Information Bridge

Bibliographic Citation 
Full Text PDF 0 K   View Full Text or Access Individual Pages  -   search, view and/or download individual pages
DOI 10.2172/793980
Title Improving Cache Utilization of Linear Relaxation Methods: Theory and Practice
Creator/Author Bassetti, F. ; Dawis, K. ; Maratha, M.
Publication Date1999 Dec 01
OSTI IdentifierOSTI ID: 793980
Report Number(s)UCRL-ID-136661
DOE Contract NumberW-7405-Eng-48
DOI10.2172/793980
Other Number(s)TRN: US200302%%694
Resource TypeTechnical Report
Resource RelationOther Information: PBD: 1 Dec 1999
CoverageTopical
Research OrgLawrence Livermore National Lab., CA (US)
Sponsoring OrgUSDOE Office of Defense Programs (DP) (US)
Subject99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; PERFORMANCE; RELAXATION; LAWRENCE LIVERMORE NATIONAL LABORATORY
Description/Abstract Application codes reliably achieve performance far less than the advertised capabilities of existing architectures, and this problem is worsening with increasingly-parallel machines. For large-scale numerical applications, stencil operations often impose the greater part of the computational cost, and the primary sources of inefficiency are the costs of message passing and poor cache utilization. This paper proposes and demonstrates optimizations for stencil and stencil-like computations for both serial and parallel environments that ameliorate these sources of inefficiency. Additionally, they argue that when stencil-like computations are encoded at a high level using object-oriented parallel array class libraries, these optimizations, which are beyond the capability of compilers, may be automated.
Country of PublicationUnited States
LanguageEnglish
FormatSize: 908 Kilobytes pages
System Entry Date2007 Nov 30
Document Discussions
 (for display)
 (Email address will NOT be displayed.)

   (All fields required. Document Discussions not displayed until approved.)

Top