Section: New Results

Energy-efficient Computing Infrastructures

Participants : Christine Morin, Anne-Cécile Orgerie, Martin Quinson.

Simulating the impact of DVFS within SimGrid

Participants : Christine Morin, Anne-Cécile Orgerie, Martin Quinson.

Simulation is a a popular approach for studying the performance of HPC applications in a variety of scenarios. However, simulators do not typically provide insights on the energy consumption of the simulated platforms. Furthermore, studying the impact of application configuration choices on energy is a difficult task, as not many platforms are equipped with the proper power measurement tools. The goal of this work is to enable energy-aware experimentation within the SimGrid simulation toolkit, by introducing a model of application energy consumption and enabling the use of Dynamic Voltage and Frequency Scaling (DVFS) techniques for the simulated platforms. We provide the methodology used to obtain accurate energy estimations, highlighting the simulator calibration phase. The proposed energy model is validated by means of a large set of experiments featuring several benchmarks and scientific applications. This work is available in the latest SimGrid release. This work is done in collaboration with the Mescal team from LIG in Grenoble. A paper is currently under preparation on this work.

Simulating Energy Consumption of Wired Networks

Participants : Timothée Haudebourg, Anne-Cécile Orgerie.

Predicting the performance of applications, in terms of completion time and resource usage for instance, is critical to appropriately dimensioning resources that will be allocated to these applications. Current applications, such as web servers and Cloud services, require lots of computing and networking resources. Yet, these resource demands are highly fluctuating over time. Thus, adequately and dynamically dimensioning these resources is challenging and crucial to guarantee performance and cost-effectiveness. In the same manner, estimating the energy consumption of applications deployed over heterogeneous cloud resources is important in order to provision power resources and make use of renewable energies. Concerning the consumption of entire infrastructures, some studies show that computing resources represent the biggest part in the Cloud’s consumption, while others show that, depending on the studied scenario, the energy cost of the network infrastructure that links the user to the computing resources can be bigger than the energy cost of the servers.

In this work, we aim at simulating the energy consumption of wired networks which receive little attention in the Cloud computing community even though they represent key elements of these distributed architectures. To this end, we are contributing to the well-known open-source simulator ns3 by developing an energy consumption module named ECOFEN.

In 2015, this simulator has been extended to integrate two more green levers: low power idle (IEEE 802.3az) and adaptive link rate. This work has been done during the internship of Timothée Haudebourg (L3 ENS Rennes) and a publication is currently under preparation.

Multicriteria scheduling for large-scale HPC environments

Participant : Anne-Cécile Orgerie.

Energy consumption is one of the main limiting factor for the design and deployment of large scale numerical infrastructures. The road towards "Sustainable Exascale" is a challenge with a target of 50 Gflops per watt. Energy efficiency must be taken into account and must be combined with other criteria like performance, resilience, Quality of Service.

As platforms become more and more heterogeneous (co-processors, GPUs, low power processors...), an efficient scheduling of applications and services at large scale remains a challenge. In this context, we will explore and propose a multicriteria scheduling model and framework for large scale HPC systems. Based on real energy measurements and calibrations, we will propose some performance and energy models and will build a multi criteria scheduler. Simulation on selected scenario will be explored and a prototype will be designed for ensuring experimental validation.

This work is done in collaboration with ROMA and Avalon teams from LIP in Lyon.