Do-loop-surface: An abstract performance data visualization

Abstract

Performance is a critical issue in current massively parallel processors. However, to get adequate performance is not an easy task and performance tools are required in order to help the programmer to understand the behaviour of a parallel program. In recent years, a wide variety of tools have been developed for this purpose including tools for monitoring and evaluating performance and visualization tools. However, these tools do not provide an abstract representation of performance. Massively parallel processors could produce a huge amount of performance data and sophisticated methods for representing and displaying this data are required. The Do-Loop-Surface (DLS) display is an abstract representation of the performance of a particular do-loop in a program and it is implemented using AVS, a Data Visualization Tool. In this paper, this new representation is presented and used in order to improve the performance of a Matrix Multiply parallel algorithm.