This freaks me out

[Skinny48]

To all crypto miners on here for how long can a GPU lasted before they give up the ghost? I'm talking about well cared for GPUs that mine 24/7 with 70°C max, the reason why I asked is because I saw a video on YouTube about a GPU repairer that have over 150 GPUs to fix and they are mostly RTX3080 and 3090 with few 3070 or is heat not the only thing that can kill a GPU?

[1714825328]

1714825328

[1714825328]

1714825328

[RentGPU]

6 years mining and still , i can tell you as long as you stay away from OC or modding and know your gpus limits and never push them for some little extra bucks they will take care of you for years no worries

[sxemini]

5 Years with high mem OC no problem for my cards. I lost 1 of 100 cards in 5 years.

[Lafu]

Depends on a lot of things !
The way you maintenance them like cleaning from time to time to get away the dust from the fans.
What kind of Room temperature it has where the Gpus are running , and how they get cooled.
And for sure as the others already have written how much you tweak them and what.
If you take care of them they will tank you it. I am also mining for years now with Gpus.

[swogerino]

I have some old Rx 470 mining since 2016 and they are running fine,it means 5 and half years running full time 24/7 with temperatures during hot summer at 75 degree Celsius max.In winter they run at 47-50 max.I still mine Ethereum Classic with them and they are still working great.

The fact that the cards to repair were mostly 3090 and 3080 means they have not been taken good care of,if they follow the best mining settings which are all over the Internet for each and every card and also do not push the cards they can run really well.I think it is 100% users fault for this faulty 3090 and 3080,no need to worry if you keep good care of the cards.

[batsonxl]

I have some old Rx 470 mining since 2016 and they are running fine,it means 5 and half years running full time 24/7 with temperatures during hot summer at 75 degree Celsius max.In winter they run at 47-50 max.I still mine Ethereum Classic with them and they are still working great.

The fact that the cards to repair were mostly 3090 and 3080 means they have not been taken good care of,if they follow the best mining settings which are all over the Internet for each and every card and also do not push the cards they can run really well.I think it is 100% users fault for this faulty 3090 and 3080,no need to worry if you keep good care of the cards.

GDDR5 memory dont get so hot like GDDR6. In my measuring GDDR5 get nearly 10C more that shows gpu temps. So you max got 75C it means your memory at 85C.Ofcourse it changes to different brand but roughly it is 10C more.
But in GDDR6 is totally different they run faster and hotter.You must keep your GDDR6 memory below 95C, otherwise it will not last long like old cards.

[batsonxl]

To all crypto miners on here for how long can a GPU lasted before they give up the ghost? I'm talking about well cared for GPUs that mine 24/7 with 70°C max, the reason why I asked is because I saw a video on YouTube about a GPU repairer that have over 150 GPUs to fix and they are mostly RTX3080 and 3090 with few 3070 or is heat not the only thing that can kill a GPU?

You see it alot because most of them newbies that have no idea what they doing, dont know about memory temps.
Ihave friend who bought brand new card 3080. He called me and asked : Why my cards hashrates gets lower.It took me 5min explain how to see memory temps, Card was running under 110C for whole week. imagine some people running those cards at this temps for months.

[FP91G]

To all crypto miners on here for how long can a GPU lasted before they give up the ghost? I'm talking about well cared for GPUs that mine 24/7 with 70°C max, the reason why I asked is because I saw a video on YouTube about a GPU repairer that have over 150 GPUs to fix and they are mostly RTX3080 and 3090 with few 3070 or is heat not the only thing that can kill a GPU?

I can't say how reliable the 30xx series from Nvidia is, but their 10xx series was very good and has been mining for almost 5 years.
The video card will not burn out from overheating, it will turn on protection and your hash rate will drop. Keep track of the main indicators of the video card and think about installing voltage stabilizers

[Didz]

To all crypto miners on here for how long can a GPU lasted before they give up the ghost? I'm talking about well cared for GPUs that mine 24/7 with 70°C max, the reason why I asked is because I saw a video on YouTube about a GPU repairer that have over 150 GPUs to fix and they are mostly RTX3080 and 3090 with few 3070 or is heat not the only thing that can kill a GPU?

