How dangerous is mining for your GPU?

[jmfg187]

   I was looking at a thread in the marketplace where the poster wants to buy cards that are not working or have problems etc.

Code:

https://bitcointalk.org/index.php?topic=101624.220

 It has page after page of people with fried cards, this makes me wonder if mining does some damage to them that gaming does not? If anyone has an idea about this please discuss because I am wanting to build a rig for LTC or NVC but if I am gonna be frying cards all the time it may be more expensive than I thought. Thanks in advance!
James

[1715465208]

1715465208

[1715465208]

1715465208

[1715465208]

1715465208

[bowen151]

Running anything that has moving parts will fail eventually due to obvious reasons.

Gamers = a few hours intensive gpu usage
Miners  = 24/7 intensive GPU usage

Mining usually will take up 99% of your GPU's processing power and this generates heat as a result of drawing more power from the PSU, thats why you need to run your fans at around 60%+

some people have different theorys and prefered settings regarding fans but if you run a fan 24/7 on even a low speed then what will happen? It will need maintenance as it begins to wear down over constant use.

Fried cards = overclocked/overvolted and mined to death, GPU's running too hot for too long, you name it, its just like any other computer component, it can and will fail.

[jmfg187]

I do understand what your saying but I can mine at an 18 intensity on my 6970s and they don't break 80C, I suppose if a fan suddenly died you would be totally screwed but most of the cards I have had started clicking or something to let you know the fan was going bad. I am wondering if a lot of people don't care and just run their cards at 100C all the time or something.

[l008com]

I would think that with adequate cooling, your GPU should be fine. Just replace a fan if it dies, don't ignore it.
I would imagine that in the Mac world, it can be a bit more complicated, since Apple sometimes lets things run a little hot, so they can run a little quieter. SMCFanControl is your friend, in this case.

[Malawi]

I do understand what your saying but I can mine at an 18 intensity on my 6970s and they don't break 80C, I suppose if a fan suddenly died you would be totally screwed but most of the cards I have had started clicking or something to let you know the fan was going bad. I am wondering if a lot of people don't care and just run their cards at 100C all the time or something.

With time a small percentage of cards are bound to fail. With many cards spread among many miners some are bound to fail.
If a card is overclocked to give max output, the chance that it fails gets bigger.

Also there is the cooling thing, if a fan is failing, the card should stop by itself due to high temp, but then it fires up again as soon as it's cold enough. This is a sure way to destroy any electronic equipment, as they don't take well to much changes in temperature.

A mining card that is run at a pretty stable and not too high temp should live the longest. (Even longer than a gaming card that sees a bit of use.)

[MrBudgens]

Keep your GPU fans clean and you shouldn't have a problem.  My GPU fan got clogged with dust to the point that it stopped turning and the PC started to fail all the time.  The case fans were louder so I never noticed the change in sound.  The lesson is to check actively that you don't have dust buildup.

[Nitmeare]

  I was looking at a thread in the marketplace where the poster wants to buy cards that are not working or have problems etc.

Code:

https://bitcointalk.org/index.php?topic=101624.220

 It has page after page of people with fried cards, this makes me wonder if mining does some damage to them that gaming does not? If anyone has an idea about this please discuss because I am wanting to build a rig for LTC or NVC but if I am gonna be frying cards all the time it may be more expensive than I thought. Thanks in advance!
James

If you use your GPUs processing Unit at 100%, it will become very hot and may fail. Just use only 50-70% of your GPUs capacity for mining, this way it will not consume that much energy and will not become that hot so it may have a longer lifetime. Most miners can be configured to use GPU with less agression to prevent overheating/hardware damage.

[C_J]

Short answer... As dangerous as your ignorance is.

You can theoretically mine at 10% of your GPU's capability for decades without an issue.

Though then again, I can't even get a single client to function beyond running imaginary blocks so why listen to me?

[erk]

Some cards get really hot others don't, I have a HD5770 from Sapphire that wants to run at 85C. Where an nVidia gts450 is happy at 65C. Both have MSI afterburner auto controlling the fans.

