32 |
validity of the cube topology depends on the \file{SIZE.h} file as |
validity of the cube topology depends on the \file{SIZE.h} file as |
33 |
detailed below. The default files provided in the release configure a |
detailed below. The default files provided in the release configure a |
34 |
cubed sphere topology of six tiles, one per subdomain, each with |
cubed sphere topology of six tiles, one per subdomain, each with |
35 |
32$\times$32 grid points, all running on a single processor. Both |
32$\times$32 grid points, with all tiles running on a single processor. Both |
36 |
files are generated by Matlab scripts in |
files are generated by Matlab scripts in |
37 |
\file{utils/exch2/matlab-topology-generator}; see Section |
\file{utils/exch2/matlab-topology-generator}; see Section |
38 |
\ref{sec:topogen} \sectiontitle{Generating Topology Files for exch2} |
\ref{sec:topogen} \sectiontitle{Generating Topology Files for exch2} |
54 |
|
|
55 |
$\bullet$ An example of \file{W2\_EXCH2\_TOPOLOGY.h} and |
$\bullet$ An example of \file{W2\_EXCH2\_TOPOLOGY.h} and |
56 |
\file{w2\_e2setup.F} must reside in a directory containing files |
\file{w2\_e2setup.F} must reside in a directory containing files |
57 |
symbolically linked when \file{genmake2} runs. The safest place to |
symbolically linked by the \file{genmake2} script. The safest place to |
58 |
put these is the directory indicated in the \code{-mods=DIR} command |
put these is the directory indicated in the \code{-mods=DIR} command |
59 |
line modifier (typically \file{../code}), or the build directory. |
line modifier (typically \file{../code}), or the build directory. |
60 |
The default versions of these files reside in \file{pkg/exch2} and |
The default versions of these files reside in \file{pkg/exch2} and |
66 |
\file{tile00$n$.mitgrid} where $n$=\code{(1:6)} (one per subdomain), |
\file{tile00$n$.mitgrid} where $n$=\code{(1:6)} (one per subdomain), |
67 |
must be in the working directory when the MITgcm executable is run. |
must be in the working directory when the MITgcm executable is run. |
68 |
These files are provided in the example experiments for cubed sphere |
These files are provided in the example experiments for cubed sphere |
69 |
configurations with 32$\times$32 cube sides and are non-trivial to |
configurations with 32$\times$32 cube sides |
70 |
generate -- please contact MITgcm support if you want to generate |
-- please contact MITgcm support if you want to generate |
71 |
files for other configurations. \\ |
files for other configurations. \\ |
72 |
|
|
73 |
$\bullet$ As always when compiling MITgcm, the file \file{SIZE.h} must |
$\bullet$ As always when compiling MITgcm, the file \file{SIZE.h} must |
77 |
\file{W2\_EXCH2\_TOPOLOGY.h} and \file{w2\_e2setup.F}. Domain |
\file{W2\_EXCH2\_TOPOLOGY.h} and \file{w2\_e2setup.F}. Domain |
78 |
decomposition issues particular to exch2 are addressed in Section |
decomposition issues particular to exch2 are addressed in Section |
79 |
\ref{sec:topogen} \sectiontitle{Generating Topology Files for exch2} |
\ref{sec:topogen} \sectiontitle{Generating Topology Files for exch2} |
80 |
and \ref{sec:exch2mpi} \sectiontitle{exch2, SIZE.h, and MPI}; a more |
and \ref{sec:exch2mpi} \sectiontitle{exch2, SIZE.h, and Multiprocessing}; a more |
81 |
general background on the subject relevant to MITgcm is presented in |
general background on the subject relevant to MITgcm is presented in |
82 |
Section \ref{sect:specifying_a_decomposition} |
Section \ref{sect:specifying_a_decomposition} |
83 |
\sectiontitle{Specifying a decomposition}.\\ |
\sectiontitle{Specifying a decomposition}.\\ |
106 |
from the Matlab prompt (there are no parameters to pass) generates |
from the Matlab prompt (there are no parameters to pass) generates |
107 |
exch2 topology files \file{W2\_EXCH2\_TOPOLOGY.h} and |
exch2 topology files \file{W2\_EXCH2\_TOPOLOGY.h} and |
