Scalable I/O for Irregular Loosely Synchronous Problems.

Abstract

To achieve good I/O performance on irregular, loosely synchronous problems it is necessary to work both at the application and the system support level. The first section describes our effort at developing an efficient out-of-core parallel sparse cholesky solver as an example of an irregular, loosely synchronous application and the second section describes the Jovian parallel I/O library which provides support for collective I/O operations.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1997
Accession Number
ADA327943

Entities

People

  • Anurag Acharya

Organizations

  • University of Maryland

Tags

DTIC Thesaurus Topics

  • Agreements
  • Bandwidth
  • Buildings And Structures
  • Computer Science
  • Computers
  • Department Of Defense
  • Load Monitoring
  • Maryland
  • Military Research
  • Research Facilities
  • Sequences
  • Steady State
  • Students
  • Technical Information Centers
  • Universities
  • Virginia

Fields of Study

  • Computer science

Readers

  • Astronomy and Astrophysics.
  • Computational Fluid Dynamics (CFD)
  • Linear Algebra