SC16 Working Group Presentations

Energy Efficiency Considerations and HPC Procurement

 

ABSTRACT:  The predominant goal for procurement of HPC systems is to identify the optimal solution to both technical and financial targets that maximizes the contribution of that system to the organization’s mission. Beyond the acquisition cost of the system, it is also important to consider the total costs of ownership, including the improvements necessary to host the system, the infrastructure that supports its operation, and the significant operational costs associated with that new HPC system. In this BoF, HPC leaders will discuss and debate key procurement requirements and lessons learned that can contribute to greater energy efficiency and reduced operational costs.

 

The Green500: Trends for Energy-Efficient Supercomputing

 

ABSTRACT:  With power a first-order design constraint on par with performance, it is important to measure and analyze trends for energy-efficient supercomputing. This BoF will discuss trends across the Green500 and highlights from the latest Green500 list. The Green500 and TOP500, in collaboration with the Energy-Efficient HPC WG, have a newly integrated submission process designed to streamline future submissions and to provide a consistent set of data for the historical record. It will also provide a forum for community review of the integrated submission process. The BoF will close with an awards presentation, recognizing the most energy-efficient supercomputers in the world.

 

 

Date: November 13, 2016

 

Opening Remarks & Agenda Review - M. Patterson

 

 

State of the WG - N. Bates

 

 

Future-proofing your DC Design

 

 

Energy efficiency economics

 

 

Architectural

 

 

System software

 

 

Closing

 

 

Birds of Feather

 

 

Session Panel

 

 

Abstract:

 

Power requirements for high-performance computers far surpass those of enterprise-servers, creating substantial loads and challenges. With enterprise racks drawing 4-15kW, their HPC counterparts can draw 40-150kW. Typical servers require 120volts, single-phase to operate, whereas high-performance-computers up the ante to 480volts, three-phase. This allows manageable wire sizes and improved efficiency. Alternating current (ac) is ubiquitous, but is direct current (dc) more efficient, reliable and less expensive? Will this trend accelerate with use of renewables? Many high-performance-computers use high-voltage dc internally; why not in the data center?

 

The implications for these new power distribution systems demand major shifts for both the infrastructure and operations of the data-center. Are we prepared for this shift? Will dc be compelling enough to drive the eco-system? What needs to be done to bridge the gap and ease the transition? This panel promises a lively discussion with experts in power distribution from the vendor and user community.

Lawrence Livermore National Laboratory

7000 East Avenue • Livermore, CA 94550

Operated by Lawrence Livermore National Security, LLC, for the Department of Energy's National Nuclear Security Administration

LLNL-WEB-670983  |  Privacy & Legal Notice | Site Search

Energy Efficient HPC Working Group

Supercomputing Conference 2016