GETNODES (2)
Get array of LAM node identifiers.
C SYNOPSIS
#include <net.h>
int getnodes (int *nodes, int nnodes, int nodetype, int typemask);
int getall (int *nodes, int nnodes);
int getcomps (int *nodes, int nnodes);
int getjones (int *nodes, int nnodes);
int getotbs (int *nodes, int nnodes);
FORTRAN SYNOPSIS
subroutine GNODES (nodes, nnodes, nodetype, typemask, ierror)
subroutine GALL (nodes, nnodes, ierror)
subroutine GCOMPS (nodes, nnodes, ierror)
subroutine GJONES (nodes, nnodes, ierror)
subroutine GOTBS (nodes, nnodes, ierror)
integer nodes(*), nnodes, nodetype, typemask, ierror
DESCRIPTION
getnodes()
fills the array,
nodes ,
with no more than
nnodes
node identifiers whose types match
nodetype
at the bit positions specified by
typemask .
Given sufficient space and a sufficiently large
nnodes ,
all node identifiers having a node type field that satisfies
the relation imposed by the
nodetype
and
typemask
arguments will be returned.
The remaining functions are wrappers for
getnodes()
that match commonly used node types.
getotbs()
getall()
getcomps()
nodes in main computing group (see getntype(2))
getjones()
neighbour nodes of the caller
The node type is a combination of flags, defined in <net.h>,
describing the capabilities of the node.
Usage of the
nodetype
and
typemask
arguments in
getnodes()
is identical to getntype(2).
Indeed, getntype(2) is usually called before
getnodes()
to count the matching nodes and thus ascertain the space requirement
for the node array.
No indication of the number of returned node identifiers is given by
getnodes() .
SEE ALSO
|