GTX 1080 ti crashes unless underclocked

[izy]

It seems more like a power problem to me than degradation because its only working if both are set lower but it could be either way, can you try to lower your power limit and see if its working like that?

 

[Deleted member 182555]

It's very unlikely it's any caps. This symptom is consistent with spent GPU ASICs. I've got a GTX 780 that does the exact same thing.

Heat and voltage over the years will damage the silicon and it will lose its capacity to operate at higher speeds due to electromigration.

This phenomenon will become more pronounced with smaller lithography chips as time goes on.

how do you know? are you a fortune teller?
capacitors age and over time and eventually behave like resistors. It's impossible to know without checking them one by one with an ESR meter. I've already found some that are damaged and would have been impossible to recognize by visual inspection.
I advise you to be less of a professor or doctor.
with a simple tester it is also possible to check voltages under load to see how they behave.

 

[Dr. Dro]

how do you know? are you a fortune teller?
capacitors age and over time and eventually behave like resistors. It's impossible to know without checking them one by one with an ESR meter. I've already found some that are damaged and would have been impossible to recognize by visual inspection.
I advise you to be less of a professor or doctor.

Years of experience. Solid-state capacitors such as those used in these high-end graphics cards shouldn't go bad after only 6 years, even if they were operating at their rated temperature through the entire lifetime of the device. There's significant redundancy in a high-end GPU's board so that a weakened capacitor or two won't cause it to malfunction in stock, failsafe settings.

What you're seeing here is a symptom consistent with ASIC degradation due to electromigration stress, GPUs that have operated very hot and under high current for extended periods of time. OP bought it 6 months ago off eBay, and I betcha this card mined more than its fair share of ETH in the past 4 or 5 years.

 

[Deleted member 182555]

Years of experience. Solid-state capacitors such as those used in these high-end graphics cards shouldn't go bad after only 6 years, even if they were operating at their rated temperature through the entire lifetime of the device. There's significant redundancy in a high-end GPU's board so that a weakened capacitor or two won't cause it to malfunction in stock, failsafe settings.

What you're seeing here is a symptom consistent with ASIC degradation due to electromigration stress, GPUs that have operated very hot and under high current for extended periods of time. OP bought it 6 months ago off eBay, and I betcha this card mined more than its fair share of ETH in the past 4 or 5 years.

still with this story? things are checked for a correct diagnosis... the rest is just talk.
I'm starting to think your famous experience is worth less than nothing.
I remember you are the one who wanted to make known about biosmods on pascal and turing talking about Socrate.

 

[Dr. Dro]

still with this story? things are checked for a correct diagnosis... the rest is just talk.
I'm starting to think your famous experience is worth less than nothing.
I remember you are the one who wanted to make known about biosmods on pascal and turing talking about Socrate.

You were the one who made those claims. Frankly I never believed you, but I gave you the benefit of the doubt and you ran away a complete no show. You never proved us that you had a working Pascal BIOS mod. Still waiting on that, btw

 

[Deleted member 182555]

You were the one who made those claims. Frankly I never believed you, but I gave you the benefit of the doubt and you ran away a complete no show. You never proved us that you had a working Pascal BIOS mod. Still waiting on that, btw

I didn't run away anywhere...
I can easily prove it in 1000 ways both on pascal and turing without revealing how to do it ....
I happily accept challenges from you professors and doctors of Nothing.
I wrote to you privately... you organize this challenge so we'll end this story too.

It seems more like a power problem to me than degradation because its only working if both are set lower but it could be either way, can you try to lower your power limit and see if its working like that?

I believe a gpu repairman could easily diagnose what in the power supply is malfunctioning definitely some worn component.

 

[Deleted member 182555]

nothing can be organized with the doctor... he doesn't collaborate... I asked him to provide me some personalized parameters
so that bios shown by me was custom and not crossflash therefore incontestable but nothing missing does not respond...
discussion ends here...

 

[jrx02]

@kamakashi
i literally made an account to comment on this post.

I experienced an almost identical situation as this. I have a 1080ti as well. I believe the driver before this latest one messed mines up. Same as you I tried my gpu on multiple pcs. I know it was my gpu because my other 1080ti worked well. That driver crashed my pc and the only way to run my pc is to underclock it. I almost gave up but I was able to get mines working again. I'll list the things I did that got mines to work again.

