Quote from: mother_of_another on June 06, 2011, 12:59:19 AM

Quote from: Veldy on June 05, 2011, 03:32:51 PM

Quote from: Kehaar on June 05, 2011, 01:07:08 PM

Quote from: Allenssmart on June 05, 2011, 01:00:58 PM

I've been mining for 13 hours straight with 100% usage on both gpus and temps at 65C and 67C. Is it safe to run this 24/7?

That temperature is fine for chips. Worry when it gets to 80C+ (even then it'll probably be fine). Some cards idle at like 60+

Hardware will stop itself from overheating - either by drawing less power and slowing down / increasing fan speed - or turning off the computer alltogether.

Fatigue happens if the the hardware expands and contracts a lot (increase/decrease in temperature).

I imagine going between 40 - 70C wouldn't be much of a problem tbh, though I'm not an expert.

Temperature cycling is NOT fine for anything. It causes "fatigue" wear. Any 3rd year mechanical, civil or aerospace engineer or materials science major will know this without thinking 1 second about it. Since there are MANY MANY parts in a board and only one needs to fail to kill it, fatigue is a nightmare for them. Mining is not the norm for a gaming card ... gaming is. Gaming runs at whatever is required while you are playing and then goes back to typical idle when finished. This is cycling over long periods of time and much few times during the life of the card. With mining, it is happening as often as once per minute in this pool.

So, sorry, but you are absolutely incorrect about the damage temperature cycling can do, even if below 70C the entire time.

Which is why I posted up this little python PID controller for temperature control. Put in the set point and good tuning parameters and it should be able to keep temp. pretty constant even with work 'turbulence' ... needs some patching work is all (minimal py skills me)

https://forum.bitcoin.org/index.php?topic=10062.20

Nice too perhaps. Still, if you aren't driving the card through usage, no controlling of fan or other cooling will keep it hot.

FWIW, I ran a test today and did not see one connection issue and every miner idle issue was near a long poll push and further, it idle was short enough that the GPU usage % was a short drop and then back up to full again and the temperature barely fluctuated at all. So, I don't think this pool is a risk to hardware. Unfortunately, I was getting 3+% stale shares and that is NOT good.

Soon enough there will be two, then three servers and perhaps this issue will hopefully be resolved as well. I use the same mining parameters with every pool [other than login of course].