A floating point technique for extending the available precision (1971)

by T J Dekker
Venue:Numerische Mathematik