IMADA - Department of Mathematics and Computer Science |
We present two recently developed (almost-)optimal algorithms for the (1-ported, full-duplex) multi-processor one-to-all broadcast operation. Both algorithms can be used for efficient implementation of the MPI_Bcast collective operation in message-passing libraries like MPI (the Message Passing Interface). We discuss and contrast the two algorithms with respect also to issues of implementability and maintenance, and present further generalizations. Host: Kim Skak Larsen SDU HOME | IMADA HOME | Previous Page Last modified: Wed Jan 17 09:21:42 CET 2007 Joan Boyar (joan@imada.sdu.dk) |