Test-set partitioning for multi-weighted random LFSRs

Weighted random LFSRs are widely used for reproducing a deterministic test-set, Large test sets are partitioned into subsets and more than one weight-sets is necessary, one weight-set per subset. It has been proposed to partition by taking into consideration the Hamming distance between pairs of vectors. This paper introduces a systematic approach for partitioning a given deterministic test-set into subsets so that the reproduction time is minimized. It presents and evaluates properties that each set in the partition must satisfy. Efficient algorithms are also proposed for partitioning the test-set into subsets that satisfy the proposed criteria. Experimental studies on the ISCAS'85 benchmarks show significant improvement over previous work. (C) 2000 Elsevier Science B.V. All rights reserved.