HELP: Dont know whats wrong with my card

[Breathtake]

Card in question = Asus RTX2080Ti Turbo : https://www.techpowerup.com/gpu-specs/asus-turbo-rtx-2080-ti.b6281

A strange thing I notice was my cards bios linked to 2 different bios through GPU-Z lookup button,

And I should have the one on the right (TURBO)'s bios, but when I look on the details of two bios, I see that the turbo's bios has a power range of +20% meanwhile the asus dual have a power range of +12%, and when I was going to overclock my card I noticed in MSI afterburner that I only had a range of +12% (112%) which leads me to believe that from the get-go I had the "wrong" bios for my card. And thats where I encounter my first problem which is it would not let me flash the bios to the correct turbo one, even with the nvflash id mismatch version (I've tried several versions of nvflash and the normal one aswell) and steps as seen below

So instead of flashing to the "correct" bios which I was unsuccessful I tried flashing to this bios : https://www.techpowerup.com/vgabios/208274/palit-rtx2080ti-11264-190131

Which worked, I could crank the slider to +24% (124%) and start overclocking, but somehow the performance in my benchmarks and games went down alot, for example the Superposition Benchmark with 4k optimized settings before flashing or tinkering, was at aprox, 10.000-11.000 score, and with only regular OC without flashing, I got a max of like 12.400 score with the +12% of my original bios (112% power limit) and then after I tried flashing and had the power limit on 124% with the new Palit bios, (so the same overclock but with higher powerlimit) I got a score of between 6000-7000 points, which I got worried about, and in for example Total War Warhammer 2 if I launch a custom battle and just dont move after loading into a battle, I would initially get 129-131fps, and then almost instantly it would continuously tick down to around 95fps, slow ish, and after a lot of back and forth I realised this strange thing where if I limit my power setting in msi afterburner (power limit) to 80% I would get a stable fps of around 123-125fps (and low 9500-10.000 in superposition) and not drop below that, but the stability of the gpu was worse, if I would go above 80% power limit I would get lower fps and lower benchmark scores, and same goes with below 80%, so either how, it would still be worse than my original bios, so I scrapped that idea and tried to run my backup.rom bios that I had saved before my first flash, but then somehow the same thing started to occur on my original bios also, if I give the gpu more power (increase the power limit above 80% or below) the fps or benchmark score would greatly decrease. Which I cannot understand how giving it more "juice" / power decreases the performance, and before touching any of this my temp was not going above 63C, and after all this my temps peak at 67C. And neither do I understand how this problem would persist after flashing to my backup.rom. And also it would not matter if now I overclocked it or not, it perform worse with any setting other than 80% power limit

I've tried to solve this for about 3-4 full days and I cannot figure out why and how this can be fixed.

So there's 2 problems I would like to try to get solved :

First : fix the issue with the weird power limit and decreased performance

Secondly : getting the "correct" bios for my card as in the Turbo's Bios

And if I should / need to contact for example Asus for this issue.

Ty for any help in advance.

 

[Caring1]

The card had the correct Bios, you flashed a wrong one to it.
Flash back to the original Bios.
Bios can be used over several cards and manufacturers, it does not make the card the same or mean you have the wrong one.
When looking for a Bios that is suitable, compare Manufacturer and subsystem ID's and ensure they match.

 

[Breathtake]

Hi, ty for replying.
I have tried reverting back to the original bios with the same steps and commands as the flash I did but put Backup.rom instead, but the issues (decreased performance, power limit issues) persisted and when I had the "correct" bios as mentioned the issues persisted so I checked the asus dual 2080ti card bios and the subsystem etc matched, and I thought I already had that bios enabled, but the bios versions differed, so I tried flashing to that card just to see if anything would change now that the ID's matched, and unfortunately it didnt change anything, all the previous issues was still there,

Any other thoughts on what to do?

