If it isn't power limited and doesn't need more power, you can't just tell it to use more power. To make it use more power, increase utilization (if bottlenecked below 99% utilization [unlikely in this case], increase CPU performance), increase the core frequency, or increase the core voltage.

What is the "perfcap reason" in GPU-z sensor tab while the video card is under load?

12ianma
What should I look for and expect when investigating the specified reason?

You look at the reason, and they you tell us.  There's nothing more to it.  What should you expect?  It's impossible to say.  That's why I asked.  I expect that it won't say PWR, because I don't think that you are power limited.  But you could prove me wrong.
 
Once you know the reason it displays, you can make adjustments to try to minimize that being the reason it isn't boosting higher.  Until you know why it isn't boosting higher, you can only guess.  Guessing is an inefficient way to overcome an obstacle.

12ianma
Hey boss I more meant what codes pop up.

I'm not offended.  Just annoyed.  You want to argue about whether the card is being power limited, but provide no evidence that it is being power limited.  If you want to know all the possible perfcap reasons, you can get the answer in seconds just by using Google.
 
PerfCap Reasons [GPU-z] | Overclock.net
 
It really doesn't matter what possible reasons could be displayed.  It more matters what reason is displayed.
 
It sounds to me that the card is not power limited.  If you want to overclock it higher, I would recommend getting it colder.

I have a suspicion that you are not power limited, and that you are up against silicon lottery at this point, until you get it colder.  Adding voltage might help, but usually doesn't help until you first get it colder.  Some examples of getting it colder are placing the AIO radiator directly in front of an AC outlet.  Or, placing the AIO radiator in a bucket of ice water.  Either method would require insulating/sealing the card against condensation damage.

Nisco3000
Are they audible at all?

Look at the size of their hubs versus the length of their blades.  They are obviously designed to spin fast.  They must be quite loud.

Ok during a bench the 2 reasons I get from gpu-z are power and voltage reliability. Both being quite constant. I guess it is not putting out it’s maximum hardware voltage based on codes given. Any ideas?also back to my original question, does anyone do custom bios?

I cannot give it more core voltage, which is also odd I don’t pop the code for being at voltage limit. I modded the card’s power shunts correctly including the PCI slot one, as I was worried for the exact reason you posted.

I know it’s still drawing exactly the same total power despite it reading lower, as I have a milliohm meter and know it’s new total resistance at each shunt, and when the math is done the power consumed is exactly the same as before.

For example the new resistance total on my 2060 is ≈ 2.92 instead of 5.

5/2.92 is ≈1.712x what reading is. Max reading ever is 68-70% power usage, which 70x1.712 = 120% power which is the setting cap on the card lol. Also worth noting if I set power target on card lower, it will indeed drop to that. Like 100%, it will drop to reading lower than 70% and limit to its previous exact same actual consumption as 100% would be.

I know the mod is done correctly as it reads exactly as much lower as it should. I also know the solder is good because it’s been checked with my milliohm meter.


Basically despite sensors all reading wattage consumption as lower than actual, it still draws up to the exact same wattage it was allowed to. I don’t understand lol. If the sensors are not what determines the cards logic, then what does? It could also calculate based on voltages somewhere else as a shunt or power sense is just reading a voltage differential. However I would be surprised they would even implement a sense system if it also used something else as the final check.

I’m talking about not being at VOp. That’s why I want a bios that allows a higher voltage setting.

No one cares about these cards enough to have already made a bios, and I don’t personally know enough about the rest of the board to alter it further.

So I’m hard limited at current results without a bios or further voltage modification, which is what I started this thread for

Can you provide a GPU-Z sensor tab screenshot? It would be interesting to see the readings.

I am sorry about not believing you. I simply don't trust anything at face value. I believe you now, but I would still find the GPU-Z screenshot interesting.

I have volt-modded a lot of cards. Most didn't require a guide and were pretty easy to figure out. I even modded 2 laptops. The theory is simple for analog VRM with a FB signal. For smart digital VRM, it is more complicated and I have never had a chance to mod one.

Interesting. It is mostly limited by Vrel, so increasing the core voltage is a logical pursuit.

Without the card in hand, I guess I can't be too helpful. I can only give general guidance. I am sorry that I haven't been helpful.

If you are able to determine that they are smart power stages based on part number, you may find information about how to modify those power stages for voltage and power limit. It wouldn't necessarily even need to be related to the same video card.

On the other hand, if you are able to determine that they are just dumb analog MOSFETs, volt modding them should be really simple. But that wouldn't explain how the card is able to figure out its power draw. If they are dumb MOSFETs, you should find this thread helpful:
https://forums.evga.com/V...Volt-Mods-m572578.aspx
Post #22 especially...