Eliminate False-Negative Methods Base On ATPG

Equivalence verification is a method to ensure the correctness of circuitwhen circuit was designed. have very important position in the integratedcircuit design. Now,there are a lot of verification technology. but due to willappear many problems when the related techniques used in the validationprocess.so,it is becomes a current and important subject of research that solvethe problem generated when we use of related techniques to verify equivalencecircuit. for example, the false-negative is easy to produce when using themethod of cut-point based, so it will become the main problem to improve theefficiency of verification that how to solve the problem of false-negativewhen using the method of cut-point based to verify equivalence circuit.In this paper we puts forward a method to eliminate the false-negative byusing the weighted value to improving ATPG algorithm. Mainly is to use the methodto improving some faults of PODEM algorithm.When false-negative.we can usePODEM algorithm of ATPG technique to counter counter example.but the PODEMalgorithm is assignment at the original input and will be also going back tothe original input when a conflict taken place,so the efficiency is not high.So we combined the method of weighted value to find a circuit of the internalnodes as PODEM start to input and back. it will reduce number of times to improvethe efficiency in backtracking.Weighting method is mainly based on the structureof the circuit,for each internal node and the original input of the circuit todistribute the weights and the allocation principles will be introduced in thearticle. To make each node has a right to their own weights, we can use theseweights to compare original input as well as the importance of the internal nodein a circuit, so as to find the biggest influence internal nodes and the originalinput on circuit and this point as a key to operate node in circuit. Weighting method combined with structure characteristics of the circuit.this method canimproving many algorithm to solve some problems that the algorithm itself existed.In PODEM algorithm also has many defects, such as assignment and backtrackingmust to the original input, assignment and backtracking path is not the best.So we use the internal nodes as input for the assignment and backtracking withthe weighted method, and find the sorting of weights. We use the highest weightsof the input as the first assignment. so as to find and resolve conflicts assoon as possible.But the method also has some shortcomings that using weights to improvePODEM.The efficiency is very low under the condition of the circuit is realinequivalence.because we need to assignment,propagation and backtracking forall input node of circuit under the condition of the circuit is realinequivalence.we can only be concluded two circuit is inequivalence after allnodes have verified.so the efficiency is lower compared with otherauthentication methods.