[Next] [Up]

Parallel Processing

In FALCON, as in PL/I and APL, it is possible to multiply every element in one array by the corresponding element of another array with a single multiplication operator, instead of using the equivalent of a FOR loop.

This makes it easier to generate code for one modern type of computer called a 'vector processor', which can fetch several numbers from memory at the same time, then perform the same arithmetic operation on all of them. These computers are also known as SIMD architectures: single instruction stream, multiple data stream. Such computers belong to a family of modern computer designs that try to achieve faster operation by doing more than one thing at a time.

Several other types of parallel processing are possible, and even in the SIMD case, it is possible, with special statements, to make more extensive use of the available parallel processing capabilities with additional statements designed for that purpose. This chapter describes the statements that work with the various forms of parallel processing.


[Next] [Up]