3 |
|
|
4 |
%\section{Getting started} |
%\section{Getting started} |
5 |
|
|
6 |
In this section, we describe how to use the model. In the first |
We believe the best way to familiarize yourself with the |
|
section, we provide enough information to help you get started with |
|
|
the model. We believe the best way to familiarize yourself with the |
|
7 |
model is to run the case study examples provided with the base |
model is to run the case study examples provided with the base |
8 |
version. Information on how to obtain, compile, and run the code is |
version. Information on how to obtain, compile, and run the code is |
9 |
found there as well as a brief description of the model structure |
found here as well as a brief description of the model structure |
10 |
directory and the case study examples. The latter and the code |
directory and the case study examples. Information is also provided |
11 |
structure are described more fully in chapters |
here on how to customize the code when you are ready to try implementing |
12 |
\ref{chap:discretization} and \ref{chap:sarch}, respectively. Here, in |
the configuration you have in mind. The code and algorithm |
13 |
this section, we provide information on how to customize the code when |
are described more fully in chapters \ref{chap:discretization} and |
14 |
you are ready to try implementing the configuration you have in mind. |
\ref{chap:sarch}. |
|
|
|
15 |
|
|
16 |
\section{Where to find information} |
\section{Where to find information} |
17 |
\label{sect:whereToFindInfo} |
\label{sect:whereToFindInfo} |
19 |
<!-- CMIREDIR:whereToFindInfo: --> |
<!-- CMIREDIR:whereToFindInfo: --> |
20 |
\end{rawhtml} |
\end{rawhtml} |
21 |
|
|
22 |
A web site is maintained for release 2 (``Pelican'') of MITgcm: |
There is a web-archived support mailing list for the model that |
|
\begin{rawhtml} <A href=http://mitgcm.org/pelican/ target="idontexist"> \end{rawhtml} |
|
|
\begin{verbatim} |
|
|
http://mitgcm.org/pelican |
|
|
\end{verbatim} |
|
|
\begin{rawhtml} </A> \end{rawhtml} |
|
|
Here you will find an on-line version of this document, a |
|
|
``browsable'' copy of the code and a searchable database of the model |
|
|
and site, as well as links for downloading the model and |
|
|
documentation, to data-sources, and other related sites. |
|
|
|
|
|
There is also a web-archived support mailing list for the model that |
|
23 |
you can email at \texttt{MITgcm-support@mitgcm.org} or browse at: |
you can email at \texttt{MITgcm-support@mitgcm.org} or browse at: |
24 |
\begin{rawhtml} <A href=http://mitgcm.org/mailman/listinfo/mitgcm-support/ target="idontexist"> \end{rawhtml} |
\begin{rawhtml} <A href=http://mitgcm.org/mailman/listinfo/mitgcm-support/ target="idontexist"> \end{rawhtml} |
25 |
\begin{verbatim} |
\begin{verbatim} |
27 |
http://mitgcm.org/pipermail/mitgcm-support/ |
http://mitgcm.org/pipermail/mitgcm-support/ |
28 |
\end{verbatim} |
\end{verbatim} |
29 |
\begin{rawhtml} </A> \end{rawhtml} |
\begin{rawhtml} </A> \end{rawhtml} |
|
Essentially all of the MITgcm web pages can be searched using a |
|
|
popular web crawler such as Google or through our own search facility: |
|
|
\begin{rawhtml} <A href=http://mitgcm.org/mailman/htdig/ target="idontexist"> \end{rawhtml} |
|
|
\begin{verbatim} |
|
|
http://mitgcm.org/htdig/ |
|
|
\end{verbatim} |
|
|
\begin{rawhtml} </A> \end{rawhtml} |
|
|
%%% http://www.google.com/search?q=hydrostatic+site%3Amitgcm.org |
|
|
|
|
|
|
|
30 |
|
|
31 |
\section{Obtaining the code} |
\section{Obtaining the code} |
32 |
\label{sect:obtainingCode} |
\label{sect:obtainingCode} |
102 |
code and CVS. It also contains a web interface to our CVS archive so |
code and CVS. It also contains a web interface to our CVS archive so |
103 |
that one may easily view the state of files, revisions, and other |
that one may easily view the state of files, revisions, and other |
104 |
development milestones: |
development milestones: |
105 |
\begin{rawhtml} <A href="http://mitgcm.org/download" target="idontexist"> \end{rawhtml} |
%\begin{rawhtml} <A href="http://mitgcm.org/download" target="idontexist"> \end{rawhtml} |