[Rikai]

The damage is usually from insufficient cooling, as is the case with most computer issues. I have four large fans (two mounted directly over my video cards) to help facilitate cooling. On top of this, if I am mining for an extended period of time, I run a temperature monitoring program on my computer, and open the case and place a box fan against it. Not the best idea, but I have had no failure as of yet

[achilles78]

Some cards get really hot others don't, I have a HD5770 from Sapphire that wants to run at 85C. Where an nVidia gts450 is happy at 65C. Both have MSI afterburner auto controlling the fans.

my   Sapphire HD5770's temperature reaches 85C

[SimonQ]

As others have said, the main damage comes from the moving parts (fans) breaking down with constant use, but also high heat can cause premature failure of integrated circuits and connections.  In the end, heat is the enemy.

Two things you can do to prolong the life of your cards:

• Don't overclock!
• Don't even run them at 100%.  Run at 80% or 60% or whatever.  This allows the fans to run slower too.
• Add additional external cooling: fans, air conditioner, etc.

Happy mining!

[blackbox]

  I was looking at a thread in the marketplace where the poster wants to buy cards that are not working or have problems etc.

Code:

https://bitcointalk.org/index.php?topic=101624.220

 It has page after page of people with fried cards, this makes me wonder if mining does some damage to them that gaming does not? If anyone has an idea about this please discuss because I am wanting to build a rig for LTC or NVC but if I am gonna be frying cards all the time it may be more expensive than I thought. Thanks in advance!
James

It shouldn't damage them.

Even if it did, If you think bitcoin will take off, the profit you eventually make will outweigh the price of a new card.

I wouldn't worry about it

[pipensa]

You can come to russia, this is very cold, your computer will feel nice there!

[Wesly]

I have 2 AMD 6870 cards OC to 980Mhz that had been running 24/7 for a whole year without any issue.

[blackbox]

I have 2 AMD 6870 cards OC to 980Mhz that had been running 24/7 for a whole year without any issue.

but do you play games with them and are they still ok, no artifacts?

[spurs32]

 

[Qdentica]

Use proper cooling!!! The better the cooling, the longer your Graphics Card will last.

[ft73]

Tried with my Nvidia Quadro, temps are fine, but mhash are low

[TeqnoHaxor]

One of my customers cooked their graphics card recently mining for Bitcoins. Keeping your chassis free of dust and debris is a big help, also make sure you have sufficient ventilation to remove hot air - consider installing side exhaust fans as many graphics card coolers simply push air outwards towards to the side panel.

Also make sure you have a power supply capable of delivering your graphics cards power requirements - peak power is not acceptable. Use the table on the side of your PSU to check the current rating (amperage) of your +12V, then multiply by the efficiency rating. For example, 25A on a 80PLUS Bronze certified is 25 x 0.8 = 20A sustained.

Expect 24 months life out of modern graphics cards, the RAM will almost certainly fail before the GPU does, giving colorful visual artifacting in games etc.

[kojo23]

Check for dust!
Thats probably the most important part of the game...

[stormtide]

I'm thinking TriForce setup with water cooling 

[vm1990]

everything fails at some point most card are only suppose to have a life span of around 100,000 hours but that said most do last longer people keep saying run your card so its nice and cool but they also forget that all the small circuits and resisters and VRMs are all points of failure after time they will fail (most of the time youve upgraded cards long before this happens)  alot of fails are down to neglect and to much heat, in some way mining with a GPU can extend the life of the card as it dosnt have a cooling down period (this is where things like older GPUs, xbox360s and even PS'3 fail) without the card cooling down the metals such as solders dont expand and contract constantly so they dont crack or break causing blank sreen and 3red lights of death (xbox) or yellow light of death (PS3)

over all there are pros and cons to it but done correctly and well maintained you can get very long life spans out of GPUs

[BitFollower]

Heed all the warnings about dust. It makes a HUGE difference in temperature. My card started having its fans go crazy just from gaming. I ignored it for a while then checked the temperature one day and it was over 90 degrees! Took some compressed air to it, back to normal.

