Observation for << The program decision logic approach to predicated execution>>
       The ideal senario for compiler is let programmer only care for
correctness and understandability, and compiler cares for the efficiency. Dynamic scheduling is
one of this ideas. The approach in this paper push this idea to a new extent and got
very promising result.
       In Section 4, when C0 = (32>=r4), C1 = (R4>=127), C3 = (r4!=10)  , we
have
p6 = C0C3' + C0'C1C3'. According to logic property, we can get
p6 = (C0+C0'C1)C3' = (C0+C1)C3'. How the program can further reason this
and get p6 = C3' ? There is more than just logic simpilcaiton here. How they
look into the inside of the logic definition of C0, C1 and C3?