Originally Posted by lynxxyarly
There is no 'scam', it's people having a knee-jerk reaction to something that was made VERY clear in the readme. The dev has a fee that he takes out for development of the miner (5 shares for every 500 you mine). People who don't RTFM are at fault here. If you're on a 600 series card, use the 0.3c miner. The arcminer is better for 700 series cards.
My guess is since you didn't have a full cover waterblock, the VRMs just didn't get any cooling. Need to actively cool those VRMs, guys!
Originally Posted by Aparition
I did some testing on my 780 for heat management since I am interested in mining.
The Core is fine, I tried with a Kuhler 920 block mod, but the real issue are the VRM's.
The ACX cooler is adequate because it blows directly on the PCB plate above the VRM's. With just a block (any) on the core the VRM's will overheat without active cooling. Just relying on the PCB heat plate and case airflow isn't enough.
The PCB heat plate started to show signs of intense heat accumulation, the metal started to discolor, looks wet. Very VERY hot to the finger!
So for the 780 (dunno about the titan or ti) you'll want to add some active cooling to the VRM's.
I tested just with benchmarks and games. The Miner should be a bit harder on the VRM's.
Originally Posted by yanks8981
Are you saying the stock cooler is not adequate, or that the Kuhler 920 is not adequate?
Originally Posted by theilya
Interesting. There is no way to monitor VRM temps on 780 but I suppose warranty covers that?
How would I monitor the VRM temps ?
Originally Posted by feartheturtle
I have a EVGA GTX780 Classified on air and my GPU-Z shows VRM temperature of 61C (64C max). Is this a safe 24/7 mining temperature if the temperature readings are to be accurate? The core temps are around 67C-69C.
I wanna clear a little misconception that people may have or not understand about the ACX vs reference designs.
EVGA's ACX coolers have a better and bigger design for cooling the GPU (compared to the reference), as the heatsink size is bigger, and more heat can be dissipated from the sides (inside the case). However, the ACX cooler has NO proper VRM cooling. Just by attaching a metal plate above the VRM's and memory chips, wont necessary cool it down. Now, however, the reference baseplate does have a dedicated heatsink welded to it and which does act to cool the VRM's. And ACX heatsinks send hot air to the baseplate, when the GPU is underload, and that's never a good thing, to cool something down with hot air, that's already hot.
My case was actually different, I was mining scrypt (heavier than PTS), overclocked to the max, 150% TDP on EVGA precisions graph, and runnig Skyn3ts unlocked bios, meaning, 1.212V at load.
After I got my refund from amazon, I then bought three 780's, all from different manufacturers, so see what's up with this VRM heating thing. One was a reference EVGA, the other was Galaxy's HOF series, and last was a lightning. This time I ran the stock EVGA bios with a full-cover XSPC waterblock and found that the VRM was still getting hot (not to the extent of my ti, though, particular of not overvolting it). The HOF and lightning were running under universal GPU blocks and only the lightning's VRM's are actually touchable (on load).
It makes sense as it has twice the no. of phases, which splits the load between all of them. The lightning's VRM's were running cooler on a universal GPU block than EVGA's reference card, with a full cover block. So you can figure it out yourselves.
People use HWinfo to monitor VRM temps, but I simply touch the back of the PCB as that gives me a good idea.
The classified shouldn't be a problem with those VRM temps on air, it's normal.