Basic HTML version of Foils prepared 6 December 96

Foil 30 The Pipe Broadcast Operation

From Full Matrices - December 4, 1995 CPS615 Basic Simulation Track for Computational Science -- Fall Semester 95. by Geoffrey C. Fox, Nancy J. McCracken
1 In the case that the size of the message is large, other implementation optimizations are possible, since it will be necessary for the broadcast message to be broken into a sequence of smaller messages. The broadcast can set up a path (or paths) from the source processor that visits every processor in the group. The message is sent from the source along the path in a pipeline, where each processor receives a block of the message from its predecessor and sends it to its successor. This is a similar concept to wormhole routing, but where each intermediate processor also saves the message in addition to passing it along. (See diagram next page.)
2 The performance of this broadcast is then the time to send the message to the processor on the end of the path plus the overhead of starting and finishing the pipeline.
3 For sufficiently large grain size the pipe broadcast is better than the log broadcast, namely when
in Table To:

© on Tue Oct 7 1997