Logo tutelle


Sur ce site

Accueil du site > Productions et Résultats > Tutorial à DAC 2015

Tutorial à DAC 2015

Synopsys a participé au Tutorial suivant de la DAC 2015 :

Alan Gibbons (Synopsys), Vita Vishnyakov, Krishna Sekar, Architecture and Design of Energy Efficient Heterogeneous Multi-Core Platforms, Tutorial 7, Track : Embedded Systems, june 8, DAC 2015, San Francisco.

Abstract :

Energy efficiency is one of the primary design metrics for heterogeneous multi-core mobile compute platforms. These compute platforms comprise complex software and hardware architectures as well as multi-layered system power management and the energy efficiency of the platform will be determined by the way in which these components interact under real software load.

Decisions made early in the platform development cycle will yield the greatest impact in terms of energy efficiency, hence it is essential to address this at the system level. Energy aware system level design techniques and methodologies enable concurrent optimization of the hardware, software and system power management aspects of the platform. Since the way in which a platform is used determines its energy consumption, it is vital that we understand the energy profile of the platform under software load. Once there is visibility into this energy profile, intelligent architecture and design decisions can then be made that consider the associated performance and power trade-offs to improve energy efficiency of the platform.

IP power models can provide this level of visibility into the energy behavior of the platform and are critical in enabling this design methodology. The use of IP power models during system level design allows exploration of the power impact of various hardware and software architectures, as well as to dynamically tune system resources and power management strategies for optimal energy efficient performance.

This tutorial will highlight the challenges associated with energy efficient platform design, the energy aware system level design tools and methodologies that can be used to deal with these challenges, as well as the IP power characterization and modelling approaches that enable these methodologies.