A common technique widely used to achieve fault tolerance in systolic
arrays consists in incorporating in the array additional processing el
ements (PEs) and extra bypass links. Given a sufficient number of PEs
and a large enough set of bypass links, it might seem that the array c
an easily tolerate a large number of faults provided they do not occur
in consecutive locations. It is not always the case as shown in this
paper. In fact, certain fault patterns exist and may occur which would
prevent any kind of restructuring of the aray, thus making the struct
ure unusable. For a given set of bypass links from each PE in the arra
y, it is possible to identify such fault patterns which will prevent a
ny reconfiguration. In this paper, we identify the class of fault patt
erns that are catastrophic for linear systolic arrays, examine their c
haracteristics, and describe a method for constructing such fault patt
erns.