Abstract
In this paper we derive a parallel program for a nontrivial problem, viz. computing the Dirichlet convolution of two arithmetical functions. The derivation proceeds in a calculational manner. Advantages of this method are that design decisions can be stated clearly and that operational reasoning about programs can be avoided.
A program for the inverse convolution problem is also presented, as well as a program for computing the Möbius sequence, which is an instance of the inverse convolution problem.
Original language | English |
---|---|
Pages (from-to) | 185-200 |
Journal | Science of Computer Programming |
Volume | 15 |
Issue number | 2-3 |
DOIs | |
Publication status | Published - 1990 |