(eng) In this paper we deal with a new high precision computation of the dot product. The key idea is to use hundreds of digit-serial arithmetic units that allow a massive digit-level pipelining. Parallel discrete-event ...