Investigating Real Power Usage on Red Storm
نویسندگان
چکیده
High Performance Computing (HPC) has historically been the impetus for new technologies. However, in dealing with power related issues HPC has lagged behind. Power has recently been recognized as one of the major obstacles to fielding a Peta-Flop class system. In this paper we will discuss power related topics by leveraging what is currently a rare capability of examining real power usage at a very granular level on an HPC platform. We will discuss what we have observed, initial efforts to implement power conserving measures, quantifications and comparisons between Light Weight and General Purpose Operating Systems and, finally, areas of investigation that have been enabled by this novel capability.
منابع مشابه
Red Storm Capability Computing Queuing Policy
Red Storm will be the first Tri-Lab [Sandia National Laboratories (SNL), Los Alamos National Laboratory (LANL), and Lawrence Livermore National Laboratory (LLNL)], U.S. Department of Energy/National Nuclear Security Administration, Advanced Simulation and Computing (ASC) platform to be managed under an explicit capability computing policy directive. Instead of allocating nodes among SNL:LANL:LL...
متن کاملChopper-Based Real-Time Load Emulator with Feed-Forward and Hysteresis Current Controller
Nowadays, the usage of load emulators, is the best method for implementation and analysis of different electrical load change scenarios in laboratories. This paper presents an improved programmable load emulator which can emulate both reference active and reactive power simultaneously. The proposed control system can track the dynamic load changes rapidly and accurately in addition to pulse cha...
متن کاملRed Storm Systems Management: Topics on Extending Current Capabilities
Systems like the Red Storm system at Sandia Laboratories provide grand challenges for developers of systems software. While many of the challenges associated with systems like Red Storm are experienced on other systems, like commodity clusters, the sheer magnitude of the system magnifies issues that may seem trivial on less complex systems. The value of taking an integrated approach to many of ...
متن کاملScalable Collection of Large MPI Traces on Red Storm
Gathering large MPI traces and statistics is important for performance analysis and trouble shooting of applications. Traces, with detailed information about each single message an application has sent, are crucial to characterize the message passing behavior of an application. On massively parallel systems like Red Storm the amount of data collected impacts the performance and behavior of the ...
متن کاملA Comparison of Three MPI Implementations for Red Storm
Cray Red Storm is a new distributed memory massively parallel computing platform designed to scale to tens of thousands of nodes. Red Storm has a custom network designed around the Cray SeaStar network interface and router. In this paper, we present an evaluation of three different MPI implementations for Red Storm: the vendor-supported MPICH2 implementation, and two other implementations based...
متن کامل