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

Title: Orbitmpi Documentation

Abstract

Orbitmpi is a parallelized version of Roscoe White's Orbit code. The code has been parallelized using MPI, which makes it portable to many types of machines. The guidelines used for the parallelization were to increase code performance with minimal changes to the code's original structure. This document gives a general description of how the parallel sections of the code run. It discusses the changes made to the original code and comments on the general procedure for future additions to Orbitmpi, as well as describing the effects of a parallelized random number generator on the code's output. Finally, the scaling results from Hecate and from Puffin are presented. Hecate is a 64-processor Origin 2000 machine, with MIPS R12000 processors and 16GB of memory, and Puffin is a PC cluster with 9 dual-processor 450 MHz Pentium III (18 processors max.), with 100Mbits ethernet communication.

Authors:
Publication Date:
Research Org.:
Princeton Plasma Physics Lab., NJ (US)
Sponsoring Org.:
USDOE Office of Energy Research (ER) (US)
OSTI Identifier:
764942
Report Number(s):
PPPL-3489
TRN: US0005040
DOE Contract Number:  
AC02-76CH03073
Resource Type:
Technical Report
Resource Relation:
Other Information: PBD: 5 Oct 2000
Country of Publication:
United States
Language:
English
Subject:
70 PLASMA PHYSICS AND FUSION TECHNOLOGY; 99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; O CODES; DOCUMENTATION; PERFORMANCE; RANDOMNESS; PARALLEL PROCESSING

Citation Formats

Lisa L. Lowe. Orbitmpi Documentation. United States: N. p., 2000. Web. doi:10.2172/764942.
Lisa L. Lowe. Orbitmpi Documentation. United States. doi:10.2172/764942.
Lisa L. Lowe. Thu . "Orbitmpi Documentation". United States. doi:10.2172/764942. https://www.osti.gov/servlets/purl/764942.
@article{osti_764942,
title = {Orbitmpi Documentation},
author = {Lisa L. Lowe},
abstractNote = {Orbitmpi is a parallelized version of Roscoe White's Orbit code. The code has been parallelized using MPI, which makes it portable to many types of machines. The guidelines used for the parallelization were to increase code performance with minimal changes to the code's original structure. This document gives a general description of how the parallel sections of the code run. It discusses the changes made to the original code and comments on the general procedure for future additions to Orbitmpi, as well as describing the effects of a parallelized random number generator on the code's output. Finally, the scaling results from Hecate and from Puffin are presented. Hecate is a 64-processor Origin 2000 machine, with MIPS R12000 processors and 16GB of memory, and Puffin is a PC cluster with 9 dual-processor 450 MHz Pentium III (18 processors max.), with 100Mbits ethernet communication.},
doi = {10.2172/764942},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Thu Oct 05 00:00:00 EDT 2000},
month = {Thu Oct 05 00:00:00 EDT 2000}
}

Technical Report:

Save / Share: