Add Block-Strided RMA Operations

content/shmem_ibget.tex

@@ -0,0 +1,59 @@
+\apisummary{
+    Copies blocks of strided data from a specified \ac{PE}.
+}
+
+\begin{apidefinition}
+
+\begin{C11synopsis}
+void @\FuncDecl{shmem\_ibget}@(TYPE *dest, const TYPE *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+void @\FuncDecl{shmem\_ibget}@(shmem_ctx_t ctx, TYPE *dest, const TYPE *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+\end{C11synopsis}
+where \TYPE{} is one of the standard \ac{RMA} types specified by Table \ref{stdrmatypes}.
+
+\begin{Csynopsis}
+void @\FuncDecl{shmem\_\FuncParam{TYPENAME}\_ibget}@(TYPE *dest, const TYPE *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+void @\FuncDecl{shmem\_ctx\_\FuncParam{TYPENAME}\_ibget}@(shmem_ctx_t ctx, TYPE *dest, const TYPE *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+\end{Csynopsis}
+where \TYPE{} is one of the standard \ac{RMA} types and has a corresponding \TYPENAME{} specified by Table \ref{stdrmatypes}.
+
+\begin{CsynopsisCol}
+void @\FuncDecl{shmem\_ibget\FuncParam{SIZE}}@(void *dest, const void *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+void @\FuncDecl{shmem\_ctx\_ibget\FuncParam{SIZE}}@(shmem_ctx_t ctx, void *dest, const void *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+\end{CsynopsisCol}
+where \SIZE{} is one of \CONST{8, 16, 32, 64, 128}.
+
+\begin{apiarguments}
+    \apiargument{IN}{ctx}{A context handle specifying the context on which to perform the operation.
+        When this argument is not provided, the operation is performed on
+        the default context.}
+    \apiargument{OUT}{dest}{Local address of the array to be updated.
+        The type of \dest{} should match that implied in the SYNOPSIS section.}
+    \apiargument{IN}{source}{Symmetric address of the source array data object.
+        The type of \source{} should match that implied in the SYNOPSIS section.}
+    \apiargument{IN}{dst}{The stride between consecutive elements of the \dest{}
+        array. The stride must be greater than or equal to \CONST{1} and is
+        scaled by the element size of the \dest{} array. A value of \CONST{1}
+        indicates contiguous data.}
+    \apiargument{IN}{sst}{The stride between consecutive elements of the \source{}
+        array. The stride must be greater than or equal to \CONST{1} and is
+        scaled by the element size of the \source{} array. A value of \CONST{1}
+        indicates contiguous data.}
+    \apiargument{IN}{bsize}{Number of elements per block in the \dest{} and \source{}
+        arrays.}
+    \apiargument{IN}{nblocks}{Number of blocks in the \dest{} and \source{}
+        arrays.}
+    \apiargument{IN}{pe}{\ac{PE} number of the remote \ac{PE}.}
+\end{apiarguments}
+
+\apidescription{
+    The \FUNC{shmem\_ibget} routines provide a method for copying strided data blocks from
+    a symmetric array from a specified remote \ac{PE} to strided locations on a
+    local array.  The routines return when the data has been copied into the local
+    \VAR{dest} array.
+}
+
+\apireturnvalues{
+    None.
+}
+
+\end{apidefinition}