106 |
|
\begin{rawhtml} <A href="http://mitgcm.org/viewvc/MITgcm/MITgcm/" target="idontexist"> \end{rawhtml} |
107 |
\begin{verbatim} |
\begin{verbatim} |
108 |
http://mitgcm.org/source_code.html |
http://mitgcm.org/source_code.html |
109 |
\end{verbatim} |
\end{verbatim} |
140 |
the files in \texttt{CVS}! You can also use CVS to download code |
the files in \texttt{CVS}! You can also use CVS to download code |
141 |
updates. More extensive information on using CVS for maintaining |
updates. More extensive information on using CVS for maintaining |
142 |
MITgcm code can be found |
MITgcm code can be found |
143 |
\begin{rawhtml} <A href="http://mitgcm.org/usingcvstoget.html" target="idontexist"> \end{rawhtml} |
\begin{rawhtml} <A href="http://mitgcm.org/public/using_cvs.html" target="idontexist"> \end{rawhtml} |
144 |
here |
here |
145 |
\begin{rawhtml} </A> \end{rawhtml} |
\begin{rawhtml} </A> \end{rawhtml} |
146 |
. |
. |
153 |
% mv MITgcm MITgcm_verif_basic |
% mv MITgcm MITgcm_verif_basic |
154 |
\end{verbatim} |
\end{verbatim} |
155 |
|
|
|
|
|
|
\subsection{Method 2 - Tar file download} |
|
|
\label{sect:conventionalDownload} |
|
|
|
|
|
If you do not have CVS on your system, you can download the model as a |
|
|
tar file from the web site at: |
|
|
\begin{rawhtml} <A href=http://mitgcm.org/download target="idontexist"> \end{rawhtml} |
|
|
\begin{verbatim} |
|
|
http://mitgcm.org/download/ |
|
|
\end{verbatim} |
|
|
\begin{rawhtml} </A> \end{rawhtml} |
|
|
The tar file still contains CVS information which we urge you not to |
|
|
delete; even if you do not use CVS yourself the information can help |
|
|
us if you should need to send us your copy of the code. If a recent |
|
|
tar file does not exist, then please contact the developers through |
|
|
the |
|
|
\begin{rawhtml} <A href="mailto:MITgcm-support@mitgcm.org"> \end{rawhtml} |
|
|
MITgcm-support@mitgcm.org |
|
|
\begin{rawhtml} </A> \end{rawhtml} |
|
|
mailing list. |
|
|
|
|
156 |
\subsubsection{Upgrading from an earlier version} |
\subsubsection{Upgrading from an earlier version} |
157 |
|
|
158 |
If you already have an earlier version of the code you can ``upgrade'' |
If you already have an earlier version of the code you can ``upgrade'' |
219 |
also means we can't tell what version of the code you are working |
also means we can't tell what version of the code you are working |
220 |
with. So please be sure you understand what you're doing. |
with. So please be sure you understand what you're doing. |
221 |
|
|
222 |
|
\subsection{Method 2 - Tar file download} |
223 |
|
\label{sect:conventionalDownload} |
224 |
|
|
225 |
|
If you do not have CVS on your system, you can download the model as a |
226 |
|
tar file from the web site at: |
227 |
|
\begin{rawhtml} <A href=http://mitgcm.org/download/ target="idontexist"> \end{rawhtml} |
228 |
|
\begin{verbatim} |
229 |
|
http://mitgcm.org/download/ |
230 |
|
\end{verbatim} |
231 |
|
\begin{rawhtml} </A> \end{rawhtml} |
232 |
|
The tar file still contains CVS information which we urge you not to |
233 |
|
delete; even if you do not use CVS yourself the information can help |
234 |
|
us if you should need to send us your copy of the code. If a recent |
235 |
|
tar file does not exist, then please contact the developers through |
236 |
|
the |
237 |
|
\begin{rawhtml} <A href="mailto:MITgcm-support@mitgcm.org"> \end{rawhtml} |
238 |
|
MITgcm-support@mitgcm.org |
239 |
|
\begin{rawhtml} </A> \end{rawhtml} |
240 |
|
mailing list. |
241 |
|
|
242 |
\section{Model and directory structure} |
\section{Model and directory structure} |
243 |
\begin{rawhtml} |
\begin{rawhtml} |
244 |
<!-- CMIREDIR:directory_structure: --> |
<!-- CMIREDIR:directory_structure: --> |
263 |
|
|
264 |
\begin{itemize} |
\begin{itemize} |
265 |
|
|
|
\item \texttt{bin}: this directory is initially empty. It is the |
|
|
default directory in which to compile the code. |
|
|
|
|
|
\item \texttt{diags}: contains the code relative to time-averaged |
|
|
diagnostics. It is subdivided into two subdirectories \texttt{inc} |
|
|
and \texttt{src} that contain include files (\texttt{*.h} files) and |
|
|
