Implementation of Apply on a transputer array

Abstract

SUMMARYApply is a machine-independent, low-level image proressing language for expressing local window operations. It has two main advantages: (1) It significantly reduces the prcqpuumiq effort, and (2) it can be mapped onto a wide range of parallel computers. In this paper we report our recent experience on implementing Apply on a Meiko Computing Surface (transputer array machine) using a farmer/ga&er model. The performance of the Meiko Implementation on a number of edge detection algorithms includhg the papular Canny operator is analysed. Results suggest that the perfomance of Apply generated code represents peak performance and it is likely to be as efficient as hand-written code.

PARADOX—a heterogeneous machine for the 3D vision algorithm

“…On an MIMD computer, such as a transputer network, the replication can be most easily accomplished by mapping each process r onto one processor. The throughput of such a system is n times greater than a single processor minus the input/output overhead associated with partitioning and delivering images [6,17].…”

“…A wide range of network topologies can be implemented, including parallel one-dimensional arrays [6], a 2D array or a ring structure. This board delivers a peak performance of 320 MIPS.…”