skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Supercomputer and cluster performance modeling and analysis efforts:2004-2006.

Abstract

This report describes efforts by the Performance Modeling and Analysis Team to investigate performance characteristics of Sandia's engineering and scientific applications on the ASC capability and advanced architecture supercomputers, and Sandia's capacity Linux clusters. Efforts to model various aspects of these computers are also discussed. The goals of these efforts are to quantify and compare Sandia's supercomputer and cluster performance characteristics; to reveal strengths and weaknesses in such systems; and to predict performance characteristics of, and provide guidelines for, future acquisitions and follow-on systems. Described herein are the results obtained from running benchmarks and applications to extract performance characteristics and comparisons, as well as modeling efforts, obtained during the time period 2004-2006. The format of the report, with hypertext links to numerous additional documents, purposefully minimizes the document size needed to disseminate the extensive results from our research.

Authors:
; ;  [1]; ;  [2]; ; ; ; ; ; ; ; ; ; ; ; ; ;
  1. (Hal) Edward
  2. (.,
Publication Date:
Research Org.:
Sandia National Laboratories
Sponsoring Org.:
USDOE
OSTI Identifier:
903425
Report Number(s):
SAND2007-0601
TRN: US200722%%49
DOE Contract Number:
AC04-94AL85000
Resource Type:
Technical Report
Country of Publication:
United States
Language:
English
Subject:
99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; BENCHMARKS; PERFORMANCE; SUPERCOMPUTERS; COMPUTER ARCHITECTURE; MATHEMATICAL MODELS; Cluster analysis.; Supercomputers-Programming.

Citation Formats

Sturtevant, Judith E., Ganti, Anand, Meyer, Harold, Stevenson, Joel O., Benner, Robert E., Jr., .), Goudy, Susan Phelps, Doerfler, Douglas W., Domino, Stefan Paul, Taylor, Mark A., Malins, Robert Joseph, Scott, Ryan T., Barnette, Daniel Wayne, Rajan, Mahesh, Ang, James Alfred, Black, Amalia Rebecca, Laub, Thomas William, Vaughan, Courtenay Thomas, and Franke, Brian Claude. Supercomputer and cluster performance modeling and analysis efforts:2004-2006.. United States: N. p., 2007. Web. doi:10.2172/903425.
Sturtevant, Judith E., Ganti, Anand, Meyer, Harold, Stevenson, Joel O., Benner, Robert E., Jr., .), Goudy, Susan Phelps, Doerfler, Douglas W., Domino, Stefan Paul, Taylor, Mark A., Malins, Robert Joseph, Scott, Ryan T., Barnette, Daniel Wayne, Rajan, Mahesh, Ang, James Alfred, Black, Amalia Rebecca, Laub, Thomas William, Vaughan, Courtenay Thomas, & Franke, Brian Claude. Supercomputer and cluster performance modeling and analysis efforts:2004-2006.. United States. doi:10.2172/903425.
Sturtevant, Judith E., Ganti, Anand, Meyer, Harold, Stevenson, Joel O., Benner, Robert E., Jr., .), Goudy, Susan Phelps, Doerfler, Douglas W., Domino, Stefan Paul, Taylor, Mark A., Malins, Robert Joseph, Scott, Ryan T., Barnette, Daniel Wayne, Rajan, Mahesh, Ang, James Alfred, Black, Amalia Rebecca, Laub, Thomas William, Vaughan, Courtenay Thomas, and Franke, Brian Claude. Thu . "Supercomputer and cluster performance modeling and analysis efforts:2004-2006.". United States. doi:10.2172/903425. https://www.osti.gov/servlets/purl/903425.
@article{osti_903425,
title = {Supercomputer and cluster performance modeling and analysis efforts:2004-2006.},
author = {Sturtevant, Judith E. and Ganti, Anand and Meyer, Harold and Stevenson, Joel O. and Benner, Robert E., Jr. and .) and Goudy, Susan Phelps and Doerfler, Douglas W. and Domino, Stefan Paul and Taylor, Mark A. and Malins, Robert Joseph and Scott, Ryan T. and Barnette, Daniel Wayne and Rajan, Mahesh and Ang, James Alfred and Black, Amalia Rebecca and Laub, Thomas William and Vaughan, Courtenay Thomas and Franke, Brian Claude},
abstractNote = {This report describes efforts by the Performance Modeling and Analysis Team to investigate performance characteristics of Sandia's engineering and scientific applications on the ASC capability and advanced architecture supercomputers, and Sandia's capacity Linux clusters. Efforts to model various aspects of these computers are also discussed. The goals of these efforts are to quantify and compare Sandia's supercomputer and cluster performance characteristics; to reveal strengths and weaknesses in such systems; and to predict performance characteristics of, and provide guidelines for, future acquisitions and follow-on systems. Described herein are the results obtained from running benchmarks and applications to extract performance characteristics and comparisons, as well as modeling efforts, obtained during the time period 2004-2006. The format of the report, with hypertext links to numerous additional documents, purposefully minimizes the document size needed to disseminate the extensive results from our research.},
doi = {10.2172/903425},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Thu Feb 01 00:00:00 EST 2007},
month = {Thu Feb 01 00:00:00 EST 2007}
}

Technical Report:

Save / Share:
  • In today's world, the need for computing power is becoming more pressing daily. Our need to process, analyze, and store data is quickly exceeding the capabilities of small self-contained serial machines, such as the modern desktop PC. Initially, this gap was filled by the creation of supercomputers: large-scale self-contained parallel machines. However, current markets, as well as the costs to develop and maintain such machines, are quickly making such machines a rarity, used only in highly specialized environments. A third type of machine exists, however. This relatively new type of machine, known as a cluster, is built from common, andmore » often inexpensive, commodity self-contained desktop machines. But how well do these clustered machines work? There have been many attempts to quantify the performance of clustered computers. One approach, Queueing Network Modeling (QNM), appears to be a potentially useful and rarely tried method of modeling such systems. QNM, which has its beginnings in the modeling of traffic patterns, has expanded, and is now used to model everything from CPU and disk services, to computer systems, to service rates in store checkout lines. This history of successful usage, as well as the correspondence of QNM components to commodity clusters, suggests that QNM can be a useful tool for both the cluster designer, interested in the best value for the cost, and the user of existing machines, interested in performance rates and time-to-solution. So, what is QNM? Queueing Network Modeling is an approach to computer system modeling where the computer is represented as a network of queues and evaluated analytically. How does this correspond to clusters? There is a neat one-to-one relationship between the components of a QNM model and a cluster. For example: A cluster is made from a combination of computational nodes and network switches. Both of these fit nicely with the QNM descriptions of service centers (delay, queueing, and load-dependent). Other examples include relationships between different classes of customers in QNM and different types of messages passed on clustered systems, and the obvious relationship between the QNM model queues and message queueing in switches and network cards. Even the parameterization of QNM components lends itself well to cluster modeling. Numbers of service centers (computational nodes and switches) is generally well known for existing systems, and can be estimated for potential systems. Number of customers in the system can be estimated based on application call traces or profiles. Timing rates and service demands, too, can be estimated based on device specifications, or through application tracing or profiling. Typical results reported include throughputs, queue lengths, and response times, all of which are important to determining how well a system is utilized. In this research, QNM is applied to the Keck Cluster as a strong scaling problem. Strong scaling is where the size of the problem to be solved remains constant even as the number of processors allocated to the solution increases. QNM could also be applied in a weak scaling manner, meaning the problem size increases as the number of allocated processors increases, but this application is not investigated here.« less
  • This report summarizes the work performed for the program entitled “High Performance Flexible Reversible Solid Oxide Fuel Cell” under Cooperative Agreement DE-FC36-04GO14351 for the U. S. Department of Energy. The overall objective of this project is to demonstrate a single modular stack that generates electricity from a variety of fuels (hydrogen and other fuels such as biomass, distributed natural gas, etc.) and when operated in the reverse mode, produces hydrogen from steam. This project has evaluated and selected baseline cell materials, developed a set of materials for oxygen and hydrogen electrodes, and optimized electrode microstructures for reversible solid oxide fuelmore » cells (RSOFCs); and demonstrated the feasibility and operation of a RSOFC multi-cell stack. A 10-cell reversible SOFC stack was operated over 1000 hours alternating between fuel cell (with hydrogen and methane as fuel) and steam electrolysis modes. The stack ran very successfully with high power density of 480 mW/cm2 at 0.7V and 80% fuel utilization in fuel cell mode and >6 SLPM hydrogen production in steam electrolysis mode using about 1.1 kW electrical power. The hydrogen generation is equivalent to a specific capability of 2.59 Nm3/m2 with electrical energy demand of 3 kWh/Nm3. The performance stability in electrolysis mode was improved vastly during the program with a degradation rate reduction from 8000 to 200 mohm-cm2/1000 hrs. This was accomplished by increasing the activity and improving microstructure of the oxygen electrode. Both cost estimate and technology assessment were conducted. Besides the flexibility running under both fuel cell mode and electrolysis mode, the reversible SOFC system has the potentials for low cost and high efficient hydrogen production through steam electrolysis. The cost for hydrogen production at large scale was estimated at ~$2.7/kg H2, comparing favorably with other electrolysis techology.« less
  • The Nez Perce Tribe, through funding provided by the Bonneville Power Administration, has implemented a small scale chinook salmon supplementation program on Johnson Creek, a tributary in the South Fork of the Salmon River, Idaho. The Johnson Creek Artificial Propagation Enhancement project was established to enhance the number of threatened Snake River spring/summer chinook salmon (Oncorhynchus tshawytscha) returning to Johnson Creek to spawn through artificial propagation. This was the sixth season of adult chinook broodstock collection in Johnson Creek following collections in 1998, 2000, 2001, 2002, and 2003. Weir installation was completed on June 21, 2004 with the first chinookmore » captured on June 22, 2004 and the last fish captured on September 6, 2004. The weir was removed on September 18, 2004. A total of 338 adult chinook, including jacks, were captured during the season. Of these, 211 were of natural origin, 111 were hatchery origin Johnson Creek supplementation fish, and 16 were adipose fin clipped fish from other hatchery operations and therefore strays into Johnson Creek. Over the course of the run, 57 natural origin Johnson Creek adult chinook were retained for broodstock, transported to the South Fork Salmon River adult holding and spawning facility and held until spawned. The remaining natural origin Johnson Creek fish along with all the Johnson Creek supplementation fish were released upstream of the weir to spawn naturally. Twenty-seven Johnson Creek females were artificially spawned with 25 Johnson Creek males. Four females were diagnosed with high bacterial kidney disease levels resulting in their eggs being culled. The 27 females produced 116,598 green eggs, 16,531 green eggs were culled, with an average eye-up rate of 90.6% resulting in 90,647 eyed eggs. Juvenile fish were reared indoors at the McCall Fish Hatchery until November 2005 and then transferred to the outdoor rearing facilities during the Visual Implant Elastomer tagging operation. These fish continued rearing in the outdoor collection basin until release in March 2006. All of these fish were marked with Coded Wire Tags and Visual Implant Elastomer tags. In addition 12,056 of the smolts released were also tagged with Passive Integrated Transponder tags. Hand counts provided by marking crews were used to amend the number of juvenile salmon released from the original egg count. A total of 90,450 smolts were released directly into Johnson Creek on March 13 through 15, 2006.« less
  • The Building America residential systems research project uses an analysis-based system research approach to (1) Identify research priorities, (2) Identify technology gaps and opportunities, (3) Establish a consistent basis to track research progress, (4) Increase the cost effectiveness of research investments by identifying system solutions that are most likely to succeed as the initial targets for residential system research projects. This report describes the technical approach used by Building America to determine the most cost effective pathways to achieve whole-house energy savings goals. The report provides an overview of design/technology strategies leading to net zero energy buildings as the basismore » for analysis of future residential system performance. The analysis approach is demonstrated by providing an initial comparison of the least-cost options required to achieve 40% energy savings in five climate zones. The preliminary results from this study will be validated against field studies and updated on an annual basis to reflect best available residential system cost/performance data from ongoing Building America research activities.« less
  • This project was designed to document existing habitat conditions and fish populations within the Rattlesnake Creek watershed (White Salmon River subbasin, Washington) before major habitat restoration activities are implemented and prior to the reintroduction of salmon and steelhead above Condit Dam. Returning adult salmon Oncorhynchus spp. and steelhead O. mykiss have not had access to Rattlesnake Creek since 1913. An assessment of resident trout populations should serve as a good surrogate for evaluation of factors that would limit salmon and steelhead production in the watershed. Personnel from United States Geological Survey's Columbia River Research Laboratory (USGS-CRRL) attended to three mainmore » objectives of the Rattlesnake Creek project. The first objective was to characterize stream and riparian habitat conditions. This effort included measures of water quality, water quantity, stream habitat, and riparian conditions. The second objective was to determine the status of fish populations in the Rattlesnake Creek drainage. To accomplish this, we derived estimates of salmonid population abundance, determined fish species composition, assessed distribution and life history attributes, obtained tissue samples for genetic analysis, and assessed fish diseases in the watershed. The third objective was to use the collected habitat and fisheries information to help identify and prioritize areas in need of restoration. As this report covers the fourth year of a five-year study, it is largely restricted to describing our efforts and findings for the first two objectives.« less