content/shmem_ibput.tex

@@ -0,0 +1,62 @@
+\apisummary{
+    Copies strided data blocks to a specified \ac{PE}.
+}
+
+\begin{apidefinition}
+
+\begin{C11synopsis}
+void @\FuncDecl{shmem\_ibput}@(TYPE *dest, const TYPE *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+void @\FuncDecl{shmem\_ibput}@(shmem_ctx_t ctx, TYPE *dest, const TYPE *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+\end{C11synopsis}
+where \TYPE{} is one of the standard \ac{RMA} types specified by Table \ref{stdrmatypes}.
+
+\begin{Csynopsis}
+void @\FuncDecl{shmem\_\FuncParam{TYPENAME}\_ibput}@(TYPE *dest, const TYPE *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+void @\FuncDecl{shmem\_ctx\_\FuncParam{TYPENAME}\_ibput}@(shmem_ctx_t ctx, TYPE *dest, const TYPE *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+\end{Csynopsis}
+where \TYPE{} is one of the standard \ac{RMA} types and has a corresponding \TYPENAME{} specified by Table \ref{stdrmatypes}.
+
+\begin{CsynopsisCol}
+void @\FuncDecl{shmem\_ibput\FuncParam{SIZE}}@(void *dest, const void *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+void @\FuncDecl{shmem\_ctx\_ibput\FuncParam{SIZE}}@(shmem_ctx_t ctx, void *dest, const void *source, ptrdiff_t dst, ptrdiff_t sst, size_t bsize, size_t nblocks, int pe);
+\end{CsynopsisCol}
+where \SIZE{} is one of \CONST{8, 16, 32, 64, 128}.
+
+\begin{apiarguments}
+    \apiargument{IN}{ctx}{A context handle specifying the context on which to perform the operation.
+        When this argument is not provided, the operation is performed on
+        the default context.}
+    \apiargument{OUT}{dest}{Symmetric address of the destination array data object.
+        The type of \dest{} should match that implied in the SYNOPSIS section.}
+    \apiargument{IN}{source}{Local address of the array containing the data to be copied.
+        The type of \source{} should match that implied in the SYNOPSIS section.}
+    \apiargument{IN}{dst}{The stride between consecutive elements of the \dest{}
+        array. The stride must be greater than or equal to \CONST{1} and is
+        scaled by the element size of the \dest{} array. A value of \CONST{1}
+        indicates contiguous data.}
+    \apiargument{IN}{sst}{The stride between consecutive elements of the \source{}
+        array. The stride must be greater than or equal to \CONST{1} and is
+        scaled by the element size of the \source{} array. A value of \CONST{1}
+        indicates contiguous data.}
+    \apiargument{IN}{bsize}{Number of elements per block in the \dest{} and \source{}
+        arrays.}
+    \apiargument{IN}{nblocks}{Number of blocks in the \dest{} and \source{}
+        arrays.}
+    \apiargument{IN}{pe}{\ac{PE} number of the remote \ac{PE}.}
+\end{apiarguments}
+
+
+\apidescription{
+    The \FUNC{shmem\_ibput} routines provide a method  for  copying strided data
+    elements (specified by \VAR{sst}) of an array from a \source{} array on the
+    local \ac{PE} to locations specified by stride \VAR{dst} on a \dest{} array
+    on specified remote \ac{PE}. The routines return when the data has
+    been copied out of the \VAR{source} array on the local \ac{PE} but not
+    necessarily before the data has been delivered to the remote data object.
+}
+
+\apireturnvalues{
+    None.
+}
+
+\end{apidefinition}

content/shmem_iget.tex

@@ -31,18 +31,20 @@
     \apiargument{IN}{source}{Symmetric address of the source array data object.
         The type of \source{} should match that implied in the SYNOPSIS section.}
     \apiargument{IN}{dst}{The stride between consecutive elements of the \dest{}
-        array.  The stride is scaled by the element size of the \dest{} array.
-        A  value of \CONST{1} indicates contiguous data.}
-    \apiargument{IN}{sst}{The stride between consecutive elements of the
-        \source{} array.  The stride is scaled by the element size of the \source{}
-        array.  A  value of \CONST{1} indicates contiguous data.}
+        array. The stride must be greater than or equal to \CONST{1} and is
+        scaled by the element size of the \dest{} array. A value of \CONST{1}
+        indicates contiguous data.}
+    \apiargument{IN}{sst}{The stride between consecutive elements of the \source{}
+        array. The stride must be greater than or equal to \CONST{1} and is
+        scaled by the element size of the \source{} array. A value of \CONST{1}
+        indicates contiguous data.}
     \apiargument{IN}{nelems}{Number of elements in the \dest{} and \source{}
         arrays.}
     \apiargument{IN}{pe}{\ac{PE} number of the remote \ac{PE}.}
 \end{apiarguments}
 
