Find this at http://www.npac.syr.edu/users/gcf/cps615nic95/

Parallel Computation Illustrated with Adaptive Integration

This talk assumes the material found in the CPS615 lecture notes, Introduction to Numerical Integration.

Going on from the static data decomposition for integration on page 27 of those notes, we examine further parallel algorithms for computing adaptive Romberg (recursive Simpson) integration.

The data decomposition and remapping techniques discussed are generally applicable to parallel adaptive algorithms for other applications.

Table of Contents for Parallel Computation Illustrated with Adaptive Integration

001 Computing Techniques for Adaptive Integration
002 Computing Techniques for Adaptive Integration
003 Sequential Romberg Integration
004 Program for Romberg Integration
    Parallel Romberg Integration
005 Parallel Romberg Integration, Continued
006 Adaptive Parallel Romberg - CommAll Version
    At each iteration, data is decomposed again by all processors.
007 Adaptive Parallel Romberg - program initialization
    Adaptive Parallel Romberg - compute integral iteration and 
    communicate values to all processors
    Adaptive Parallel Romberg - recompute decomposition
    Adaptive Parallel Romberg - Neighbors Version
    At each iteration, processors average load with neighbors
008 More discussion of Neighbors version