So watch temperature, and clean your card regularly.

[deadweasel]

Heed all the warnings about dust. It makes a HUGE difference in temperature. My card started having its fans go crazy just from gaming. I ignored it for a while then checked the temperature one day and it was over 90 degrees! Took some compressed air to it, back to normal.

So watch temperature, and clean your card regularly.

+1 +1 +1

Dust your PC bits once a month, especially when you are maxxing your GPU out constantly.  It will get caked in dust very quickly. 

I've been dreaming of a pc air filter... Remove the side of your case, tape air filters to it and mount fans to it.  Still a poor idea, so it stays in my head.

[vm1990]

Heed all the warnings about dust. It makes a HUGE difference in temperature. My card started having its fans go crazy just from gaming. I ignored it for a while then checked the temperature one day and it was over 90 degrees! Took some compressed air to it, back to normal.

So watch temperature, and clean your card regularly.

+1 +1 +1

Dust your PC bits once a month, especially when you are maxxing your GPU out constantly.  It will get caked in dust very quickly.  

I've been dreaming of a pc air filter... Remove the side of your case, tape air filters to it and mount fans to it.  Still a poor idea, so it stays in my head.

your not as crazy as you sound, some high end cases do come with dust filters on the front in fans reduces the amount of dust very nicely to

[blackbox]

Well I just had my fan on my GPU go.  Took it out, vaccummed it, but still makes a terrible noise when it spins.  Does not seem user replaceable...

[BitshireHashaway]

I don't see it as dangerous at all as long as you make sure that you have adequate cooling for your video card. That being said anything with moving parts that is constantly running is bound to fail one day.

[greenboy84]

There are commands you can use with cgminer to ensure your temp stays good, I use this:

 --temp-target 82 --auto-fan

[bigkillerstorm]

Yeah watch that dust its evil clean ur gpu whit high pressure air every 3-6 months specialy if u mine 24/7.

[jmfg187]

Well I just had my fan on my GPU go.  Took it out, vaccummed it, but still makes a terrible noise when it spins.  Does not seem user replaceable...

What model is your GPU, I should be able to tell you if you can replace the fan.

[mvrak]

With adequate cooling your GPUs are fine.  Fans wear out.  Running GPUs too hot increases the chance of localized temperature spikes that make kill some part of some transistor.

It only takes one weakest link to bring down the whole card.

Mining is not dangerous, overheating is dangerous.

[JohnnyBrown]

If you choose to OC badly (eg. nonsensical voltages), then you can fry the cores. If you have good temps & OC ranges, then nothing bad may happen (personal experience). If you have a good cooling solution (Like XFX's Double Dissipation or some nice aftermarket ones (eg. GELID)), then you are all set for some conservative or normal OC.

[mountaineer]

 Mining will stress components more so then regular use. As others have said, keeping everything as cool as possible will prevent things from failing.

[yugo23]

I keep all my GPUs below 75 C and they seem fine so far.

[Malawi]

If your fan looks like this, it's time to do a little cleaning.

[Catcher]

Some cards get really hot others don't, I have a HD5770 from Sapphire that wants to run at 85C. Where an nVidia gts450 is happy at 65C. Both have MSI afterburner auto controlling the fans.

My HD5770 from Sapphire runs on temperature 72 °C. No overclock.

[Kruniac]

Burning up a card is a huge concern of mine. I'm really, really curious as to how the new chips being developed will stand up to constant use.

[SimonJones]

Losing a card at this time would be a big loss! Bitcoin mining using them is not so profitable anyway.

[oddwh]

If you think the temperature is too high, just reduce the frequency. Even if your hashrate will not be the best, you will not burn your GPU 
Me, I just make my possible to stay arount 70-75°C, not more. I don't know if it's the best choice but in my opinion, it seems fine.

[mika1977]

I care for my cards to stay with temperatures below 55 degrees Celsius with my AC room temperature is 24 - 25 degrees, and i have adequate cooling in my rig.