 \apidescription{
-    The \FUNC{iget} routines provide a method for copying strided data elements from
+    The \FUNC{shmem\_iget} routines provide a method for copying strided data elements from
     a symmetric array from a specified remote \ac{PE} to strided locations on a
     local array.  The routines return when the data has been copied into the local
     \VAR{dest} array.

content/shmem_iput.tex

@@ -31,22 +31,23 @@
     \apiargument{IN}{source}{Local address of the array containing the data to be copied.
         The type of \source{} should match that implied in the SYNOPSIS section.}
     \apiargument{IN}{dst}{The stride between consecutive elements of the \dest{}
-        array.  The stride is scaled by the element size of the \dest{} array.  A
-        value of \CONST{1} indicates contiguous data.}
-    \apiargument{IN}{sst}{The  stride between consecutive elements of the
-        \source{} array.  The stride is scaled by the element size of the \source{}
-        array.  A  value of \CONST{1} indicates contiguous data.}
+        array. The stride must be greater than or equal to \CONST{1} and is
+        scaled by the element size of the \dest{} array. A value of \CONST{1}
+        indicates contiguous data.}
+    \apiargument{IN}{sst}{The stride between consecutive elements of the \source{}
+        array. The stride must be greater than or equal to \CONST{1} and is
+        scaled by the element size of the \source{} array. A value of \CONST{1}
+        indicates contiguous data.}
     \apiargument{IN}{nelems}{Number of elements in the \dest{} and \source{} arrays.}
     \apiargument{IN}{pe}{\ac{PE} number of the remote \ac{PE}.}
 \end{apiarguments}
 
 
 \apidescription{
-    The \FUNC{iput} routines provide a method  for  copying strided data
+    The \FUNC{shmem\_iput} routines provide a method  for  copying strided data
     elements (specified by \VAR{sst}) of an array from a \source{} array on the
     local \ac{PE} to locations specified by stride \VAR{dst} on a \dest{} array
-    on specified remote \ac{PE}. Both strides, \VAR{dst} and \VAR{sst}, must be
-    greater than or equal to \CONST{1}. The routines return when the data has
+    on specified remote \ac{PE}. The routines return when the data has
     been copied out of the \VAR{source} array on the local \ac{PE} but not
     necessarily before the data has been delivered to the remote data object.
 }
@@ -55,11 +56,6 @@
     None.
 }
 
-\apinotes{
-    See Section \ref{subsec:memory_model} for a definition of the term
-    remotely accessible.
-}
-
 \begin{apiexamples}
 
 \apicexample

main_spec.tex

@@ -171,6 +171,9 @@ \subsubsubsection{\textbf{SHMEM\_P}}\label{subsec:shmem_p}
 \subsubsubsection{\textbf{SHMEM\_IPUT}}\label{subsec:shmem_iput}
 \input{content/shmem_iput.tex}
 
+\subsubsubsection{\textbf{SHMEM\_IBPUT}}\label{subsec:shmem_ibput}
+\input{content/shmem_ibput.tex}
+
 \subsubsubsection{\textbf{SHMEM\_GET}}\label{subsec:shmem_get}
 \input{content/shmem_get.tex}
 
@@ -180,6 +183,9 @@ \subsubsubsection{\textbf{SHMEM\_G}}\label{subsec:shmem_g}
 \subsubsubsection{\textbf{SHMEM\_IGET}}\label{subsec:shmem_iget}
 \input{content/shmem_iget.tex}
 
+\subsubsubsection{\textbf{SHMEM\_IBGET}}\label{subsec:shmem_ibget}
+\input{content/shmem_ibget.tex}
+
 \subsubsection{Nonblocking Remote Memory Access Routines}\label{subsec:rma_nbi}
 
 \subsubsubsection{\textbf{SHMEM\_PUT\_NBI}}\label{subsec:shmem_put_nbi}