Modeling of applications and hardware to explore task mapping and scheduling strategies on a heterogeneous micro-server system
Abstract
Many of todays important applications of our everyday lives, e.g. weather forecast, design of plane and car shapes, medical analysis or even search engine queries depend on massively-parallel computer programs that are executed in data centers hosting thousands of computers. A large amount of electrical energy is used to power them, and it is of primary importance to compute more efficiently to sustain the increasing demand of computing power while keeping energy consumption reasonable. One promising research path in this domain is heterogeneous systems. The rationale for that is that at least parts of applications execute more efficiently depending on the computing resource (processors, accelerators, etc.). Nevertheless, the exploitation of these heterogeneous platforms raises new challenges in terms of application management optimization on available computing resources. The aim of our work is to determine effective algorithms to exploit these heterogeneous platforms by finding the best mapping and scheduling of an application to optimize the execution time and energy consumption with respect to various constraints. To achieve this goal, there is a need of a detailed modeling of the applications and the underlying hardware to be able to find realistic solutions. In this paper, we propose such as model, provide two implementations with state-of-the-art tools and preliminary mapping and scheduling numerical results.
Keywords
Application programs
Computer hardware
Energy utilization
Green computing
Hardware
Mapping
Search engines
Weather forecasting
Application management
Effective algorithms
Heterogeneous computing
Heterogeneous platforms
Heterogeneous systems
Massively parallel computers
Resource management
Scheduling strategies
Scheduling