Also I find it weird that the Turbo card (exactly what I have) differs in subsystem ID's and not being able to flash to it, as it litterly is the exact card I have

 

[purecain]

Card in question = Asus RTX2080Ti Turbo : https://www.techpowerup.com/gpu-specs/asus-turbo-rtx-2080-ti.b6281

A strange thing I notice was my cards bios linked to 2 different bios through GPU-Z lookup button, View attachment 170392

And I should have the one on the right (TURBO)'s bios, but when I look on the details of two bios, I see that the turbo's bios has a power range of +20% meanwhile the asus dual have a power range of +12%, and when I was going to overclock my card I noticed in MSI afterburner that I only had a range of +12% (112%) which leads me to believe that from the get-go I had the "wrong" bios for my card. And thats where I encounter my first problem which is it would not let me flash the bios to the correct turbo one, even with the nvflash id mismatch version (I've tried several versions of nvflash and the normal one aswell) and steps as seen below

View attachment 170396

View attachment 170398

So instead of flashing to the "correct" bios which I was unsuccessful I tried flashing to this bios : https://www.techpowerup.com/vgabios/208274/palit-rtx2080ti-11264-190131

Which worked, I could crank the slider to +24% (124%) and start overclocking, but somehow the performance in my benchmarks and games went down alot, for example the Superposition Benchmark with 4k optimized settings before flashing or tinkering, was at aprox, 10.000-11.000 score, and with only regular OC without flashing, I got a max of like 12.400 score with the +12% of my original bios (112% power limit) and then after I tried flashing and had the power limit on 124% with the new Palit bios, (so the same overclock but with higher powerlimit) I got a score of between 6000-7000 points, which I got worried about, and in for example Total War Warhammer 2 if I launch a custom battle and just dont move after loading into a battle, I would initially get 129-131fps, and then almost instantly it would continuously tick down to around 95fps, slow ish, and after a lot of back and forth I realised this strange thing where if I limit my power setting in msi afterburner (power limit) to 80% I would get a stable fps of around 123-125fps (and low 9500-10.000 in superposition) and not drop below that, but the stability of the gpu was worse, if I would go above 80% power limit I would get lower fps and lower benchmark scores, and same goes with below 80%, so either how, it would still be worse than my original bios, so I scrapped that idea and tried to run my backup.rom bios that I had saved before my first flash, but then somehow the same thing started to occur on my original bios also, if I give the gpu more power (increase the power limit above 80% or below) the fps or benchmark score would greatly decrease. Which I cannot understand how giving it more "juice" / power decreases the performance, and before touching any of this my temp was not going above 63C, and after all this my temps peak at 67C. And neither do I understand how this problem would persist after flashing to my backup.rom. And also it would not matter if now I overclocked it or not, it perform worse with any setting other than 80% power limit

I've tried to solve this for about 3-4 full days and I cannot figure out why and how this can be fixed.

So there's 2 problems I would like to try to get solved :

First : fix the issue with the weird power limit and decreased performance

Secondly : getting the "correct" bios for my card as in the Turbo's Bios

And if I should / need to contact for example Asus for this issue.

Ty for any help in advance.

It happens to us all m8. Virtually every enthusiast on the forum once flashed the wrong bios on launch day which bricked their cards me included. Until w1zard saved the day.
It looks like youve been really unfortunate. This kind of problem with the original bios settings not taking effect as they did before would lead me to think that the power portion of the card could be/is malfunctioning.
Do you know how to test the voltage on your card itself.
I would definitely contact asus for advice. Or at least use the forum and post up whats happened there so that if anyone else who has bought the same card but solved the issue can offer help.
Ask asus to send you a bios in the mean time. ahh i see youve downloaded another bios with the same issue. Go back to your original bios for now.

 

[Breathtake]

If you mean testing as in outside the pc/softwares and with tools, sadly I dont know how to test it. But I will take your advice and contact Asus for the correct Bios.

But have I potentially broken the power portion of the card? making it a hardware issue?

 

[Nike_486DX]

Did u use DDU to reinstall the drivers? Wipe them nicely and then make a clean install. Then test everything again. What are the temps, voltages and clocks?