*if you haven't replaced thermal paste and pads for a while I suggest you do that!!*

#1: Even with the original vbios I reflashed my gpu again with the same vbios.
Find your gpus vbios online, download and flash it. You'll have to Google how to do it unless you already know how to.
#2: Once that's done, I reset my motherboards bios. What I did was take out the cmos battery with everything unplugged went to make a peepee and then put it back on and I also re updated my mobo bios. (Not sure if re updating helped but I did it anyway to make sure my mobo is cleaned from covid)
#3: (I really didn't wanna do this) But I had to do a clean reinstall of windows 10, took me hours since i had 2tb of important stuff.
#4: After that, once windows 10 and all the updates are done, make sure you install all drivers for your motherboard. You'll have to find your motherboards make and download them from their site. (*I also signed into Microsoft Store and made sure to update everything in there as well. Idk if that does anything but I did it anyway)
#5: Once I got all that done I went over to nvidia website and downloaded the latest driver that's out now. (Not sure if this makes a difference, but I only installed dlthe driver. I did not install GeForce experience)
#6: After that I installed msi afterburner and my gpu was back to working like it was before. I was able to run it at stock speeds and overclock it without any crashes.
#7: *if you're using display port 1.4 you'll also wanna download that and update your 1080ti before installing nvidia latest drivers*

I almost lost hope! I thought it was the psu but for some reason this is what I did!

 

[eidairaman1]

Sorry but Too many wuds bork the os.

 

[Deleted member 182555]

I'm revisiting this old post where we talked about semi-dead GPUs...
I have a dying 1650 that has worn out a lot.
I measured voltages with tester they are 0.2v lower than they should be.
on the board I found the melted Thermalpad. temperatures are also higher than usual.
vrms are no longer working as they should or something is causing resistance.
this gpu seems to now support only stock frequencies and no longer likes overclocking.

I'll try to do a permanent undervolt and see how it evolves.

 

[Space Lynx]

much like real life, the older we get, the more underclocked we need to be.

 

[Deleted member 182555]

I actually caused it, it's my fault. imprudent voltage experiments...

 

[Shrek]

Ouch; I feel your pain.

 

[Space Lynx]

I actually caused it, it's my fault. imprudent voltage experiments...

happens to all of us that like to tinker. i ruined my gtx 1070 laptop for 4 months one time, it wouldn't boot i gave up on it. (liquid metal gone wrong)

then one day i took an electric toothbrush to it with some isporpyl alcohol, soft bristle toothbrish high speed electric toothbrush, ran over the entire area of the laptop circuit area, and voila she worked just fine after that. lmao

 

[Deleted member 182555]

Now it's working again without any problems. the last move was a --fullrecovery of original bios with .bnv file.
or it's a coincidence or he rested after a period of stress.
everything seems to be back to normal... maybe the malfunction is intermittent? very strange.

 

[Deleted member 182555]

placebo effect... unfortunately the card is damaged and no longer reaches the previous frequencies and temperatures.

I'm missing about 100 mhz and 15w, I made a permanent undervolt to 0.9v@1962mhz and removed zerofan to keep the temperatures under control. It looks like it passes the stress test.

I'll keep an eye on it for a couple of weeks to see if it freezes again and if it gets worse over time, after which I'll park it somewhere.
I'd like to understand what was damaged but given the low value I don't think it makes sense to spend money on it.

 

[Lew Zealand]

placebo effect... unfortunately the card is damaged and no longer reaches the previous frequencies and temperatures.

I'm missing about 100 mhz and 15w, I made a permanent undervolt to 0.9v@1962mhz and removed zerofan to keep the temperatures under control. It looks like it passes the stress test.

I'll keep an eye on it for a couple of weeks to see if it freezes again and if it gets worse over time, after which I'll park it somewhere.
I'd like to understand what was damaged but given the low value I don't think it makes sense to spend money on it.

View attachment 322750

Looking at the 72C temps, if you're interested in squeezing a bit more extended life out of that 1650, maybe take another 50-100 Hz off with associated undervolt and hopefully the degradation will be reduced. Yeah giving up any performance on a lower end card is tough but I'm thinking on the card's longevity.