Fortran subroutines (\texttt{*.F} files), respectively. |
|
|
|
|
266 |
\item \texttt{doc}: contains brief documentation notes. |
\item \texttt{doc}: contains brief documentation notes. |
267 |
|
|
268 |
\item \texttt{eesupp}: contains the execution environment source code. |
\item \texttt{eesupp}: contains the execution environment source code. |
269 |
Also subdivided into two subdirectories \texttt{inc} and |
Also subdivided into two subdirectories \texttt{inc} and |
270 |
\texttt{src}. |
\texttt{src}. |
271 |
|
|
|
\item \texttt{exe}: this directory is initially empty. It is the |
|
|
default directory in which to execute the code. |
|
|
|
|
272 |
\item \texttt{model}: this directory contains the main source code. |
\item \texttt{model}: this directory contains the main source code. |
273 |
Also subdivided into two subdirectories \texttt{inc} and |
Also subdivided into two subdirectories \texttt{inc} and |
274 |
\texttt{src}. |
\texttt{src}. |
277 |
package corresponds to a subdirectory. For example, \texttt{gmredi} |
package corresponds to a subdirectory. For example, \texttt{gmredi} |
278 |
contains the code related to the Gent-McWilliams/Redi scheme, |
contains the code related to the Gent-McWilliams/Redi scheme, |
279 |
\texttt{aim} the code relative to the atmospheric intermediate |
\texttt{aim} the code relative to the atmospheric intermediate |
280 |
physics. The packages are described in detail in section 3. |
physics. The packages are described in detail in chapter \ref{chap.packagesI}. |
281 |
|
|
282 |
\item \texttt{tools}: this directory contains various useful tools. |
\item \texttt{tools}: this directory contains various useful tools. |
283 |
For example, \texttt{genmake2} is a script written in csh (C-shell) |
For example, \texttt{genmake2} is a script written in csh (C-shell) |
284 |
that should be used to generate your makefile. The directory |
that should be used to generate your makefile. The directory |
285 |
\texttt{adjoint} contains the makefile specific to the Tangent |
\texttt{adjoint} contains the makefile specific to the Tangent |
286 |
linear and Adjoint Compiler (TAMC) that generates the adjoint code. |
linear and Adjoint Compiler (TAMC) that generates the adjoint code. |
287 |
The latter is described in details in part V. |
The latter is described in detail in part \ref{chap.ecco}. |
288 |
|
This directory also contains the subdirectory build\_options, which |
289 |
|
contains the `optfiles' with the compiler options for the different |
290 |
|
compilers and machines that can run MITgcm. |
291 |
|
|
292 |
\item \texttt{utils}: this directory contains various utilities. The |
\item \texttt{utils}: this directory contains various utilities. The |
293 |
subdirectory \texttt{knudsen2} contains code and a makefile that |
subdirectory \texttt{knudsen2} contains code and a makefile that |
296 |
\texttt{matlab} subdirectory contains matlab scripts for reading |
\texttt{matlab} subdirectory contains matlab scripts for reading |
297 |
model output directly into matlab. \texttt{scripts} contains C-shell |
model output directly into matlab. \texttt{scripts} contains C-shell |
298 |
post-processing scripts for joining processor-based and tiled-based |
post-processing scripts for joining processor-based and tiled-based |
299 |
model output. |
model output. The subdirectory exch2 contains the code needed for |
300 |
|
the exch2 package to work with different combinations of domain |
301 |
|
decompositions. |
302 |
|
|
303 |
\item \texttt{verification}: this directory contains the model |
\item \texttt{verification}: this directory contains the model |
304 |
examples. See section \ref{sect:modelExamples}. |
examples. See section \ref{sect:modelExamples}. |
305 |
|
|
306 |
\end{itemize} |
\item \texttt{jobs}: contains sample job scripts for running MITgcm. |
|
|
|
|
\section[MITgcm Example Experiments]{Example experiments} |
|
|
\label{sect:modelExamples} |
|
|
\begin{rawhtml} |
|
|
<!-- CMIREDIR:modelExamples: --> |
|
|
\end{rawhtml} |
|
|
|
|
|
%% a set of twenty-four pre-configured numerical experiments |
|
|
|
|
|
The full MITgcm distribution comes with more than a dozen |
