This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Identifying Performance Bottlenecks in Work-Stealing Computations
December 2009 (vol. 42 no. 12)
pp. 44-50
Nathan R. Tallent, Rice University
John M. Mellor-Crummey, Rice University
Work stealing is an effective load-balancing strategy for multithreading, but when comp- utations based on it underperform, traditional tools can't explain why. To resolve a computation's key performance obstacles, tools must pinpoint and quantify parallel idleness and overhead.

1. N.R. Tallent and J. Mellor-Crummey, "Effective Performance Measurement and Analysis of Multithreaded Applications," Proc. Symp. Principles and Practice of Parallel Programming (SIGPLAN 09), ACM Press, 2009, pp. 229-240.
2. M. Frigo, C.E. Leiserson, and K.H. Randall, "The Implementation of the Cilk-5 Multithreaded Language," Proc. Conf. Programming Language Design and Implementation (SIGPLAN 98), ACM Press, 1998, pp. 212-223.
3. Supercomputing Technologies Group, MIT Laboratory for Computer Science, Cilk Reference Manual; http://supertech.csail.mit.educilk.

Index Terms:
Performance analysis, Call path profiling, Multithreaded programming models, HPCToolkit, Multicore processors
Citation:
Nathan R. Tallent, John M. Mellor-Crummey, "Identifying Performance Bottlenecks in Work-Stealing Computations," Computer, vol. 42, no. 12, pp. 44-50, Dec. 2009, doi:10.1109/MC.2009.396
Usage of this product signifies your acceptance of the Terms of Use.