108 |
\file{w2\_e2setup.F} in the working directory and displays a figure of |
\file{w2\_e2setup.F} in the working directory and displays a figure of |
109 |
the topology via Matlab. The other m-files in the directory are |
the topology via Matlab -- figures \ref{fig:6tile}, \ref{fig:12tile}, |
110 |
subroutines of \file{driver.m} and should not be run ``bare'' except |
and \ref{fig:24tile} are examples of the generated diagrams. The other |
111 |
|
m-files in the directory are |
112 |
|
subroutines called from \file{driver.m} and should not be run ``bare'' except |
113 |
for development purposes. \\ |
for development purposes. \\ |
114 |
|
|
115 |
The parameters that determine the dimensions and topology of the |
The parameters that determine the dimensions and topology of the |
116 |
generated configuration are \code{nr}, \code{nb}, \code{ng}, |
generated configuration are \code{nr}, \code{nb}, \code{ng}, |
117 |
\code{tnx} and \code{tny}, and all are assigned early in the script. \\ |
\code{tnx} and \code{tny}, and all are assigned early in the script. \\ |
118 |
|
|
119 |
The first three determine the size of the subdomains and |
The first three determine the height and width of the subdomains and |
120 |
hence the size of the overall domain. Each one determines the number |
hence the size of the overall domain. Each one determines the number |
121 |
of grid points, and therefore the resolution, along the subdomain |
of grid points, and therefore the resolution, along the subdomain |
122 |
sides in a ``great circle'' around an axis of the cube. At the time |
sides in a ``great circle'' around each the three spatial axes of the cube. At the time |
123 |
of this writing MITgcm requires these three parameters to be equal, |
of this writing MITgcm requires these three parameters to be equal, |
124 |
but they provide for future releases to accomodate different |
but they provide for future releases to accomodate different |
125 |
resolutions around the axes to allow (for example) greater resolution |
resolutions around the axes to allow subdomains with differing resolutions.\\ |
|
around the equator.\\ |
|
126 |
|
|
127 |
The parameters \code{tnx} and \code{tny} determine the dimensions of |
The parameters \code{tnx} and \code{tny} determine the width and height of |
128 |
the tiles into which the subdomains are decomposed, and must evenly |
the tiles into which the subdomains are decomposed, and must evenly |
129 |
divide the integer assigned to \code{nr}, \code{nb} and \code{ng}. |
divide the integer assigned to \code{nr}, \code{nb} and \code{ng}. |
130 |
The result is a rectangular tiling of the subdomain. Figure |
The result is a rectangular tiling of the subdomain. Figure |
140 |
|
|
141 |
\caption{Plot of a cubed sphere topology with a 32$\times$192 domain |
\caption{Plot of a cubed sphere topology with a 32$\times$192 domain |
142 |
divided into six 32$\times$32 subdomains, each of which is divided |
divided into six 32$\times$32 subdomains, each of which is divided |
143 |
into four tiles (\code{tnx=16, tny=16}) for a total of twenty-four |
into four tiles of width \code{tnx=16} and height \code{tny=16} for a |
144 |
tiles. } \label{fig:24tile} |
total of twenty-four tiles. The colored borders of the subdomains |
145 |
|
represent the parameters \code{nr} (red), \code{nb} (blue), and |
146 |
|
\code{ng} (green). } \label{fig:24tile} |
147 |
\end{figure} |
\end{figure} |
148 |
|
|
149 |
\begin{figure} |
\begin{figure} |
182 |
|
|
183 |
|
|
184 |
|
|
185 |
\subsection{exch2, SIZE.h, and multiprocessing} |
\subsection{exch2, SIZE.h, and Multiprocessing} |
186 |
\label{sec:exch2mpi} |
\label{sec:exch2mpi} |
187 |
|
|
188 |
Once the topology configuration files are created, the Fortran |
Once the topology configuration files are created, the Fortran |
190 |
Section \ref{sect:specifying_a_decomposition} \sectiontitle{Specifying |
