Quote:
It would take fps_max being above 16,384 and actually having the frame rate like this in order to run into these problems.
You're right, my example is too simple to give a divide by zero error at anything less than a fps > 16000.
But take acceleration which is length/time^2:
Code:
a = dist * (time_step*time_step);
At 1024fps, time_step = 0.015625, time_step^2 = 0.000244140625!
At 640fps, time_step^2 = 0.000625