Tool Overview

TORSCHE (Time Optimisation of Resources, SCHEduling) is a MATLAB-based toolbox including various scheduling algorithms, that are used for various applications as high level synthesis of parallel algorithms, optimized production of manufacturing lines, etc. Using the toolbox, one can easily and quickly obtain an optimal code of computing intensive applications running on specific hardware architectures. The tool can also be used to investigate application performance prior to its implementation and to use these valeus (e.g. the shortest achievable sampling period of the filter implemented on given set of processors) in the control system design process performed in Matlab/Simulink. The main contribution of the toolbox, which is built on well-known disciplines of the graph theory and operation research, is to make it easy to apply this type of reasoning to a wide range of problems. Many of them are combinatorial optimisation problems, and as such they are challenging from the theoretical point of view. TORSCHE offers a collection of MATLAB routines that allow the user to formalize the scheduling problem, while considering appropriate configuration of resources (e.g. HW architecture performing filter algorithm), task parameters (e.g. deadlines, release dates, preemption) and optimisation criterion (e.g. makespawn minimisation, maximum lateness minimisation). The toolbox enables to solve these problems by their reformulation or to solve them directly while choosing appropriate scheduling algorithm. The input data of the problem instance are typically represented by an oriented graph and the output data are represented by a Gantt chart. The input data might be automatically generated from the problem description (e.g. equations of the filter algorithm) and output data, the schedule, may be used to automatically generate an implementation of embedded system (e.g. parallel code for dedicated processing units).

More information on TORSCHE (p.64)
TORSCHE website

Réalisation Axome - Création de sites Internet