How long has your GPU lasted mining 24/7?

[phorensic]

Two 5830's running 100% @ 75-80C constantly since mid 2011.  A fan just burned out on one of them last week.

7970 running since it was released and no problems yet, but I consider it still pretty new.

Just dust 'em and keep them under 85C and they seem to run forever.  I have 17 years of computer experience and these AMD cards seems to be REALLY well built.  I've gone through hdd's, ssd's, and blown up PSU's while these GPU's just keep churning.

[1713553289]

1713553289

[1713553289]

1713553289

[Littleshop]

About 15 cards.  I have bought and sold some but half of my cards started having fan issues after about two years.  I have 'refurbished' them by replacing fans, heat paste and heat pads and they work great.  I have never had a card totally die never to be repaired but I also do not overvolt my cards.

[Damnsammit]

I've been actively trying to burnout my 5850... lol it's old and my wife says I can only have 2 GPUs so this is how I must get a new 7950

It's been running 24/7 at 90-95 degrees for a little over a month.

That damn thing gets SICK and DEAD, but a restart revives it. 

It's never gonna die.

[freshzive]

Back in the day (last year) I ran 6 X 5970 and 3x 5830. Ran them for 13 months in the 80c range running 24/7.

All cards were purchased used on Ebay.

5830s never had a single issue and sold them for more than I bought them for. Had two fans die on 5970s, both were easily replaced. One 5970 died after about 11 months of mining, got the impression  that it was pretty heavily used prior to purchase though.

[lunchb0x]

I have a 5770 that has been running for 2 years and 4 months now.

It still runs fine, except that whenever the miner application isn't running, the fan starts clicking for some reason.

[squall1066]

I have seen MANY card come and go, Though I do have 2 5870's xfx reference cards that have been going for just over a year now, They have only been on 300mem 800core on a 1v (think its 1.08 really) with 75% fan, So nice and easy on these cards and it has paid off, I did blow a 5970 in a month with 200/875 overvolted 100% fan Do'h

[superfastkyle]

How long your video cards for the most part, relies almost entirely on the power being provided to it. You will have the best success with clean power from the outlet, from the power supply, from the motherboard and finally from the voltage regulators on the card itself. Many people think its the temperature itself but high temps just put your voltage regulators closer to their highest tolerances, which then results in failing components

[Trillium]

How long your video cards for the most part, relies almost entirely on the power being provided to it. You will have the best success with clean power from the outlet, from the power supply, from the motherboard and finally from the voltage regulators on the card itself. Many people think its the temperature itself but high temps just put your voltage regulators closer to their highest tolerances, which then results in failing components

Unfortunantly if you don't have a (high quality, line active) UPS, ensuring good power from the outlet is difficult or just impossible. I am very skeptical of cheap (<$20) surge suppression devices that a lot of people will think are actively doing something, when in fact they are just a $0.50 MOV or two and some wires on the inside.

The AC side of a quality switch mode PSU is usually pretty good at sorting some kinds of problems with the AC line. Many of the expensive ones can these days can even shut themselves down if they detect strange transient voltages, to protect themselves. There should be some good energy storage on both sides of the high frequency transformer, that helps give good smooth DC power on the output.

For example my Corsair HX1000 has even survived ~1 second complete power outages where most of my other computers and appliances have shut down, no doubt thanks to the huge electrolytic caps that must be inside it, and with not even 1% change on the DC output mind you!

[worldinacoin]

My advice never use new GPUs for mining, the ROI hard to sustain, get the second hands ones cheap.   Also mining does seems to affect my graphic cards that I used for playing games esp running 24/7

[Unacceptable]

Since June 2011: 1-XFX6950,2-XFX-6970's.OC'ed a little to get 400mhs from the 6970's & 365mhs from the 6950    Going on ebay in a week or 2  

All have double lifetime warranty 

Just shut em down last week due to electric costs & my 30GH from BFL arrived,along with 1.2GH in BE's  

[superfastkyle]

Your definitely right about having an active ups and I do plan on buying one when I get more asics, more than just the few block erupters I have now. It just never was reasonable to do when running thousand of watts of gpu's

How long your video cards for the most part, relies almost entirely on the power being provided to it. You will have the best success with clean power from the outlet, from the power supply, from the motherboard and finally from the voltage regulators on the card itself. Many people think its the temperature itself but high temps just put your voltage regulators closer to their highest tolerances, which then results in failing components

Unfortunantly if you don't have a (high quality, line active) UPS, ensuring good power from the outlet is difficult or just impossible. I am very skeptical of cheap (<$20) surge suppression devices that a lot of people will think are actively doing something, when in fact they are just a $0.50 MOV or two and some wires on the inside.

