Above are the basic card profiles to give you an idea of what to expect speed wise, along with the heat and power consumption while performing these tasks.
Notice in the test below that this is total system power consumption. We tried our best to push the load as much as possible on the video card alone and nothing else. We ran FurMark at the highest settings our system would allow. We feel FurMark gives the most reliable readings on GPU load and on how much wattage the card puts out without stressing other components in the system too much. (Total consumption measured at outlet)
At stock speeds, the load power consumption was an improvement over the older GTX 480. It is good to see that, so far, the card is performing as it was envisioned to do. The load has been brought down to a respectable range, which is now only around 30-40 watts above the GTX 460 and HD 5870.
Now with the voltage maxed (1.21v), the ASUS GTX 580 is still hanging in there. At 457w, it is still a little higher than we would like to see. However, this is a great improvement over the older GTX 480. With this progress, we really cannot complain too much as clearly NVIDIA is really trying to improve on this as you can see.
We chose to sort the decibel test by fan levels at 50%.
Why? A lot of fan profiles are set to different percentages at stock speeds. This will give you a false breakdown for the decibel levels, as some fans may be running at 25% and some at 40%. That is why we chose 50% so that every card is on a level playing field, yielding the truest results.
The ASUS GTX 580 came in right in the middle of the pack. The idle temperatures were higher than most, but unless you have a decibel meter the difference would be negligible. The GTX 580 only let u set a max speed of 80% however, and this was a big reason in helping the GTX 580 stay at a lower decibel rating in our max testing.
To see how this stacks up with everyday noises, please look below at the chart.