DocumentCode
2537355
Title
Detailed Load Balance Analysis of Large Scale Parallel Applications
Author
Huck, Kevin A. ; Labarta, Jesus
Author_Institution
Centro Nac. de Supercomputacion, Barcelona Supercomput. Center, Barcelona, Spain
fYear
2010
fDate
13-16 Sept. 2010
Firstpage
535
Lastpage
544
Abstract
Balancing the workload in parallel applications is a difficult task, even in conventional cases. Many computing cycles are wasted when the load is not evenly balanced across processing nodes. Global load balance analysis may determine that an application is well balanced, when in fact the application has hidden inefficiencies. In this paper, we consider the load balance of parallel applications which present unique challenges in the analysis process. We have performed trace analysis and simulation to demonstrate the existence of otherwise undiscovered performance issues. We also demonstrate that by collecting dynamic phase profiles, we are able to approximate the analysis results of trace analysis and simulation, and more accurately represent the performance behavior of complex parallel applications than through flat or callpath profiles alone.
Keywords
data mining; parallel processing; processor scheduling; program diagnostics; software performance evaluation; dynamic phase profile collection; global load balance analysis; large scale parallel applications; trace analysis; workload balancing; Analytical models; Benchmark testing; Computational modeling; Instruments; Load modeling; Synchronization; data mining; micro load imbalance; parallel performance analysis; performance tools;
fLanguage
English
Publisher
ieee
Conference_Titel
Parallel Processing (ICPP), 2010 39th International Conference on
Conference_Location
San Diego, CA
ISSN
0190-3918
Print_ISBN
978-1-4244-7913-9
Electronic_ISBN
0190-3918
Type
conf
DOI
10.1109/ICPP.2010.61
Filename
5599254
Link To Document