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

Title: PEBBL: An object-oriented framework for scalable parallel branch and bound.

Abstract

Abstract not provided.

Authors:
; ;
Publication Date:
Research Org.:
Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
Sponsoring Org.:
USDOE National Nuclear Security Administration (NNSA)
OSTI Identifier:
1095816
Report Number(s):
SAND2013-8562J
476646
DOE Contract Number:
AC04-94AL85000
Resource Type:
Journal Article
Resource Relation:
Journal Name: Mathematical Programming Computation; Related Information: Proposed for publication in Mathematical Programming Computation.
Country of Publication:
United States
Language:
English

Citation Formats

Hart, William E., Phillips, Cynthia A., and Eckstein, Jonathan. PEBBL: An object-oriented framework for scalable parallel branch and bound.. United States: N. p., 2013. Web.
Hart, William E., Phillips, Cynthia A., & Eckstein, Jonathan. PEBBL: An object-oriented framework for scalable parallel branch and bound.. United States.
Hart, William E., Phillips, Cynthia A., and Eckstein, Jonathan. Tue . "PEBBL: An object-oriented framework for scalable parallel branch and bound.". United States. doi:.
@article{osti_1095816,
title = {PEBBL: An object-oriented framework for scalable parallel branch and bound.},
author = {Hart, William E. and Phillips, Cynthia A. and Eckstein, Jonathan},
abstractNote = {Abstract not provided.},
doi = {},
journal = {Mathematical Programming Computation},
number = ,
volume = ,
place = {United States},
year = {Tue Oct 01 00:00:00 EDT 2013},
month = {Tue Oct 01 00:00:00 EDT 2013}
}
  • This report describes the design of PICO, a C++ framework for implementing general parallel branch-and-bound algorithms. The PICO framework provides a mechanism for the efficient implementation of a wide range of branch-and-bound methods on an equally wide range of parallel computing platforms. We first discuss the basic architecture of PICO, including the application class hierarchy and the package's serial and parallel layers. We next describe the design of the serial layer, and its central notion of manipulating subproblem states. Then, we discuss the design of the parallel layer, which includes flexible processor clustering and communication rates, various load balancing mechanisms,more » and a non-preemptive task scheduler running on each processor. We describe the application of the package to a branch-and-bound method for mixed integer programming, along with computational results on the ASCI Red massively parallel computer. Finally we describe the application of the branch-and-bound mixed-integer programming code to a resource constrained project scheduling problem for Pantex.« less
  • Abstract not provided.
  • The phase field approach is a powerful and popular method for modeling microstructure evolution. In this work, advanced numerical tools are used to create a phase field framework that facilitates rapid model development. This framework, called MARMOT, is based on Idaho National Laboratory's finite element Multiphysics Object-Oriented Simulation Environment. In MARMOT, the system of phase field partial differential equations (PDEs) are solved simultaneously with PDEs describing additional physics, such as solid mechanics and heat conduction, using the Jacobian-Free Newton Krylov Method. An object-oriented architecture is created by taking advantage of commonalities in phase fields models to facilitate development of newmore » models with very little written code. In addition, MARMOT provides access to mesh and time step adaptivity, reducing the cost for performing simulations with large disparities in both spatial and temporal scales. In this work, phase separation simulations are used to show the numerical performance of MARMOT. Deformation-induced grain growth and void growth simulations are included to demonstrate the muliphysics capability.« less
  • The Parallel Object-Oriented Methods and Applications (POOMA) FrameWork provides a rich variety of objects to enable rapid prototyping and development of scientific applications on high-performance computing platforms. Applications built with components from the POOMA FrameWork can move between serial, distributed, and parallel architectures with no change to the high-level source code. Thus, a developer is able to develop applications on a serial workstation and, with no change to the source code, move the application to a parallel supercomputer for production. Simulations that are developed under this paradigm are easily moved to new architectures because the details of message-passing algorithms aremore » encapsulated from the application developer. Thus, the investment made in refining a simulation with FrameWork components enables the scientist`s application to run on the next generation of high-performance parallel architectures with little or no effort.« less
  • In this paper, we derive an object-oriented parallel algorithm for three-dimensional isopycnal flow simulations. The matrix formulation is central to the algorithm. It enables us to apply an efficient preconditioned conjugate gradient linear solver for the global system of equations, and leads naturally to an object-oriented data structure design and parallel implementation. We discuss as well, in less detail, a similar algorithm based on the reduced system, suitable also for parallel computation. Favorable performances are observed on test problems.