MPP implementation of abstract data parallel architectures for declarative programming languages

A method of implementing declarative data structures efficiently on a data parallel architecture is illustrated by the implementation of a data parallel algorithm for functional arrays on the massively parallel processor (MPP). Functional arrays are defined, and it is shown why the basic operations on them are slow when implemented on a sequential machine. Functional arrays cannot be implemented efficiently on conventional architectures without severely restricting the way a program can access the array. Conventional unrestricted algorithms typically require O(log n) time to update an array and access an element. The MPP algorithm performs these operations in a constant time of about 100 μs