*CMZ : 11/11/94 14.40.52 by John Apostolakis CERN GP-MIMD 2 *-- Author : John Apostolakis CERN GP-MIMD 2 13/07/94 C------------------------------------------------------------------------ C gpdynamic C Function: informs parallel Geant that events will be C dynamically distributed between nodes, ie C each "worker" node will request (the data of) an event C from a "master" or "leader". C C called by: , (if CERNLIB_PARA switch is used) C------------------------------------------------------------------------ #if defined(CERNLIB_PARA) subroutine gpdynamic implicit none #include "geant321/multiprox.inc" c c This routine overrides the default (set in gpinit) c of a static decomposition of events , with which c each processor would get 1/(number of processors) of the share c of events. c It should be used if an input file exists or the time per event c has a large standard deviation, the events can be dynamically assigned. c You will need to look at the routines gukine and muxread c in gexam3 for how to distribute events. c c The following initialises the "dynamic" configuration, c in which there is no "master" or "leader". c c -- Id of task expected to be a "master" npleader = 0 c -- Id of task expected to be 1st worker c (representative for printout ...) nfirstworker= 1 return end #else SUBROUTINE GPDYNAMIC_DUMMY END #endif