An Approach to Locality-Conscious Load Balancing and Transparent Memory Hierarchy Management with a Global-Address-Space Parallel Programming Model

Conference · Tue Apr 25 00:00:00 EDT 2006

DOI:https://doi.org/10.1109/IPDPS.2006.1639719· OSTI ID:912986

Krishnamoorthy, Sriram; Catalyurek, Umit; Nieplocha, Jarek; Sadayappan, Ponnuswamy

This paper describes a global-addressspace framework for the convenient specification and effi- cient execution of parallel out-of-core applications operating on block-sparse data. The programming model provides a global view of block-sparse matrices and a mechanism for the expression of parallel tasks that operate on blocksparse data. The tasks are automatically partitioned into phases that operate on memory-resident data, and mapped onto processors to optimize load balance and data locality. Experimental results are presented that demonstrate the utility of the approach.

• Research Organization: Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC05-76RL01830
• OSTI ID: 912986
• Report Number(s): PNNL-SA-48518; KJ0101030; TRN: US200802%%321
• Resource Relation: Conference: 20th International Parallel and Distributed Symposium (IPDPS'06), 25-29 April 2006, , 8
• Country of Publication: United States
• Language: English

Related Subjects

99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE
MEMORY MANAGEMENT
PARALLEL PROCESSING
LOAD MANAGEMENT
OPTIMIZATION
load balancing
global address space
distributed arrays