PS
Next time u would start tinkering with a graphics card it would be great if u could set the temp/clocks/tdp/voltage/scores baselines first before going to modify anything.
That way u would easily troubleshoot and have 0 placebo effect

 

[Breathtake]

Yes I used DDU to reinstall the drivers but no difference. The temps as stated have gone up from 63C max to 67C max, and the voltages now only hovers around 750-850 ish and wont go above it, if i put on max power limit the voltage will spike to max and then come back down to 800 ish again, before I did any of this it usually stayed around 1050 and or high 900's and clocks before was around 2145-2115, anything above 2000, but now they sit around 1500-1600 with 80% powerlimit which i figured out for now is the best result i can get, if i were to go max power limit the clocks would go up to 1700-1820 ish but with decreased performance and fps sadly

Edit: and in my first post i also mentioned my scores in Superposition benchmark, from around 12500 ish score down to 6000-7000 points with same settings, and with the bad bios, if i decreased the power limit to 80% i managed to get a score of around 10.000-11.500, but way less stable and would crash in games

And these bad effects followed even though i flashed back to my Backup.rom

 

[Nike_486DX]

Does ur card have 2 bioses? If yes then maybe u might want to try the other (unflashed) bios to test the performance.

Try to flash from dos...
It is weird that despite of lower tdp u are getting higher temps, something is going really wrong with the fan curve or the driver.

+ reinstall msi afterburner (dont keep the profiles, restore it to 0).

 

[nguyen]

so from your system specs your 2080 Ti is under water ?
If so this issue is not related to BIOS at all, it's because of the TIM application, a part of the GPU die doesn't make contact with the waterblock, 67C is just too high for watercooling.
It is possible that the TIM is dripping down when you have your GPU in the vertical position, at high temperature the TIM just liquidfy and flown down.
Just repaste your GPU with a higher viscosity TIM, i recommend Thermalright TF-X for this kind of situation.

 

[Breathtake]

Okey, so first, i dont think i have a dual bios since i cannot see a switch around the gpu,
and i would have to search up how to flash gpu from dos,
and lastly, I used Thermal Grizzly's Kryonaut on the gpu, and i agree the temps should be lower, but could it really be a problem connected to TIM as you said, since i had no issue whatsoever before i flashed my card to the wrong bios? and its not thermal throttling?

 

[nguyen]

Okey, so first, i dont think i have a dual bios since i cannot see a switch around the gpu,
and i would have to search up how to flash gpu from dos,
and lastly, I used Thermal Grizzly's Kryonaut on the gpu, and i agree the temps should be lower, but could it really be a problem connected to TIM as you said, since i had no issue whatsoever before i flashed my card to the wrong bios? and its not thermal throttling?

I always knew you would be using Kryonaut LOL, that TIM has low viscosity and not suitable for vertical mounted GPU.
Well the temperature you are reading is an average temperature from a lot of thermal sensors throughout the die, a part of the die without contact to the waterblock could be running at 100C while the temperature reading is only 60C.
This problem show itself when you increase the power limit, leading to higher thermal output and that part of the die that was just at the edge of stability before is now pushed past that.
Just try repasting first to see if the stuttering stop.
BTW I have the same Asus 2080 Ti Turbo underwater , with the horizontal orientation I have no issue using Conductonaut, temp under maximum load is 44C.

 

[Breathtake]

I've contacted Asus for their bios version and waiting for a reply, and I ordered the Thermalright TF-X TIM from Amazon.co.uk so I will have it on wednesday,

Will update as soon as possible.

 

[nguyen]

I've contacted Asus for their bios version and waiting for a reply, and I ordered the Thermalright TF-X TIM from Amazon.co.uk so I will have it on wednesday,
Will update as soon as possible.

There is nothing wrong with BIOS flashing, if there were something wrong your card wouldn't be working.
Now let me explain a little bit more with the power limits:
Basically you have the Non-A 2080 Ti chip with maximum power limit of 280W with stock BIOS, flashing it with the Palit one give you 310W PL. That is the highest PL you can get with non-A chip. Furthermore you can't flash non-A chip BIOS with A chip BIOS (GPU MISMATCH in Nvflash)
non-A chip will have device ID: 10DE 1E04
A chip will have device ID: 10DE 1E07
The Palit BIOS is the best BIOS for non-A chip, as I am using that one also, there is no need to reflash to your stock BIOS, just repaste the GPU and it will be fine.

