This paper is a study of the practical value of the various approaches
, proposed so far, to achieve fault tolerance in arithmetic units usin
g the residue code technique and to find the best design in terms of c
ost and error coverage. The results have shown that the error correcti
on approaches can treat a one-bit error (E = +/- 2(i)) using relativel
y small hardware cost and time delay. It is also shown that no more th
an a single error, of the one-bit type, can be treated at reasonable c
ost using error correction approaches. However, a combination of N-mod
ular redundancy (NMR) and residue codes can be used to cover a wide ra
nge of errors at considerably lower cost. Here we suggest an approach
based on model duplication and residue codes (DAR) which is shown to h
ave better error coverage than error correction approaches and lower c
ost than both triple modular redundancy (TMR) and error correction app
roaches.