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

Title: Aces4: A Platform for Computational Chemistry Calculations with Extremely Large Block-Sparse Arrays

Authors:
 [1];  [2];  [2]; ORCiD logo [3];  [2];  [2];  [2]
  1. ENSCO, Inc.
  2. University of Florida, Gainesville
  3. ORNL
Publication Date:
Research Org.:
Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR) (SC-21)
OSTI Identifier:
1376383
DOE Contract Number:
AC05-00OR22725
Resource Type:
Conference
Resource Relation:
Conference: IEEE International Parallel and Distributed Processing Symposium (IPDPS'17) - Orlando, Florida, United States of America - 5/29/2017 8:00:00 AM-
Country of Publication:
United States
Language:
English

Citation Formats

Byrd, Jason, Jindal, Nakul, Lotrich, Victor, Liakh, Dmitry I., Perera, Ajith, Bartlett, Rodney, and Sanders, Beverly. Aces4: A Platform for Computational Chemistry Calculations with Extremely Large Block-Sparse Arrays. United States: N. p., 2017. Web.
Byrd, Jason, Jindal, Nakul, Lotrich, Victor, Liakh, Dmitry I., Perera, Ajith, Bartlett, Rodney, & Sanders, Beverly. Aces4: A Platform for Computational Chemistry Calculations with Extremely Large Block-Sparse Arrays. United States.
Byrd, Jason, Jindal, Nakul, Lotrich, Victor, Liakh, Dmitry I., Perera, Ajith, Bartlett, Rodney, and Sanders, Beverly. Mon . "Aces4: A Platform for Computational Chemistry Calculations with Extremely Large Block-Sparse Arrays". United States. doi:. https://www.osti.gov/servlets/purl/1376383.
@article{osti_1376383,
title = {Aces4: A Platform for Computational Chemistry Calculations with Extremely Large Block-Sparse Arrays},
author = {Byrd, Jason and Jindal, Nakul and Lotrich, Victor and Liakh, Dmitry I. and Perera, Ajith and Bartlett, Rodney and Sanders, Beverly},
abstractNote = {},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Mon May 01 00:00:00 EDT 2017},
month = {Mon May 01 00:00:00 EDT 2017}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • Development of complex scientific applications for high-end systems is a challenging task. Addressing complexity of software and algorithms involved is becoming increasingly difficult and requires appropriate software engineering approaches to address interoperability, maintenance, and software composition challenges. At the same time, the requirements for performance and scalability to thousand processor configurations magnifies the level of difficulties facing the scientific programmer due to the variable levels of parallelism available in different algorithms or functional modules of the application. The current paper demonstrates how the Common Component Architecture (CCA) can be used in context of computational chemistry to express and manage multi-levelmore » parallelism through the use of processor groups. Our target application is NWChem, a large (2.5millions lines of code) suite of computational chemistry algorithms that was developed based on multiple languages (Fortran, C, C++, Python) and programming models (MPI, Global Arrays) approach. The experimental results for numerical Hessian calculation show that multilevel parallelism expressed and managed through the CCA component model can be very effective for improving performance and scalability of NWChem. For example, numerical Hessian calculation using three levels of parallelism outperformed the original version of the NWChem code based on one level parallelism by a factor of 90% when running on 256 processors.« less
  • This paper explores the use of a sub-block decomposition strategy for parallel sparse Cholesky factorization, in which the sparse matrix is decomposed into rectangular blocks. Such a strategy has enormous theoretical scalability advantages over more traditional column-oriented and panel-oriented decompositions. However, little progress has been made in producing a practical sub-block method. This paper proposes and evaluates an approach that is simple to implement, provides slightly higher performance than column (and panel) methods on small parallel machines, and has the potential to provide much higher performance on large parallel machines.
  • Abstract not provided.
  • Abstract not provided.