Index: topol.tex
===================================================================
--- topol.tex (revision 2006)
+++ topol.tex (working copy)
@@ -153,9 +153,10 @@
auxiliary function \mpifunc{MPI\_DIMS\_CREATE} can be used to compute a balanced
distribution of processes among a given number of dimensions.
-%%% Do we really need this rationale - it is over 20 years old now.
-\begin{rationale}Similar functions are contained in
-EXPRESS \cite{express} and PARMACS. \end{rationale}
+%%% WDG: Do we really need this rationale - it is over 20 years old now.
+% htor: agreed, removed as ticket-0 change
+%\begin{rationale}Similar functions are contained in
+%EXPRESS \cite{express} and PARMACS. \end{rationale}
MPI defines functions to query a communicator for topology information.
The function \mpifunc{MPI\_TOPO\_TEST} is used to query for the type of topology
@@ -640,14 +641,12 @@
\const{MPI\_BOTTOM} (not usable for initialization or assignment).
See Section~\ref{subsec:named-constants}. % Subsection 2.5.4.
-%% These three statements need to be re-written - together they make little
-%% sends.
\begin{users}
In the case of an empty weights array argument passed while constructing
a weighted graph, one should not pass \const{NULL} because the value of
\const{MPI\_UNWEIGHTED} may be equal to \const{NULL}. The value of this argument
would then be indistinguishable from \const{MPI\_UNWEIGHTED} to the
-implementation. In this case \const{MPI\_WEIGHTS\_EMPTY} should be used
+implementation. \const{MPI\_WEIGHTS\_EMPTY} should be used
instead.
\end{users}
@@ -995,9 +994,7 @@
Specifically, \mpifunc{MPI\_GRAPH\_NEIGHBORS\_COUNT} and
\mpifunc{MPI\_GRAPH\_NEIGHBORS} will return values based on the original
\mpicode{index} and \mpicode{edges} array passed to \mpifunc{MPI\_GRAPH\_CREATE}
-(assuming that \mpicode{index[-1]} effectively equals zero):
-%% What does effectively mean here? Why not simply say that
-%% for the purposes of this example, assume index[-1] is zero?
+(for the purpose of this example, we assume that \mpicode{index[-1]} is zero):
\begin{itemize}
\item The number of neighbors (\mpiarg{nneighbors}) returned from\flushline % fix for margins
@@ -1199,12 +1196,9 @@
\subsection{Cartesian Shift Coordinates}
\label{subsec:topol-shift}
-%%
-%% I don't believe ``likely'' is appropriate here.
-%% Perhaps it is better to say ``an MPI_SENDRECV operation may be used ...
If the process topology is a Cartesian structure, an
\mpifunc{MPI\_SENDRECV}
-operation is likely to be used along a coordinate direction to perform a shift
+operation may be used along a coordinate direction to perform a shift
of data. As input, \mpifunc{MPI\_SENDRECV} takes the rank of a source process
for the receive, and the rank of a destination process for the send. If the
function \mpifunc{MPI\_CART\_SHIFT} is called for a Cartesian process group, it