Depending on a lot of factors - brand of card (and therefore quality of components), luck, environmental setting the cards are in, etc - running gpu fans at 100% all the time will wear the fans out much more quickly. I believe they are spec'd to ideally run for the life of the rest of the card's components at about 60% all the time, and going above that is supposed to be just for extreme situations in gaming - which is what these cards were actually designed for.

I have been mining for over a year in Australia which is a very tough place for GPUs in the summer without aircon and I started out with used AMD cards mining ETH (in the good old days of mining 5 ETH every three days at 250mh/s). I have had to run the fans at 90%-100% sometimes with the cards at 80 degrees plus and in the last three months I have had many fan failures. I have some Sapphire Tri-X R9 290's and I have replaced the whole triple fan array on two cards, and one individual fan on another card. R9 290s are power hungry monsters and always ran really hot compared to other cards, and all cards use more power when they are hotter - so it's a "balancing act". 3 of the 1st 290s I ever bought were reference design blower style coolers and were SO loud in summer I had to watercool them (at ridiculous expense) or my lady would have either left or made me stop mining, because my rigs are in a room next our bedroom. One card I once owned i nicknamed "the lawnmower" because it was so f'ing loud. I sold it to a friend who didn't care. Watercooled cards are awesome - silent and always running at 50-60 degrees even on 40 degree days, but it certainly isn't economical in terms of ROI. I just did it because I love mining as a hobby and love modding stuff to make it work better. I still run those three watercooled cards and they have always been overclocked to within an inch of their lives and they keep on going at low temps and no noise.

I would encourage anyone to try to lower the temperature of the room if possible, spread the cards out as much as possible so the fans can "breathe" and use external fans (like even large domestic cooling fans) to blow the hot air away from around the rigs. Hopefully if any or all of these measures drop the temperature you should definitely lower the GPU fan speeds whenever possible.

In recent months I have been slowly migrating to Nvidia cards. I used to run 2 x 6 card AMD rigs. Now I have 1 x 5 card Nvidia rig (with 2 more 1070's on back-order)  and 1 x 4 card AMD rig (I sold the most "respectable" 2 cards on that rig recently - a Sapphire 390 nitro and an MSI 290x). I have been wondering what sort of temperatures are safe cruising for Nvidia cards. Older generation R9 AMD cards were always expected to run hot and designed that way because they were big power consumers, but Pascal cards that all run at 150W or less shouldn't be too hot that often right?

What kind of temperatures are you guys comfortable with on 1060/70/80s?

I have two rigs go almost 3 weeks 24/7. Only one has done a soft reset in which the miner reset and started back up.

The second rig hasn't even had a soft reset in the three weeks.

For nanopool:

append / followed by your email address to the end of the worker; for example if my worker is nvOC and my email is bob@aol.com, I would use: