Uintah parallelism infrastructure: a performance evaluation on the SGI origin 2000

Update item information
Publication Type Manuscript
School or College Scientific Computing and Imaging Institute
Program Center for the Simulation of Accidental Fires and Explosions (C-SAFE)
Creator De St Germain, John Davison; Parker, Steven G.; Johnson, Christopher R.
Other Author McCorquodale, John
Title Uintah parallelism infrastructure: a performance evaluation on the SGI origin 2000
Date 2001
Description Uintah is a component-based visual problem solving environment (PSE) designed to specifically address the unique problems inherent in running massively parallel scientific computations on terascale computing platforms. In particular, development of the Uintah system is part of the C-SAFE [2] effort to study the interactions between hydrocarbon fires, structures and high-energy materials (explosives and propellants). In this paper we describe methods for generating meaningful performance measurements for the Uintah PSE runing on the SGI Origin 2000 multiprocessor architecture (these methods are applicable to many other applications.) These techniques include utilizing the non-intrusive performance counters built into the R10k and R12k processors, controlling process placement, controlling memory layout, and utilization of a task graph approach to specifying and solving the problem.
Type Text
Publisher Society for Modeling and Simulation International (SCS)
First Page 1
Last Page 6
Subject Uintah; Problem solving environment; Performance analysis; Parallelism; Origin 2000; C-SAFE
Subject LCSH Parallel processing (Electronic computers); Computer simulation
Language eng
Bibliographic Citation McCorquodale, J., De St. Germain, J. D., Parker, S. G., & Johnson, C. R. (2001). Uintah parallelism infrastructure: a performance evaluation on the SGI origin 2000. High Performance Computing 2001, Seattle, Washington, March.
Rights Management (c)Society for Modeling and Simulation International (SCS) Information is provided courtesy of the The Society for Modeling & Simulation International.
Format Medium application/pdf
Format Extent 176,369 bytes
Identifier ir-main,7195
ARK ark:/87278/s6b85skd
Setname ir_uspace
Date Created 2012-06-13
Date Modified 2021-05-06
ID 705316
Reference URL https://collections.lib.utah.edu/ark:/87278/s6b85skd
Back to Search Results