Then if you don't care about noise, set the fan at 100% and overvolt/overclock until 80-85°C again. Up the voltage in 0.05v steps and see how much it overclocks, there will be a point where the required voltage far exceeds the frequency increase, and there you know you're well past your card's sweet spot.
The card won't throttle usually until 95°C (more or less this is the point, depends on the particular card) but then you can be certain such temperatures aren't good at all for the rest of the components on the card, on the long term. Temperature and voltage kill cards (or any chip), but especially temperature, and even more so at >80°C AND when you're already overvolting. A cold silicon chip is a happy silicon chip, that's why better cooling lets you overvolt more and with much less of a posibility of damaging the GPU/CPU/whatever. When I talk damage, it could be either a card that won't work anymore, or in your case, not being able to do 900MHz on 1.18v, requiring more... but then that's out of reach of your cooling solution. These two are the most usual failure modes once you start treading these waters.
You've been warned, if you want to keep overvolting to overclock more and you're already running at high temperatures, whatever might (or not) happen to the card in the near or long term is up to you. I wouldn't go past where you already are because it certainly isn't worth it for one or two FPS more (not to mention you are near or past your card's sweet voltage/frequency/temperature spot), but then it isn't my card :biggrin:
I mean, your card's GPU is built on a 40nm process, it can by definition take more voltage than the past 4 years' cards built at 28nm, but you can't overvolt with peace of mind because you're already at 80°C at 80% fan speed. See what I mean? If you were running much cooler 1.3v wouldn't be a problem for your GPU. Yet in the situation you are, it is.