Turbo is a tradeoff . . .
At high temperatures, the life of the LED and other semiconductors will decrease, but in Turbo mode, the battery life is also a lot less.
Actually, I would be much more concerned about running the lithium batteries that hard.
There have been a number of very serious battery explosions in Fenix lights after running 3 volt Lithiums on Turbo for extended periods (and a few for no reason at all, or after the light was dropped, so I guessing all the vibration on a bike would NOT be good for them).
For daily use involving extended runs in Turbo, I think a NiMH based light would be a much safer choice.
As to your P3D heating up, I think the reason that it seems to heat up more is simply that Fenix is taking advantage of the fact that unlike some other LED's with similar efficiency, the Rebel can handle higher drive levels and still produce useful increases in light output.
Where some earlier LED's tend to 'top-out' the rebel can pump out up to about 200 lumens if you drive it hard enough.
But more drive means more heat, it's that simple.
A perfect LED would run at 683 Lumens/watt and emit only visable light in the useful portion of the spectrum with no IR, UV, or heat as a byproduct.
http://en.wikipedia.org/wiki/Luminous_efficacy
Unfortunately, we are not even close to that figure (and probably never will be).
At just over 100 lumens/watt at high power levels, the vast majority of energy you pump into a LED still turns to heat.
This is physics folks, and the premium LEDs from Lumiled, CREE, and Seoul are neck and neck in the Lumens/Watt race, so you can see it's just a little silly to talk about one brand of LED heating up more than another, because everything that doesn't turn into light WILL turn into heat.
The suggestion to retake the pictures with a much dimmer exposure, is a very good one.
One fellow here made the comment that his P3D 'only' seemed to measure about 1EV brighter on turbo and wasn't brighter in the lower power settings.
The lower power settings are regulated to hold a constant output and aren't really relevant, and the 1EV change in brightness he saw on his light meter in Turbo is DOUBLE the output, indicating the difference between 100 lumens and 200 lumens!
So basically he was complaining that his P3D Rebel light was doing EXACTLY what it was designed to do.