How long will my gpu last?

[Ludwish]

3 years?

I have a few HD 7990. They are already 4 years old. I do not stress them.

[1714978839]

1714978839

[1714978839]

1714978839

[QuintLeo]

The R9 290(X) were build to run at 95c. The fans wont sweat until you reach the 95c if you didn't set the max temp lower.

I'm running my 290x with an AIO. Stays at 61c full load.

Wonder what you are planning to mine with the 290.

 R9 290 thermal limits at 94, on the reference-board design.
 NOT a good idea to let them get that hot though, they're a lot happier for the long term if you can keep them down to around 80.

 290 is still a good ETH mining card, though I'm not sure that's the BEST option.

[Za1n]

3 years?

As has been said many times already, a properly cared for GPU can have a indefinite life-span with usually the part becoming obsolete before failing. This can easily mean 5+ years of operation.

Most GPUs that have some defect from manufacturing or installation damage will usually fail within the first month or so of continuous operation, which since is within the warranty period usually pose no problem other than the inconvenience of doing the RMA.

Other factors that can lead to failure have been mentioned already in the thread, but normally boils down to the care one takes to maximizing their longevity. Electrical problems (surges, brownouts, static discharge, loose connections, bad or unstable PSU) are probably near the top of the lists, along with excessive heat (higher than 80° C operating temperature), dust, dirt, humidity, and of course physical damage.

By being proactive and trying to minimize any of the above factors will lead to a long lifespan. I have many mining GPUs that have 4+ years on them and are still hashing away, and I have several non-mining GPUs that are 6+ years old that still work but are otherwise obsolete.

I think the majority of my GPUs will be replaced not because of failure, but because newer models are always becoming more efficient and a point in time arrives where the mining efficiency is no longer worth the operating costs. This is probably somewhere around the 4 to 5 year mark as some of my oldest cards while still viable for algorithms such as Equihash are really starting to fall behind newer cards, especially those such as the Nvidia 10xx series

[papasmurf1337]

Keep your cards FAN not to run more than 70% , and you will have them for like 5 years.
Main factor is to keep your fan because that mostly goes to die, not gpu or their memory.
Also it is much important to clean dust from them like once per month or something that like that.

[Za1n]

Keep your cards FAN not to run more than 70% , and you will have them for like 5 years.
Main factor is to keep your fan because that mostly goes to die, not gpu or their memory.
Also it is much important to clean dust from them like once per month or something that like that.

While this is true and a good idea, myself I do not consider a fan failure an actual failure of the card unless it got to the point you didn't catch it and it overheated causing damage.

I have had fans fail, especially those old 3 fan 7970's, and you can buy replacements for under $10 and easily fix them yourself, prolonging the lifespan of your graphics card. If it is under warranty you can send them in too, but $10 is usually less than the RMA shipping and it is simple enough to do it yourself.

[AndyM456]

I have a pair of GTX650 bought in 2012 still running BOINC at 100% 24/7. Air cooled, fans on Auto, no OC with an open case. Had various other GTX running for 5 years plus no problems all on Boinc or Folding@home.

Only added Team Red to the farm since starting Alt Coin mining in last year so not sure on those, but Team Green can certainly do 5 years no problems at all.

[civilufo]

3 years?

As has been said many times already, a properly cared for GPU can have a indefinite life-span with usually the part becoming obsolete before failing. This can easily mean 5+ years of operation.

Most GPUs that have some defect from manufacturing or installation damage will usually fail within the first month or so of continuous operation, which since is within the warranty period usually pose no problem other than the inconvenience of doing the RMA.

Other factors that can lead to failure have been mentioned already in the thread, but normally boils down to the care one takes to maximizing their longevity. Electrical problems (surges, brownouts, static discharge, loose connections, bad or unstable PSU) are probably near the top of the lists, along with excessive heat (higher than 80° C operating temperature), dust, dirt, humidity, and of course physical damage.

By being proactive and trying to minimize any of the above factors will lead to a long lifespan. I have many mining GPUs that have 4+ years on them and are still hashing away, and I have several non-mining GPUs that are 6+ years old that still work but are otherwise obsolete.

I think the majority of my GPUs will be replaced not because of failure, but because newer models are always becoming more efficient and a point in time arrives where the mining efficiency is no longer worth the operating costs. This is probably somewhere around the 4 to 5 year mark as some of my oldest cards while still viable for algorithms such as Equihash are really starting to fall behind newer cards, especially those such as the Nvidia 10xx series

Haha, I agreed... I don't think I will still using my gpu for mining after 5 years, it is old and bad hashrate at that time. But of course don't die in a few months...

