If you run the chill water full bore you get the largest potential difference between the wort temp and the water temp. It would seem like this is the ideal potential, but I my experience it is not. I always found that the fastest chill rate was when the water was coming out hot to warm and not coming out cold. When you run the water full bore, despite the max difference between wort and water, you have to account for the water moving. WIth the water on full bore a fast transit time through the coil limits contact time with the hot copper too. Therefore if the water is coming out cool, then you haven't captured as much heat as possible.
A super slow stream however just wastes time. You certainly absorb the maximal amount of heat, but it happens during the first portion of the coil and the rest of the transit time is a waste of time.
In theory, I am sure there is a really cool formula for figureing it out based on ID, thickness of the copper, water temp, wort temp, etc etc. But its pretty complicated based on the reality that the wort temp is dropping the whole time. Which involves a differential factor into the equation! Yikes!
Anyway the key is you have to get a feel for it. Don't go full blast, but be sure the water coming out is picking up a fair amount of heat. In time you'll get a sense of the best way to run it. When I started using an IC I though full blast was the best way, but I was running a lot of water and not getting much chilling. I slowed things down and it went much faster.
Swirling the IC around in the wort occasionally (when it wasn't too hot to touch) also makes the process go faster.