Download
1 / 13
160 likes | 295 Views
TOPIC : D-frontier and PODEM. UNIT 3 : Testing and test generation. Module 3.3 Test generation methodology. Introduction. There are various methods of arriving at the vector that can detect the faulty circuit. Out of these methods two methods which we are discussing here are :
E N D
TOPIC : D-frontier and PODEM UNIT 3 : Testing and test generation Module 3.3 Test generation methodology
Introduction • There are various methods of arriving at the vector that can detect the faulty circuit. • Out of these methods two methods which we are discussing here are : • D- Frontier : It propagation of D or Dbar to the Pos and then justifying the Pis. • PODEM : Path oriented decision making.
D- Frontier • The D-frontier consists of all gates whose output value is currently x but have one or more error signals (either D's or D's) on their inputs. • Error propagation consist of selecting one gate from the D-frontier and assigning values to the unspecified gate inputs so that the gate output becomes D or D.
J-frontier • To keep track of the currently unsolved line-justification problems, we use a set called the J-frontier, which consists of all gates whose output value is known but is not implied by its input values. • Let c be the controlling value and i be the inversion of a gate on the J-frontier. Then the output value is c xor i, at least two inputs must have value x, and no input can have value c.
Example explanation • In the previous example it may be seen that the gate5 has a D at the output which has been propagated to ‘S’ by applying input values to other gates. • In general, one may find the gates with D as the input. All the gates containing D or D_bar at the input generated by any error are called D-frontier. This concept of D-frontier can be made use to propagate the error (D or D_bar) to the output.
Error propagation in D frontier • It consists of selecting 1 gate from the D-frontier and identifying a gate to which this D is connected as input. • Then assigning values to the other inputs invariably C_bar, so that the D is propagated. • This process is to be repeated till the output points are reached. If the D-frontier cannot be propagated, D-frontier is said to be empty beyond the last D-frontier. • Once the empty D-frontier is encountered backtracking should be resorted to propagate D-frontier through another path.
Example of Error propagation • Let ‘b’ input be s-a-0. One has to apply b=1 which becomes in the five valued logic as D. • Gate1 becomes D-frontier. This can be propagated to ‘d’ by assuming a=1. This D-frontier has to be moved through d-g • gate3 should become D-frontier by making e=1. Thus the D has been propagated through the path ‘bdg’. • The vectors that are to be applied for this purpose are a=1,b=1,c=0.
PODEM introduction • PODEM (Path-Oriented Decision Making) is a TG algorithm characterized by a direct search process, in which decisions consist only of PI assignments. • We have seen that the problems of fault activation and error propagation lead to sets of line-justification problems.
PODEM methodology • PODEM treats a value Vkto be justified for line k as an objective (k,Vk) to be achieved via PI assignments. • A backtracing procedure (next slide) maps a desired objective into a PI assignment that is likely to contribute to achieving the objective.
PODEM algortihm Backtrace (k,vk) /* map objective into PI assignment */ begin v =vk while k is a gate output begin i = inversion of k select an input (j) of k with value x v =v xor i k=j end I* k is a PI */ return (k,v) end
