8 Matching Results

Search Results

Advanced search parameters have been applied.

Moving Large Data Sets Over High-Performance Long Distance Networks

Description: In this project we look at the performance characteristics of three tools used to move large data sets over dedicated long distance networking infrastructure. Although performance studies of wide area networks have been a frequent topic of interest, performance analyses have tended to focus on network latency characteristics and peak throughput using network traffic generators. In this study we instead perform an end-to-end long distance networking analysis that includes reading large data sets from a source file system and committing large data sets to a destination file system. An evaluation of end-to-end data movement is also an evaluation of the system configurations employed and the tools used to move the data. For this paper, we have built several storage platforms and connected them with a high performance long distance network configuration. We use these systems to analyze the capabilities of three data movement tools: BBcp, GridFTP, and XDD. Our studies demonstrate that existing data movement tools do not provide efficient performance levels or exercise the storage devices in their highest performance modes. We describe the device information required to achieve high levels of I/O performance and discuss how this data is applicable in use cases beyond data movement performance.
Date: April 1, 2011
Creator: Hodson, Stephen W; Poole, Stephen W; Ruwart, Thomas & Settlemyer, Bradley W
Partner: UNT Libraries Government Documents Department

Diagnosing Anomalous Network Performance with Confidence

Description: Variability in network performance is a major obstacle in effectively analyzing the throughput of modern high performance computer systems. High performance interconnec- tion networks offer excellent best-case network latencies; how- ever, highly parallel applications running on parallel machines typically require consistently high levels of performance to adequately leverage the massive amounts of available computing power. Performance analysts have usually quantified network performance using traditional summary statistics that assume the observational data is sampled from a normal distribution. In our examinations of network performance, we have found this method of analysis often provides too little data to under- stand anomalous network performance. Our tool, Confidence, instead uses an empirically derived probability distribution to characterize network performance. In this paper we describe several instances where the Confidence toolkit allowed us to understand and diagnose network performance anomalies that we could not adequately explore with the simple summary statis- tics provided by traditional measurement tools. In particular, we examine a multi-modal performance scenario encountered with an Infiniband interconnection network and we explore the performance repeatability on the custom Cray SeaStar2 interconnection network after a set of software and driver updates.
Date: April 1, 2011
Creator: Settlemyer, Bradley W; Hodson, Stephen W; Kuehn, Jeffery A & Poole, Stephen W
Partner: UNT Libraries Government Documents Department

Open Standards for Sensor Information Processing

Description: This document explores sensor standards, sensor data models, and computer sensor software in order to determine the specifications and data representation best suited for analyzing and monitoring computer system health using embedded sensor data. We review IEEE 1451, OGC Sensor Model Language and Transducer Model Language (TML), lm-sensors and Intelligent Platform Management Inititative (IPMI).
Date: July 1, 2009
Creator: Pouchard, Line Catherine; Poole, Stephen W & Lothian, Josh
Partner: UNT Libraries Government Documents Department

Graph Generator Survey

Description: The benchmarking effort within the Extreme Scale Systems Center at Oak Ridge National Laboratory seeks to provide High Performance Computing benchmarks and test suites of interest to the DoD sponsor. The work described in this report is a part of the effort focusing on graph generation. A previously developed benchmark, SystemBurn, allowed the emulation of dierent application behavior profiles within a single framework. To complement this effort, similar capabilities are desired for graph-centric problems. This report examines existing synthetic graph generator implementations in preparation for further study on the properties of their generated synthetic graphs.
Date: December 1, 2013
Creator: Lothian, Josh; Powers, Sarah S; Sullivan, Blair D; Baker, Matthew B; Schrock, Jonathan & Poole, Stephen W
Partner: UNT Libraries Government Documents Department

OpenSHMEM Application Programming Interface, v1.0 Final

Description: This document defines the elements of the OpenSHMEM Application Programming Interface. The purpose of the OpenSHMEM API is to provide programmers with a standard interface for writing parallel programs using C, C++ and Fortran with one-sided communication.
Date: January 1, 2012
Creator: Kuehn, Jeffery A; Chapman, Barbara; Curtis, Anthony R; Mauricio, Ricardo; Pophale, Swaroop; Nanjegowda, Ramachandra et al.
Partner: UNT Libraries Government Documents Department

SystemBurn: Principles of Design and Operation Release 3.0

Description: As high performance computing technology progresses toward the progressively more extreme scales required to address critical computational problems of both national and global interest, power and cooling for these extreme scale systems is becoming a growing concern. A standardized methodology for testing system requirements under maximal system load and validating system environmental capability to meet those requirements is critical to maintaining system stability and minimizing power and cooling risks for high end data centers. Moreover, accurate testing permits the high end data center to avoid issues of under- or over-provisioning power and cooling capacity saving resources and mitigating hazards. Previous approaches to such testing have employed an ad hoc collection of tools, which have been anecdotally perceived to produce a heavy system load. In this report, we present SystemBurn, a software tool engineered to allow a system user to methodically create a maximal system load on large scale systems for the purposes of testing and validation.
Date: September 1, 2012
Creator: Dobson, Jonathan D; Kuehn, Jeffery A; Poole, Stephen W; Hodson, Stephen W; Glandon, Steven R; Reister, David B et al.
Partner: UNT Libraries Government Documents Department

SystemBurn: Principles of Design and Operation, Release 2.0

Description: As high performance computing technology progresses toward the progressively more extreme scales required to address critical computational problems of both national and global interest, power and cooling for these extreme scale systems is becoming a growing concern. A standardized methodology for testing system requirements under maximal system load and validating system environmental capability to meet those requirements is critical to maintaining system stability and minimizing power and cooling risks for high end data centers. Moreover, accurate testing permits the high end data center to avoid issues of under- or over-provisioning power and cooling capacity saving resources and mitigating hazards. Previous approaches to such testing have employed an ad hoc collection of tools, which have been anecdotally perceived to produce a heavy system load. In this report, we present SystemBurn, a software tool engineered to allow a system user to methodically create a maximal system load on large scale systems for the purposes of testing and validation.
Date: January 1, 2012
Creator: Kuehn, Jeffery A; Poole, Stephen W; Hodson, Stephen W; Lothian, Josh; Dobson, Jonathan D; Reister, David B et al.
Partner: UNT Libraries Government Documents Department