[winniethepool]

keep them cool (low 70s or under i keep mine in the mid 60s)  undervolt them. make sure they are on a GOOD psu (for the better voltage regulation). dont run fans over 80-85% it doesnt cool any better and will just kill the bearings faster, especially cheap sleeve bearings. but buy cards with ball bearing fans as they just last and last.

no reason why you can get years out of them. my HD6870 and two HD5830s went a few years 24/7 during btc/ltc mining days. they still work.

my original mining card, a HD4850 single slot blower fan, the fan did die on it after about a year at 90+% fan speed. stuck an after market cooler on it and put it in a gaming rig i gave to a friend. worked for a couple more years gaming till he built a newer game machine.

oldest currently mining card is a sapphire 390 nitro, two years old i think. still running fine.

back in my folding@home days I had two GTX8800 gpus go for a couple years 24/7. they are still working last time i fired them up about 6 months ago.

probably the best investment you can make for gpu longevity (aside from some good fans blowing at the rig) is using a high quality psu. it will provide your cards with clean power. junk psus put trash on the lines that will slowly degrade the gpus, as well as the rest of the system. cant hurt use a good surge suppressor or do as i have done, install a whole house surge suppressor.

Thank you very much for sharing your experience with us. Amazing tips.

[Nasdam]

The R9 290(X) were build to run at 95c. The fans wont sweat until you reach the 95c if you didn't set the max temp lower.

I'm running my 290x with an AIO. Stays at 61c full load.

Wonder what you are planning to mine with the 290.

 R9 290 thermal limits at 94, on the reference-board design.
 NOT a good idea to let them get that hot though, they're a lot happier for the long term if you can keep them down to around 80.

 290 is still a good ETH mining card, though I'm not sure that's the BEST option.

I have some R9 290. But I only run them at 900mV or lower. The temperature is below 70 degree C. They are long lasting.

[QuintLeo]

3 years?

As has been said many times already, a properly cared for GPU can have a indefinite life-span with usually the part becoming obsolete before failing. This can easily mean 5+ years of operation.

Most GPUs that have some defect from manufacturing or installation damage will usually fail within the first month or so of continuous operation, which since is within the warranty period usually pose no problem other than the inconvenience of doing the RMA.

Other factors that can lead to failure have been mentioned already in the thread, but normally boils down to the care one takes to maximizing their longevity. Electrical problems (surges, brownouts, static discharge, loose connections, bad or unstable PSU) are probably near the top of the lists, along with excessive heat (higher than 80° C operating temperature), dust, dirt, humidity, and of course physical damage.

 Long term, FAN DEATH is the leading cause of card failure for cards that make it out of infant mortality.

 I've got several HD 7750 and a 7870 that date back to LITECOIN GPU mining days - they've been in pretty close to continuous usage that whole time (4 years now almost), doing mining or Distributed.Net work (which offers a VERY similar load to most cryptocoin mining algos as it IS crypto work as well).
 Had to replace the fans on all the HIS 7750 cards (junk sleeve bearing fans) in less than a year, the Sapphire HD7750 JUST had it's (ball bearing) fan die last month, and the HIS blower-type card is still going strong on the original fan (ball bearing LOW loaded big-bloody-blower, it's the IceQ model).

Everything I had before that point was either ancient pre-Stream pre-CUDA, or AMD APUs.

[papasmurf1337]

I have my R9 380 models for 2 years mining ...
but yeah it is big revolution from 380 to now models...

[Bare]

94 is too much it wont last for long set the cooler to max 100% arround 70-75 max 80 would be normal!

And that's the only reason why I would never buy a card from a miner generally because of miners like you, who treat their cards like a piece of crap,
100% fans with way hotter temps than ok for 24/7 workload.

Basically, when I see listed 2nd hand cards I wanna buy that have been used for mining (in description) I assume the worst, shit treatments from guys like you.

Who knows how much did you shorter card's/fan's lifespan with these settings.

I make sure my cards never go above 63-64°C core and VRM always below 80°C(at least for hawaii/fiji chips, older cards like 5970 were much more resilient because of bigger fabrication process -therefore VRM could go as far as 105°C for VRM, ~70°C core)
and my fans are always below 70%

And when I sell I don't even want to list them as cards used for mining because I really do take care of them, but still, it drives my price down no matter what because of guys like you!

Don't take this personally totko, this is not aimed at you but everyone who mines like this.