[penguinn]

Yeah, this is a lot of good information. Please keep it coming!

[ccc]

Cards will not fry themselves from mining 24/7. Just make sure they are not running too hot. Keep in mind GPU temperatures don't tell the whole story. Other components like the VRMs can get hot, and if you're not using risers, the heat from the cards can affect the motherboard as well. Make sure you have fans to blow all this heat away from the cards, even if you're using risers. Box fans work well.

[No 1]

mining isnt/wasnt inended for  "casual" man. just dump 5k in btc and speculate for life

[Marduk]

I've been running folding@home 24/7 for about 10 years now on my CPUs and GPUs.  Well, not quite 10 years on GPUs. I don't remember when the GPU client first came out.  However, I've run it on at least 4 GPUs 24/7 though. 

I always overclock everything (CPU/GPU/system memory) as high as I can go and still run Prime95 for 24 hours stable, and then I lower the clocks down just a little.  I used to game a lot, so I would test GPU overclocks running 3dMark in a loop for an hour or so, and then just play various games ALL day long and check for artifacting or crashes.

I would usually not run folding@home for a month or 2 in the summer simply because it would make my apartment too hot!  But in all that time, I have never fried a piece of hardware.  I usually upgrade every 2-3 years just so I have some new hardware to tweak.

[DDP]

I'm no expert, but I don't like the idea of getting the clock waaayy too high, even if I can keep the temperature low. I like to look at what others used as OC settings, and pick a setting somewhere over the middle range, but not too high either. There's always someone proud about reaching the limits of the card stability and posting it. I believe this one will have troubles in a few weeks/months.

[r1nftw]

I have 3x7950's - two of them are sapphire toxic, the 3rd one is PowerColor with two 92mm fans.My question is - the middle card(sapphire toxic) is working at the moment at 70C core temp and.....98C on VRM.I know that the VRM's could maintain normal work until 130C.So, will my card be able to cointinue normal work in the future?

[with the lemons!]

If you set fan high enough, so that core is cooled well, it should live as long as GPU that is used for web browsing. I never seen GPU with faulty fan, so I think they are very durable.

[Antropov]

How about mine HD5850 running at 775 core speed, 1108 mem speed @ 60-65 deg. Celsius? Is it going to burn? Shall I lower the settings or is it normal for aTi vgas? BTW this thread should be sticky...

[DDP]

65C is a low temp (especially if you're using air cooling). I run at 80C 24/7.

It will be better to run your card 24/7 at 90C, than working at 80C 12h/day, and 50C the 12 others, all year long. The change in temperature will change the density of components, and that is a bad thing for your hardware. Worst than just keeping it busy at a higher temperature.

130C is a bad idea. I don't like hitting the 90C, but maybe that's just me.

[Chrithu]

Actually mining might have helped me enlengthen the lifetime of my GPU itself. Setting my rig up for mining during times I don't use it for gaming or just use the main card for gaming (when the game has no crossfire support) was the first time that I actually watched the temps of my graphics devices. And I found that (possibly to enlegthen the fan lifespan) the cards ran at temps of 80C+ at stock clocks when gaming. I setup MSI afterburner to control the fans and increased the fanspeeds so now the cards run at around 70C while being overclocked by 120 MHz.

To answer the OPs question, and possibly repeat what was already said: Mining per se doesn't do more damage to the card than gaming. What is worn down faster than usual are the moving parts. The fans. While mining you should trust your ear. If it starts to sound strange it's better to look for a replacement fan and replace it before it has an outage.

If available for your card of choice it's also worth to install a fan with a bigger diameter than the standart ones. The bigger the fan the lower the speed needed to reach the same cooling and thuse the longer it's lifetime. Plus lower frequenzy is better for the loudness.

[madkind]

I think the magic number is 75, I alway try to stay under that.

[atp1916]

Anything 85C and under is fine for the GPU for 24/7 use.  It may get hot in the room, but the cards should be fine.

In the crate rig i got set up for my 3 7970s, they don't get above 75C.