I’m mining Primecoins, and Litecoins on the side (GPU). I’m currently have an i7 4770k and an GTX 780. Of course, I’m using my CPU for XPM and GPU for LTC. I’m sort of lowering my overclock since I don’t really like mining on my CPU with an average temp of 90C, when my gpu is running idle. I started to reach 90C-100C when I was running cudaminer as well. This is probably why I found black grame on the back of my socket when I was redoing the cooling paste.
So, it lowered my cpu clock to 4.7Ghz, and halfed the voltage settings from +200mV to +100mV. Now I’ve got around 80C on my CPU and about 70C on the GTX780.
Is this a save temperature to be working on? Note that I don’t leave my computer on for 24 hours a day, but more like 12-15 hours. Graphics card fan speed remains about 1850RPM, while max is 4400RPM. CPU fan sits somewhere between 2250RPM and 2500RPM, max is 3000RPM. All my case-fans will usually go 100%, unless it is really cold outside (remember, it is winter). When it is cold enough, only 1 case-fan remains spinning at 100% (intake-fan directed at the videocard), 1 stays around 65-70% (intake-fan directed at cpu fan), plus 2 smaller intake-fans will only every so often raise from 25% to 100% and stay at 100%, but one more than the other.
Sorry if this is a little bit too much of information, but I’d like to know if there is anything wrong with driving the cpu/gpu to such temperatures and fans to 100% (or not driving them to 100%) for hours and hours.
Also, it was probably better to post this in the Primecoin mining forums, sorry.
80C on overclocked haswell seems still within limits for me, these chips can run very hot. Post your voltages under load, however.
1.286V as read by HWMonitor and 1.280V-1.288V as read by CPU-Z.
Just to put my 2 cents in,
I only recently moved to a 2011-socket board in my case, but have always been running a q6600 at 3.86 ghz on a p5e3.
I’m currently overclocking my 3820 to 4.5 sometimes 4.7 ghz (in this case by upping the fsb to 125+ and lowering the multiplier) but have always preferred to key in constant voltages.
I’ve found that many times i can run my cpu on a lower vcore then the offset-based voltages would automatically assign to it.
In my specific case to run 4.7ghz table i have to disable all additional core-shutdown and turbo features but then it’s stable at 1.375V instead of 1.45 to have it all enabled. In return i have quite a bit lower core temperature. Also my switch to more FSB less multiplier seems to generate less heat in my case.
Funny things to play with!
I run WC too but the temperatures in comparison (average core temperature) 1.45V when mining XPM on 8 threads; about 85C on the core.
1.375V: 78C on the core.
 didnt make this voltage based comparison on the higher multiplier lower fsb, but i remember using 1.375v without worry and having core temperatures rougly 80-82 degrees avg. [/edit]
note that my chip’s stock vcore is 1.35v by the way, yours is probably 1.25v