Section \ref{sect:specifying_a_decomposition} \sectiontitle{Specifying |
191 |
a decomposition} provides a general description of domain |
a decomposition} provides a general description of domain |
192 |
decomposition within MITgcm and its relation to \file{SIZE.h}. The |
decomposition within MITgcm and its relation to \file{SIZE.h}. The |
193 |
current section specifies certain constraints the exch2 package |
current section specifies constraints that the exch2 package |
194 |
imposes as well as describes how to enable parallel execution with |
imposes and describes how to enable parallel execution with |
195 |
MPI. \\ |
MPI. \\ |
196 |
|
|
197 |
As in the general case, the parameters \varlink{sNx}{sNx} and |
As in the general case, the parameters \varlink{sNx}{sNx} and |
207 |
The parameters \varlink{nSx}{nSx}, \varlink{nSy}{nSy}, |
The parameters \varlink{nSx}{nSx}, \varlink{nSy}{nSy}, |
208 |
\varlink{nPx}{nPx}, and \varlink{nPy}{nPy} relate to the number of |
\varlink{nPx}{nPx}, and \varlink{nPy}{nPy} relate to the number of |
209 |
tiles and how they are distributed on processors. When using exch2, |
tiles and how they are distributed on processors. When using exch2, |
210 |
the tiles are stored in a single dimension, and so |
the tiles are stored in the $x$ dimension, and so |
211 |
\code{\varlink{nSy}{nSy}=1} in all cases. Since the tiles as |
\code{\varlink{nSy}{nSy}=1} in all cases. Since the tiles as |
212 |
configured by exch2 cannot be split up accross processors without |
configured by exch2 cannot be split up accross processors without |
213 |
regenerating the topology, \code{\varlink{nPy}{nPy}=1} as well. \\ |
regenerating the topology, \code{\varlink{nPy}{nPy}=1} as well. \\ |
215 |
The number of tiles MITgcm allocates and how they are distributed |
The number of tiles MITgcm allocates and how they are distributed |
216 |
between processors depends on \varlink{nPx}{nPx} and |
between processors depends on \varlink{nPx}{nPx} and |
217 |
\varlink{nSx}{nSx}. \varlink{nSx}{nSx} is the number of tiles per |
\varlink{nSx}{nSx}. \varlink{nSx}{nSx} is the number of tiles per |
218 |
processor and \varlink{nPx}{nPx} the number of processors. The total |
processor and \varlink{nPx}{nPx} is the number of processors. The |
219 |
number of tiles in the topology minus those listed in |
total number of tiles in the topology minus those listed in |
220 |
\file{blanklist.txt} must equal \code{nSx*nPx}. \\ |
\file{blanklist.txt} must equal \code{nSx*nPx}. Note that in order to |
221 |
|
obtain maximum usage from a given number of processors in some cases, |
222 |
|
this restriction might entail sharing a processor with a tile that |
223 |
|
would otherwise be excluded because it is topographically outside of |
224 |
|
the domain and therefore in \file{blanklist.txt}. For example, |
225 |
|
suppose you have five processors and a domain decomposition of |
226 |
|
thirty-six tiles that allows you to exclude seven tiles. To evenly |
227 |
|
distribute the remaining twenty-nine tiles among five processors, you |
228 |
|
would have to run one ``dummy'' tile to make an even six tiles per |
229 |
|
processor. Such dummy tiles are \emph{not} listed in |
230 |
|
\file{blanklist.txt}.\\ |
231 |
|
|
232 |
|
|
233 |
The following is an example of \file{SIZE.h} for the twelve-tile |
The following is an example of \file{SIZE.h} for the twelve-tile |
234 |
configuration illustrated in figure \ref{fig:12tile} running on |
configuration illustrated in figure \ref{fig:12tile} running on |
298 |
topology of twenty-four square tiles, four per subdomain (as in figure |
topology of twenty-four square tiles, four per subdomain (as in figure |
299 |
\ref{fig:24tile}), will have \code{exch2\_domain\_nxt=12} and |
\ref{fig:24tile}), will have \code{exch2\_domain\_nxt=12} and |
300 |
\code{exch2\_domain\_nyt=2}. Note that these parameters express the |
