NEW ENCODING DECODING METHODS FOR DESIGNING FAULT-TOLERANT MATRIX OPERATIONS/

Algorithm-based fault tolerance (ABFT) can provide a low-cost error pr otection for array processors and multiprocessor systems. Several ABFT techniques (weighted check-sum) have been proposed to design fault-to lerant matrix operations. In these schemes, encoding/decoding uses eit her multiplications or divisions so that overhead is high. In this pap er, new encoding/decoding methods are proposed for designing fault-tol erant matrix operations. The unique feature of these new methods is th at only additions and subtractions are used in encoding/decoding. In t his paper, new algorithms are proposed to construct error detecting/co rrecting codes with the minimum Hamming distance 3 and 4. We will show that the overhead introduced due to the incorporation of fault tolera nce is drastically reduced by using these new coding schemes.