450w and sustain it, and that there isn't a hard power limit at 400w.  How would it allow a power virus to pull 450w but know to limit a game to 400w (presuming that game SHOULD be pulling more than 400w)?

Because the card doesn't power balance properly; if you find the right power virus which loads up the right thing, you could reach the power limit before the power balance issue plays its role. It's been proven plenty of times, but Hemi won't look at anything other than his perspective.

We know that the card can use 450 watts, but often we see that power balancing problems causes one input to limit everything way before the total expected power limit.

Hemi's approach is that if he can show that one scenario doesn't exhibit the problem, it is no problem. When I push him on the subject of intended use, disregard of the other indications, and falling below expectations, his response is 'maybe that is the way it is supposed to work'. Yeah, right. Every other video card is history didn't exhibit this behavior, the problem exists primarily along lines of product segmentation, and EVGA had a special RMA program to specifically attempt to resolve this issue with the higher-end models due to overwhelming outcry, but ignored trying to improve the lower-end models which would cut into the performance marketing of the higher-end models.

IF your card is prematurely limited in performance due to an unexpectedly low power perfcap event, seeing your card reach full expected perfomance in one situation is not an indication that everything is fine. All it does is mock you about a level of performance which you will not obtain for your normal intended application.

ty_ger07

mattschlosser

ty_ger07
But in the game where it was limited to 400 watts, what was the "perfcap reason"? Was it "PWR"? If so, you can't blame vsync or computer setup or power supply or anything else other than the card itself. If vsync or framerate limiter was keeping the card from using more power, the perfcap reason wouldn't be "PWR".

Please verify the perfcap reason. It is the metric in this discussion which matters the most.

It isn't fair to use a power virus and then say that a power virus proves that the card is working as intended, when you game as intended with a perfcap reason of "PWR". The reason you bought the card was to game. It should balance and power limit well enough to reach its full advertised potential in the application intended. That is the expectation set forth.

Hemi likes to ignore the details when it comes to this matter. He loves proving that there is one situation where the card uses more power, even though that situation suits no one and the explanation ignores all other details on the matter.

I will need to look and see if there even was any perfcap reason.  I do not know.
 
But using OCCT to get the card to pull 450w does prove that its possible to hit 450w and sustain it, and that there isn't a hard power limit at 400w.  How would it allow a power virus to pull 450w but know to limit a game to 400w (presuming that game SHOULD be pulling more than 400w)?

Because the card doesn't power balance properly; if you find the right power virus which loads up the right thing, you could reach the power limit before the power balance issue plays its role. It's been proven plenty of times, but Hemi won't look at anything other than his perspective.

We know that the card can use 450 watts, but often we see that power balancing problems causes one input to limit everything way before the total expected power limit.

Hemi's approach is that if he can show that one scenario doesn't exhibit the problem, it is no problem. When I push him on the subject of intended use, disregard of the other indications, and falling below expectations, his response is 'maybe that is the way it is supposed to work'. Yeah, right. Every other video card is history didn't exhibit this behavior, the problem exists primarily along lines of product segmentation, and EVGA had a special RMA program to specifically attempt to resolve this issue with the higher-end models due to overwhelming outcry, but ignored trying to improve the lower-end models which would cut into the performance marketing of the higher-end models.

IF your card is prematurely limited in performance due to an unexpectedly low power perfcap event, seeing your card reach full expected perfomance in one situation is not an indication that everything is fine. All it does is mock you about a level of performance which you will not obtain for your normal intended application.

I ran the Superposition benchmark with GPU-Z sensors open -- you are right, it was PerfCap = pwr.  So even though it hovered around 400W the whole time, it was hitting the performance cap due to power when it should have another 50W of headroom.  So, what's the recourse? Will EVGA replace if you RMA?  Seems like you are correct and that our cards don't perform at the stated power limit according to what is expected.

Hi guys I know this is a bit old and with evga dead probably not many still care, but I realised I'm facing the same issue and would like to squeeze some more power out of my board. I notice that it doesn't really go above 400w even though it shows 99% usage in heavy games. With kombustor the max I got is 420w shown in GPU-Z. I have +100 on core and +1150 on mems, 118% voltage and 1.1 volts set in afterburner/precision x1. My card maxes out when under heavy load at around 70C. Is there any bios you would recommend over the XOC one for the EVGA 3080 FTW3 Ultra? I read something about an MSI bios, is it worth it? Does it increase performance?

"Hemi likes to ignore the details when it comes to this matter. He loves proving that there is one situation where the card uses more power, even though that situation suits no one and the supposed explanation ignores all other details on the matter and is contradictory to what makes sense. Since you haven't verified to us the perfcap reason, I can't say for sure whether he is right or wrong in this instance, but as a general rule this is the same nonsense argument he makes."

But you can sure run your yap. I realize this is old. But holy crap, it made me laugh out laugh.

Not wrong.

ty_ger07
Not wrong.

The only thing I was attempting to 'prove' was that the card would draw full power. That case was proven. The rest of your post was your usual yapping nonsense which you admit, proved nothing.
Your silly little rage fits are always wrong. Any questions?