\code{exch2\_domain\_nyt=2}. Note that these parameters express the |
301 |
tile layout to allow global data files that are tile-layout-neutral |
tile layout in order to allow global data files that are tile-layout-neutral. |
302 |
and have no bearing on the internal storage of the arrays. The tiles |
They have no bearing on the internal storage of the arrays. The tiles |
303 |
are stored internally in a range from \code{(1:\varlink{bi}{bi})} the |
are stored internally in a range from \code{\varlink{bi}{bi}=(1:NTILES)} in the |
304 |
$x$ axis, and the $y$ axis variable \varlink{bj}{bj} generally is |
$x$ axis, and the $y$ axis variable \varlink{bj}{bj} is assumed to |
305 |
ignored within the package. \\ |
equal \code{1} throughout the package. \\ |
306 |
|
|
307 |
\subsubsection{Arrays Indexed to Tile Number} |
\subsubsection{Arrays indexed to tile number} |
308 |
|
|
309 |
The following arrays are of length \code{NTILES} and are indexed to |
The following arrays are of length \code{NTILES} and are indexed to |
310 |
the tile number, which is indicated in the diagrams with the notation |
the tile number, which is indicated in the diagrams with the notation |
314 |
\varlink{exch2\_tny}{exch2_tny} express the $x$ and $y$ dimensions of |
\varlink{exch2\_tny}{exch2_tny} express the $x$ and $y$ dimensions of |
315 |
each tile. At present for each tile \texttt{exch2\_tnx=sNx} and |
each tile. At present for each tile \texttt{exch2\_tnx=sNx} and |
316 |
\texttt{exch2\_tny=sNy}, as assigned in \file{SIZE.h} and described in |
\texttt{exch2\_tny=sNy}, as assigned in \file{SIZE.h} and described in |
317 |
section \ref{sec:exch2mpi} \sectiontitle{exch2, SIZE.h, and |
Section \ref{sec:exch2mpi} \sectiontitle{exch2, SIZE.h, and |
318 |
multiprocessing}. Future releases of MITgcm are to allow varying tile |
Multiprocessing}. Future releases of MITgcm may allow varying tile |
319 |
sizes. \\ |
sizes. \\ |
320 |
|
|
321 |
The location of the tiles' Cartesian origin within a subdomain are |
The arrays \varlink{exch2\_tbasex}{exch2_tbasex} and |
322 |
determined by the arrays \varlink{exch2\_tbasex}{exch2_tbasex} and |
\varlink{exch2\_tbasey}{exch2_tbasey} determine the tiles' |
323 |
\varlink{exch2\_tbasey}{exch2_tbasey}. These variables are used to |
Cartesian origin within a subdomain |
324 |
relate the location of the edges of different tiles to each other. As |
and locate the edges of different tiles relative to each other. As |
325 |
an example, in the default six-tile topology (Fig. \ref{fig:6tile}) |
an example, in the default six-tile topology (Fig. \ref{fig:6tile}) |
326 |
each index in these arrays is set to \code{0} since a tile occupies |
each index in these arrays is set to \code{0} since a tile occupies |
327 |
its entire subdomain. The twenty-four-tile case discussed above will |
its entire subdomain. The twenty-four-tile case discussed above will |
328 |
have values of \code{0} or \code{16}, depending on the quadrant the |
have values of \code{0} or \code{16}, depending on the quadrant of the |
329 |
tile falls within the subdomain. The elements of the arrays |
tile within the subdomain. The elements of the arrays |
330 |
\varlink{exch2\_txglobalo}{exch2_txglobalo} and |
\varlink{exch2\_txglobalo}{exch2_txglobalo} and |
331 |
\varlink{exch2\_txglobalo}{exch2_txglobalo} are similar to |
\varlink{exch2\_txglobalo}{exch2_txglobalo} are similar to |
332 |
\varlink{exch2\_tbasex}{exch2_tbasex} and |
\varlink{exch2\_tbasex}{exch2_tbasex} and |
333 |
\varlink{exch2\_tbasey}{exch2_tbasey}, but locate the tiles within the |
\varlink{exch2\_tbasey}{exch2_tbasey}, but locate the tile edges within the |
334 |
global address space, similar to that used by global output and input |
global address space, similar to that used by global output and input |
335 |
files. \\ |
files. \\ |
336 |
|
|
339 |
standard cube topology and indicated by \textbf{\textsf{f}}$n$ in |
