[Abstract]
While the average power is reduced by finding the optimal signal activities (probabilities of signal switching) at the primary inputs, the peak power is reduced by restricting the number of active primary inputs
[1. introduction]
For ISCAS benchmark circuits,
while the average of power reduction by using DS-LFSR is 19% [7], the average of power reduction by using the low power random testing technique in [5] is as high as 78%.
At any time instant, about half of the primary inputs are fi xed to their preceding values, and only half of the primary inputs are made active. Hence the instanta-
noeus power is always reduced.
[2. Preliminaries and definitions]
The variable mx is the number of switching of x(t) in the time interval [-tao/2, tao/2]