Here's a very good explanation about Life of Components from the book "A quick guide on building a GPU Mining Rig" written by Ermiya Karan:



In my mining farm I set my commercial AC to 73 Fahrenheit (23 Degree Celcius) so my GPU cores stay around 60°C.

I also clean my mining rigs once every 3 months with a special blower and brushes for electronic components.

[anu1908]

that's a huge number, but you should compare it to the number of cards sold over the last few months. as far as i can remember, the rma/broken rates of those new cards are still acceptable, not as disastrous as you might think it is. if you want to be safe then buy on a trusted retailer with great support, so you don't have to rely on third-party repairer.

[Looper_U]

Graphic cards don't die easily unless you do something stupid when using them, the first time I killed a GPU was when I build my first rig, I mistakenly turned the riser the wrong way and there was smoke everywhere, apart from this no GPU dies on my watch since 2016

[FP91G]

To all crypto miners on here for how long can a GPU lasted before they give up the ghost? I'm talking about well cared for GPUs that mine 24/7 with 70°C max, the reason why I asked is because I saw a video on YouTube about a GPU repairer that have over 150 GPUs to fix and they are mostly RTX3080 and 3090 with few 3070 or is heat not the only thing that can kill a GPU?

Here's a very good explanation about Life of Components from the book "A quick guide on building a GPU Mining Rig" written by Ermiya Karan:


In my mining farm I set my commercial AC to 73 Fahrenheit (23 Degree Celcius) so my GPU cores stay around 60°C.

I also clean my mining rigs once every 3 months with a special blower and brushes for electronic components.

Condensers do not heat up to these temperatures because they do not come into contact with the radiators.
And at a temperature of 80 degrees Celsius, Capacitors will live for more than 10 years, which is a very good indicator, because the video card loses its relevance after 5 years

[HardCore12V]

I prefer GPUs with three fans on them simple because of cooling but the most important thing is keeping the cards clean from time to time, I do have a friend who use GPUs with single fans and they have been running since 2017 none stop, it all depends on care

[FP91G]

I prefer GPUs with three fans on them simple because of cooling but the most important thing is keeping the cards clean from time to time, I do have a friend who use GPUs with single fans and they have been running since 2017 none stop, it all depends on care

Modern video cards have new polymer capacitors that can withstand elevated temperatures

RTX 2060

Photo source
https://oclab.ru/obzory/videokarta-palit-geforce-rtx-2060-gaming-pro-oc-obzor-test-razgon/

But my PX 470 video cards with conventional solid-state capacitors have been working for a very long time.

[adaseb]

Its not as easy to kill as GPU as you think. Maybe out of 100 GPUs, I had 10 or so that failed. And 5 of those were user error and not the fault of the card. I basically was doing a thermal paste job and the heatsink was stuck and I guess I pulled too hard and broke the solder joints. The other 5 had some form of MOSFET or capacitor melting. However this is extremely rare and was years of mining. Most were older GPUs like R9 280X which ran at much higher temps.

What is going to happen most likely is your fans will start to die soon. And you will need to either replace with a new one or lubricate the bearing yourself. Then the GPU can keep on running again.

That video on Youtube was mostly for views. People create these projects just to get views and it doesn't mean the product they are repairing is unreliable. He just bought a bunch of bad GPUs on purpose.

[FP91G]

I have tried lubricating fans that spin poorly, but this is not a solution to the problem.
As a rule, damage is noticed at the last stage, when replacement is required.
Standard 80x80 fans perfectly replace factory fans if you install 3 80x80 fans on the radiator instead of two factory ones

[Gorosden]

It's really easy to monitor GPU, this can prolong their lifes and also make sure you have thermal pastes and pads at home just in case, I have thermal pads and pastes and also spare fans but none of my GPUs ever reached 70°C when mining not even 68°C, to maintain this temp I always get rid of dusts from the fans every month