standard cube topology and indicated by \textbf{\textsf{f}}$n$ in |
340 |
figures \ref{fig:12tile} and |
figures \ref{fig:12tile} and |
341 |
\ref{fig:24tile}. \varlink{exch2\_nNeighbours}{exch2_nNeighbours} |
\ref{fig:24tile}. \varlink{exch2\_nNeighbours}{exch2_nNeighbours} |
342 |
contains a count of the neighboring tiles each tile has, and is used |
contains a count of the neighboring tiles each tile has, and sets |
343 |
for setting bounds for looping over neighboring tiles. |
the bounds for looping over neighboring tiles. |
344 |
\varlink{exch2\_tProc}{exch2_tProc} holds the process rank of each |
\varlink{exch2\_tProc}{exch2_tProc} holds the process rank of each |
345 |
tile, and is used in interprocess communication. \\ |
tile, and is used in interprocess communication. \\ |
346 |
|
|
349 |
\varlink{exch2\_isEedge}{exch2_isEedge}, |
\varlink{exch2\_isEedge}{exch2_isEedge}, |
350 |
\varlink{exch2\_isSedge}{exch2_isSedge}, and |
\varlink{exch2\_isSedge}{exch2_isSedge}, and |
351 |
\varlink{exch2\_isNedge}{exch2_isNedge} are set to \code{1} if the |
\varlink{exch2\_isNedge}{exch2_isNedge} are set to \code{1} if the |
352 |
indexed tile lies on the respective edge of a subdomain, \code{0} if |
indexed tile lies on the edge of its subdomain, \code{0} if |
353 |
not. The values are used within the topology generator to determine |
not. The values are used within the topology generator to determine |
354 |
the orientation of neighboring tiles, and to indicate whether a tile |
the orientation of neighboring tiles, and to indicate whether a tile |
355 |
lies on the corner of a subdomain. The latter case requires special |
lies on the corner of a subdomain. The latter case requires special |
382 |
The arrays \varlink{exch2\_pi}{exch2_pi} and |
The arrays \varlink{exch2\_pi}{exch2_pi} and |
383 |
\varlink{exch2\_pj}{exch2_pj} specify the transformations of indices |
\varlink{exch2\_pj}{exch2_pj} specify the transformations of indices |
384 |
in exchanges between the neighboring tiles. These transformations are |
in exchanges between the neighboring tiles. These transformations are |
385 |
necessary in exchanges between subdomains because the array index in |
necessary in exchanges between subdomains because a horizontal dimension |
386 |
one dimension may map to the other index in an adjacent subdomain, and |
in one subdomain |
387 |
may be have its indexing reversed. This swapping arises from the |
may map to other horizonal dimension in an adjacent subdomain, and |
388 |
|
may also have its indexing reversed. This swapping arises from the |
389 |
``folding'' of two-dimensional arrays into a three-dimensional |
``folding'' of two-dimensional arrays into a three-dimensional |
390 |
cube. \\ |
cube. \\ |
391 |
|
|
393 |
are the neighbor ID \code{N} and the tile number \code{T} as explained |
are the neighbor ID \code{N} and the tile number \code{T} as explained |
394 |
above, plus a vector of length \code{2} containing transformation |
above, plus a vector of length \code{2} containing transformation |
395 |
factors \code{t}. The first element of the transformation vector |
factors \code{t}. The first element of the transformation vector |
396 |
holds the factor to multiply the index in the same axis, and the |
holds the factor to multiply the index in the same dimension, and the |
397 |
second element holds the the same for the orthogonal index. To |
second element holds the the same for the orthogonal dimension. To |
398 |
clarify, \code{exch2\_pi(1,N,T)} holds the mapping of the $x$ axis |
clarify, \code{exch2\_pi(1,N,T)} holds the mapping of the $x$ axis |
399 |
index of tile \code{T} to the $x$ axis of tile \code{T}'s neighbor |
index of tile \code{T} to the $x$ axis of tile \code{T}'s neighbor |
400 |
\code{N}, and \code{exch2\_pi(2,N,T)} holds the mapping of \code{T}'s |