The AC side of a quality switch mode PSU is usually pretty good at sorting some kinds of problems with the AC line. Many of the expensive ones can these days can even shut themselves down if they detect strange transient voltages, to protect themselves. There should be some good energy storage on both sides of the high frequency transformer, that helps give good smooth DC power on the output.

For example my Corsair HX1000 has even survived ~1 second complete power outages where most of my other computers and appliances have shut down, no doubt thanks to the huge electrolytic caps that must be inside it, and with not even 1% change on the DC output mind you!

[spartan82]

4 x 7970 (running between 65-75c)
2 x 6990 (reaches temps from 65 up to 80)
1 x 6970 (65c)
3 x 6950 (65-75c)

24/7 mining (down only for maintenance or shifting to another area) approximately 6 months now.

[HellDiverUK]

Only card I've had die was an old 5870 - got it used, mined the hell out of it for 6 months until the screen started artifacting.  Sold it on eBay as spares/repairs for the same price I paid for it.

The 6970's fan is a bit wobbly after a year, and so far so good on the 2x 7950, 7850 and 7770. All are 6 months+.

[gdpone]

I've been actively trying to burnout my 5850... lol it's old and my wife says I can only have 2 GPUs so this is how I must get a new 7950

It's been running 24/7 at 90-95 degrees for a little over a month.

That damn thing gets SICK and DEAD, but a restart revives it. 

It's never gonna die.

wow dude really, your wife determines how many GPUs you can have LOL.  I might accept it as a cost factor, but its doesn't sound like the case.  Im sorry man, but thanks for being an example of how I will never let a woman control me.

[Damnsammit]

I've been actively trying to burnout my 5850... lol it's old and my wife says I can only have 2 GPUs so this is how I must get a new 7950

It's been running 24/7 at 90-95 degrees for a little over a month.

That damn thing gets SICK and DEAD, but a restart revives it. 

It's never gonna die.

wow dude really, your wife determines how many GPUs you can have LOL.  I might accept it as a cost factor, but its doesn't sound like the case.  Im sorry man, but thanks for being an example of how I will never let a woman control me.

Well she doesn't really determine how many GPUs I can have... my mother board does.  And she is all practical and shit and says "wait until that one burns out until you get a new one" so meh.

Eventually I will get a new mobo that has more PCI-E slots, but then I have to get a new processor, ram, etc.

As for everyone who is pissing and moaning about GPUs will never have an ROI, you are wrong.  I spent $80 on my 5850 and it has paid itself off already. 

Word to the wise:  If you are mining LTC or BTC, then you are doing it wrong!

[tonto]

My pair of 7970s I got day one they were released, and they have been mining non-stop to date, around 85C to 89C

[Trillium]

As for everyone who is pissing and moaning about GPUs will never have an ROI, you are wrong.  I spent $80 on my 5850 and it has paid itself off already. 

Word to the wise:  If you are mining LTC or BTC, then you are doing it wrong!

Yeah I got a 5750 (which is not a really impressive card in any case... but still some extra K or MH/s) for $40 on ebay including delivery and it's paid for itself at least 2 or 3 times over the last few months. The thing that gets me about this card is how incredibly stable it is compared to say dual GPU cards (like my 5970) or 6870. Also I have blocked the fan up a few times and it's gotten to 100+ degree temps without crashing, happily mining away for hours in that heat. Never seen that on a AMD card before.

[GSnak]

Just keep the dust out of the heat sinks and they should last a long time. Mine start to crash after the filters clog up but cleaning it out restores them to new. My Sapphires undervolted/overclocked run ~62, my Mysts run much hotter ~76 in a closed case.

[TechCF]

7970 Asus DirectCU II lasted 14 months (mining, and some gaming time). One fan was oiled because of noise after 10 months. After 14 months the machine would BSOD every time I logged into windows (with Aero) or launced 3D apps. Was running 1100 Mhz @ 80'C. Awaiting a replacement.

[rograz]

My initial 3x5850s have been running since may 2011 and still going strong.