I randomly took your comment as a raged upon victim.

Keep your cards safe and healthy, take care of them, and they will take care of your earnings for a long time.

Cheers!

[Cryptopriest]

VRAM temperature is also important, and most of the time in my experience, its about 15-20 c higher than core temperature.

[Bare]

VRAM temperature is also important, and most of the time in my experience, its about 15-20 c higher than core temperature.

VRM - Voltage Regulator Module, to put it simply the 1st place where electricity passes before it reaches gpu core and other modules on the card, obviously it takes more heat than core since it's job is to give as much power as gpu core asks, as you said 10-20°C warmer than core.

VRAM is video ram, usually gddr5(x) or new the hbm(2) on fiji/vega gpus, they don't generate that much heat to be of a concern

[Facinat]

VRAM temperature is also important, and most of the time in my experience, its about 15-20 c higher than core temperature.

If the VRAM temperature is 90-100 degree, it is still within the operating range.

[QuintLeo]

I have a pair of GTX650 bought in 2012 still running BOINC at 100% 24/7. Air cooled, fans on Auto, no OC with an open case. Had various other GTX running for 5 years plus no problems all on Boinc or Folding@home.

Only added Team Red to the farm since starting Alt Coin mining in last year so not sure on those, but Team Green can certainly do 5 years no problems at all.

 Those date back to the days when cards had fan curves from the manufacturer designed to keep the cards COOL, instead of the current insanity of putting "quiet" as FAR MORE IMPORTANT than "cool".

 EVGA in particular had major issues with it's early 1070 models because the FAN CURVE WAS SET WAY TOO LOW (despite them pointing at "didn't have thermal pads on the VRM models", the cards last just FINE and stay plenty cool enough if you turn the fans up past the factory-default UNDER 40% AT 80C insanity - the actual final FIX out of EVGA was a BIOS change that turned the fan curve up to more like 50% at 70C).


 Factory default fan curves on ANY recent card are set crazy-low, and ignore the possibility of the card being used in a "compute" capability (the manufacturer wants you to pay the crazy-expensive price for their "professional" or "workstation" cards instead even though they tend to use the SAME GPU and the SAME ram and have an INFERIOR cooling solution so they have to clock them a TON lower to keep cool).

[Ludwish]

I have a pair of GTX650 bought in 2012 still running BOINC at 100% 24/7. Air cooled, fans on Auto, no OC with an open case. Had various other GTX running for 5 years plus no problems all on Boinc or Folding@home.

Only added Team Red to the farm since starting Alt Coin mining in last year so not sure on those, but Team Green can certainly do 5 years no problems at all.

 Those date back to the days when cards had fan curves from the manufacturer designed to keep the cards COOL, instead of the current insanity of putting "quiet" as FAR MORE IMPORTANT than "cool".

 EVGA in particular had major issues with it's early 1070 models because the FAN CURVE WAS SET WAY TOO LOW (despite them pointing at "didn't have thermal pads on the VRM models", the cards last just FINE and stay plenty cool enough if you turn the fans up past the factory-default UNDER 40% AT 80C insanity - the actual final FIX out of EVGA was a BIOS change that turned the fan curve up to more like 50% at 70C).


 Factory default fan curves on ANY recent card are set crazy-low, and ignore the possibility of the card being used in a "compute" capability (the manufacturer wants you to pay the crazy-expensive price for their "professional" or "workstation" cards instead even though they tend to use the SAME GPU and the SAME ram and have an INFERIOR cooling solution so they have to clock them a TON lower to keep cool).

I always use MSI afterburner to set the fan curve. And I only use 70% of the rated maximum power.

[Kumic]

Personally, I'm nervous when I see temperatures above 60°C. I like to push my cards in terms of overclocking, but I don't like to have big temperatures on them. I don't have so many cooling problems because I don't have a farm, so my fans are above 70% maybe just in the summertime.
Under 65°C is ok, between 65°C - 75°C tolerable, but above 75°C try to do something.

[elissa]

I have 18 sapphire r9 390 nitro working for a year. I had problems just with one card but during July when the temperature outside was 40 degrees Celsius and the card had around 80 degrees. Others are fine. Now the temperature is low and they are just fine with 80% fan.

[Nasdam]

I have 18 sapphire r9 390 nitro working for a year. I had problems just with one card but during July when the temperature outside was 40 degrees Celsius and the card had around 80 degrees. Others are fine. Now the temperature is low and they are just fine with 80% fan.

The R9 390 is very good if you do not stress them too much.