 

[Breathtake]

But why does it say on here on techpowerup on the turbo card page device id: 10DE 1E07 when its not an A chip? it doesnt have an A on the gpu variant, the only difference i see is that on the turbo card it says TU102-300-K2-A1 and on the other gpus (Dual and or Palit) it says TU102-300-K1-A1,


And also, you have the exact same card as me and flashed to the bios that i did, but it worked for you and not me? I kind of refuse to believe that the TIM problem happened as soon as i flashed to the Palit bios and it was just coincidence,

And lastly, before you flashed to the palit bios, did you have the +20% power limit as stated on the turbo bios or like me +12%?

 

[nguyen]

But why does it say on here on techpowerup on the turbo card page device id: 10DE 1E07 when its not an A chip? it doesnt have an A on the gpu variant, the only difference i see is that on the turbo card it says TU102-300-K2-A1 and on the other gpus (Dual and or Palit) it says TU102-300-K1-A1,
And also, you have the exact same card as me and flashed to the bios that i did, but it worked for you and not me? I kind of refuse to believe that the TIM problem happened as soon as i flashed to the Palit bios and it was just coincidence,
And lastly, before you flashed to the palit bios, did you have the +20% power limit as stated on the turbo bios or like me +12%?

LOL yeah Asus was being a dick, initially the Turbo version use A chip but later Asus switched to non-A chip (Yeah I also got burned), TPU just didn't update their specifications page.
A chip will have marking like this TU102-300A-K3-A1 (the A after 300)

Before flashing the Palit BIOS I also have +12% PL (280W) and after flashing it is +24% (310W). Try using GPU-Z as that would give you the full wattage value.

Well your 2080 Ti can still run with a bad thermal paste coverage, it just doesn't run very well, stuttering all over the place, the reasons are as I explained before.
Here is a techtuber do did a crap job of screwing down the waterblock, which later killed his 2080 Ti, 1/3 of the GPU die don't even make contact with the waterblock.

 

[Breathtake]

So I did what you said and I took apart my card and litterly it looked almost the same as that pc centric's card, exactly where his big "bump" was on the card where it was bent upwards, exactly on my card, and the pressure points (as in his and mine thermal pads), looked identical, where the middle thermal pads where not even touched, I replaced the kryonaut as a temporary measure, and tried to screw it in a way so the bump would be as tiny as possible, and i installed the gpu in the regular horizontal slot and not vertical since i dont have the correct TIM, and i fear no doubt i have in some way damaged my card in the same way as pc centric, as just little before i saw your post and i was watching YT, the browser just like started flickering a bit, but to my suprise after reinstalling my gpu after the adjustments "no bios tinkering" i tried the same benchmark with clock speeds and touched nothing, immediately i saw the power limit actually hitting 90-99% where it never did before, and clocks were way up, and i actually got a score of 11640, so i think the performance is back and before my temps were hitting 67C but now my peak temp after 2 runs in the benchmark my peak temp was 46C which i was shocked by, but i assume you have the palit bios with overclock and your max was 44C,

Which leads me to believe i could use some more pressure on the thermal pads connecting to the mem, but the screws wont do it, it seems, do you have any suggestion as to how to get additional pressure around the gpu chip where the 3 thermal pads are located?

and ty for replying, you litterly saved my card bro

 

[nguyen]

well if you are not experiencing any stutter during game I think your GPU temp is as good as it can get. I'm using Conductonaut TIM now, before I was using Kryonaut and max GPU temp was 48. Conductonaut is not safe for vertical mount GPU though.

If the thermal pads are not making good contact with Mem then you can buy thicker pads, refer to the waterblock manual to find out which thickness was the original thermal pads.

Though I think everything should be okay after your waterblock remounting.

 

[Breathtake]

Okey will remember that, and thanks for the amazing help and saving my card, i also tested the palit bios again after the remounting, and i got in the same benchmark a score of 13.432, which is a new record for me, and max temp 50C