\code{N}, and \code{exch2\_pi(2,N,T)} holds the mapping of \code{T}'s |
409 |
\code{(1,0)}, since all tiles on the same subdomain are oriented |
\code{(1,0)}, since all tiles on the same subdomain are oriented |
410 |
identically. An axis that corresponds to the orthogonal dimension |
identically. An axis that corresponds to the orthogonal dimension |
411 |
with the same index direction in a particular tile-neighbor |
with the same index direction in a particular tile-neighbor |
412 |
orientation will have \code{(0,1)}. Those in the opposite index |
orientation will have \code{(0,1)}. Those with the opposite index |
413 |
direction will have \code{(0,-1)} in order to reverse the ordering. \\ |
direction will have \code{(0,-1)} in order to reverse the ordering. \\ |
414 |
|
|
415 |
The arrays \varlink{exch2\_oi}{exch2_oi}, |
The arrays \varlink{exch2\_oi}{exch2_oi}, |
495 |
\code{Tn}'s $y$ axis corresponds to \code{T}'s $x$ axis, \code{T}'s |
\code{Tn}'s $y$ axis corresponds to \code{T}'s $x$ axis, \code{T}'s |
496 |
northern edge exchanges with \code{Tn}'s western edge. The western |
northern edge exchanges with \code{Tn}'s western edge. The western |
497 |
edge of the tiles corresponds to the lower bound of the $x$ axis, so |
edge of the tiles corresponds to the lower bound of the $x$ axis, so |
498 |
\code{exch2\_itlo\_c} \code{exch2\_ithi\_c} are \code{0}. The range of |
\code{exch2\_itlo\_c} and \code{exch2\_ithi\_c} are \code{0}, in the |
499 |
|
western halo region of \code{Tn}. The range of |
500 |
\code{exch2\_jtlo\_c} and \code{exch2\_jthi\_c} correspond to the |
\code{exch2\_jtlo\_c} and \code{exch2\_jthi\_c} correspond to the |
501 |
width of \code{T}'s northern edge, plus the halo. \\ |
width of \code{T}'s northern edge, expanded by one into the halo. \\ |
502 |
|
|
503 |
|
|
504 |
\subsection{Key Routines} |
\subsection{Key Routines} |
507 |
themselves and are of the same format as those described in |
themselves and are of the same format as those described in |
508 |
\ref{sect:cube_sphere_communication} \sectiontitle{Cube sphere |
\ref{sect:cube_sphere_communication} \sectiontitle{Cube sphere |
509 |
communication}. Like the original routines, they are written as |
communication}. Like the original routines, they are written as |
510 |
templates which the local Makefile converts from RX into RL and RS |
templates which the local Makefile converts from \code{RX} into |
511 |
forms. \\ |
\code{RL} and \code{RS} forms. \\ |
512 |
|
|
513 |
The interfaces with the core model subroutines are |
The interfaces with the core model subroutines are |
514 |
\code{EXCH\_UV\_XY\_RX}, \code{EXCH\_UV\_XYZ\_RX} and |
\code{EXCH\_UV\_XY\_RX}, \code{EXCH\_UV\_XYZ\_RX} and |
523 |
the singularities at the cube corners. \\ |
the singularities at the cube corners. \\ |
524 |
|
|
525 |
The separate scalar and vector forms of \code{EXCH2\_RX1\_CUBE} and |
The separate scalar and vector forms of \code{EXCH2\_RX1\_CUBE} and |
526 |
\code{EXCH2\_RX2\_CUBE} reflect that the vector-handling subrouine |
\code{EXCH2\_RX2\_CUBE} reflect that the vector-handling subroutine |
527 |
needs to pass both the $u$ and $v$ components of the phsical vectors. |
needs to pass both the $u$ and $v$ components of the physical vectors. |
528 |
This arises from the topological folding discussed above, where the |
This swapping arises from the topological folding discussed above, where the |
529 |
$x$ and $y$ axes get swapped in some cases. This swapping is not an |
$x$ and $y$ axes get swapped in some cases, and is not an |
530 |
issue with the scalar version. These subroutines call |
issue with the scalar case. These subroutines call |
531 |
\code{EXCH2\_SEND\_RX1} and \code{EXCH2\_SEND\_RX2}, which do most of |
\code{EXCH2\_SEND\_RX1} and \code{EXCH2\_SEND\_RX2}, which do most of |
532 |
the work using the variables discussed above. \\ |
the work using the variables discussed above. \\ |
533 |
|
|