 

[Deleted member 182555]

Looking at the 72C temps, if you're interested in squeezing a bit more extended life out of that 1650, maybe take another 50-100 Hz off with associated undervolt and hopefully the degradation will be reduced. Yeah giving up any performance on a lower end card is tough but I'm thinking on the card's longevity.

gpu is damaged I had to undervolt because temps are high without it and also it freezes randomly. before I was at 2055/2070@1v 75°. after doing some voltage experiments it started to give problems. I replaced melted termalpads on vrms but I didn't get speed and temps back.
If it still gives problems I'll try lowering it further.
ps. zerofan is bullshit, once the card warms up it always stays at 60° instead with 1200rpm the fan can barely be heard and the gpu drops to 30°

 

[und3rground]

1080TI MSI Gaming X 11GB bought in 2020 from a mining guy along with Corsair cs850m PSU. All good these 3+ years. Since last month a few of my most played games are behaving funny ( garage crash World of Tanks holiday ops patch, border hang Dota2, savegame load crash Witcher 3 on ultra settings). PSU going as low as 11.1v on 8pin connector wasnt the main cause. My new 850w Seasonic providing 11.9v didnt fix Witcher3 savegame load crash to desktop on ultra.
Windows, ram, ssd or GPU issue?

 

[RJARRRPCGP]

I had to undervolt because temps are high without it and also it freezes randomly.
ps. zerofan is bullshit, once the card warms up it always stays at 60° instead with 1200rpm the fan can barely be heard and the gpu drops to 30°

The symptoms could well be bad caps. Especially bad caps in bronze-rated PSUs. With the original caps on a bronze, you will be lucky to get 6-7 years before they are bad.
I consider myself grateful that the caps in my 2008 FSP 500W white PSU were fine in 2016, IIRC! But, much of the time, the PSU was used with air <70 F.

 

[izy]

1080TI MSI Gaming X 11GB bought in 2020 from a mining guy along with Corsair cs850m PSU. All good these 3+ years. Since last month a few of my most played games are behaving funny ( garage crash World of Tanks holiday ops patch, border hang Dota2, savegame load crash Witcher 3 on ultra settings). PSU going as low as 11.1v on 8pin connector wasnt the main cause. My new 850w Seasonic providing 11.9v didnt fix Witcher3 savegame load crash to desktop on ultra.
Windows, ram, ssd or GPU issue?

Have you tried to underclock your GPU/ GPU VRAM? Or set a lower power limit? I would start from here, also you shoud reset your bios to default and test if that fixes any problems.

 

[Assimilator]

1080TI MSI Gaming X 11GB bought in 2020 from a mining guy along with Corsair cs850m PSU. All good these 3+ years. Since last month a few of my most played games are behaving funny ( garage crash World of Tanks holiday ops patch, border hang Dota2, savegame load crash Witcher 3 on ultra settings). PSU going as low as 11.1v on 8pin connector wasnt the main cause. My new 850w Seasonic providing 11.9v didnt fix Witcher3 savegame load crash to desktop on ultra.
Windows, ram, ssd or GPU issue?

Why would it be anything but an issue with the GPU?

 

[und3rground]

Why would it be anything but an issue with the GPU?

Nvidia tech admin in forum asked for HW info logs and noticed 11.4v supply from PSU. He said psu is pos and needs replacing.

Have you tried to underclock your GPU/ GPU VRAM? Or set a lower power limit? I would start from here, also you shoud reset your bios to default and test if that fixes any problems.

Will try this next, tho I dont have much xp on it..

 

[izy]

Nvidia tech admin in forum asked for HW info logs and noticed 11.4v supply from PSU. He said psu is pos and needs replacing.


Will try this next, tho I dont have much xp on it..

You can use MSI Afterburner , its pretty easy. My PSU does the same goes down to about 11.4 on GPU under load and it works just fine, i thought that might be my problem too but it wasnt.

 

[und3rground]

You can use MSI Afterburner , its pretty easy. My PSU does the same goes down to about 11.4 on GPU under load and it works just fine, i thought that might be my problem too but it wasnt.

What was the problem in your case?