|
|
pre-configured numerical experiments. Some of these example |
|
|
experiments are tests of individual parts of the model code, but many |
|
|
are fully fledged numerical simulations. A few of the examples are |
|
|
used for tutorial documentation in sections \ref{sect:eg-baro} - |
|
|
\ref{sect:eg-global}. The other examples follow the same general |
|
|
structure as the tutorial examples. However, they only include brief |
|
|
instructions in a text file called {\it README}. The examples are |
|
|
located in subdirectories under the directory \texttt{verification}. |
|
|
Each example is briefly described below. |
|
|
|
|
|
\subsection{Full list of model examples} |
|
|
|
|
|
\begin{enumerate} |
|
|
|
|
|
\item \texttt{exp0} - single layer, ocean double gyre (barotropic with |
|
|
free-surface). This experiment is described in detail in section |
|
|
\ref{sect:eg-baro}. |
|
|
|
|
|
\item \texttt{exp1} - Four layer, ocean double gyre. This experiment |
|
|
is described in detail in section \ref{sect:eg-baroc}. |
|
|
|
|
|
\item \texttt{exp2} - 4x4 degree global ocean simulation with steady |
|
|
climatological forcing. This experiment is described in detail in |
|
|
section \ref{sect:eg-global}. |
|
|
|
|
|
\item \texttt{exp4} - Flow over a Gaussian bump in open-water or |
|
|
channel with open boundaries. |
|
|
|
|
|
\item \texttt{exp5} - Inhomogenously forced ocean convection in a |
|
|
doubly periodic box. |
|
|
|
|
|
\item \texttt{front\_relax} - Relaxation of an ocean thermal front (test for |
|
|
Gent/McWilliams scheme). 2D (Y-Z). |
|
|
|
|
|
\item \texttt{internal wave} - Ocean internal wave forced by open |
|
|
boundary conditions. |
|
|
|
|
|
\item \texttt{natl\_box} - Eastern subtropical North Atlantic with KPP |
|
|
scheme; 1 month integration |
|
|
|
|
|
\item \texttt{hs94.1x64x5} - Zonal averaged atmosphere using Held and |
|
|
Suarez '94 forcing. |
|
|
|
|
|
\item \texttt{hs94.128x64x5} - 3D atmosphere dynamics using Held and |
|
|
Suarez '94 forcing. |
|
|
|
|
|
\item \texttt{hs94.cs-32x32x5} - 3D atmosphere dynamics using Held and |
|
|
Suarez '94 forcing on the cubed sphere. |
|
|
|
|
|
\item \texttt{aim.5l\_zon-ave} - Intermediate Atmospheric physics. |
|
|
Global Zonal Mean configuration, 1x64x5 resolution. |
|
|
|
|
|
\item \texttt{aim.5l\_XZ\_Equatorial\_Slice} - Intermediate |
|
|
Atmospheric physics, equatorial Slice configuration. 2D (X-Z). |
|
|
|
|
|
\item \texttt{aim.5l\_Equatorial\_Channel} - Intermediate Atmospheric |
|
|
physics. 3D Equatorial Channel configuration. |
|
|
|
|
|
\item \texttt{aim.5l\_LatLon} - Intermediate Atmospheric physics. |
|
|
Global configuration, on latitude longitude grid with 128x64x5 grid |
|
|
points ($2.8^\circ$ resolution). |
|
|
|
|
|
\item \texttt{adjustment.128x64x1} Barotropic adjustment problem on |
|
|
latitude longitude grid with 128x64 grid points ($2.8^\circ$ resolution). |
|
|
|
|
|
\item \texttt{adjustment.cs-32x32x1} Barotropic adjustment problem on |
|
|
cube sphere grid with 32x32 points per face (roughly $2.8^\circ$ |
|
|
resolution). |
|
|
|
|
|
\item \texttt{advect\_cs} Two-dimensional passive advection test on |
|
|
cube sphere grid. |
|
|
|
|
|
\item \texttt{advect\_xy} Two-dimensional (horizontal plane) passive |
|
|
advection test on Cartesian grid. |
|
|
|
|
|
\item \texttt{advect\_yz} Two-dimensional (vertical plane) passive |
|
|
advection test on Cartesian grid. |
|
|
|
|
|
\item \texttt{carbon} Simple passive tracer experiment. Includes |
|
|
derivative calculation. Described in detail in section |
|
|
\ref{sect:eg-carbon-ad}. |
|
|
|
|
|
\item \texttt{flt\_example} Example of using float package. |
|
|
|
|
|
\item \texttt{global\_ocean.90x40x15} Global circulation with GM, flux |
|
|
boundary conditions and poles. |
|
|
|
|
|
\item \texttt{global\_ocean\_pressure} Global circulation in pressure |
|
|
coordinate (non-Boussinesq ocean model). Described in detail in |
|
|
section \ref{sect:eg-globalpressure}. |
|
|
|
|
|
\item \texttt{solid-body.cs-32x32x1} Solid body rotation test for cube |
|
|
sphere grid. |
|
|
|
|
|
\end{enumerate} |
|
|
|
|
|
\subsection{Directory structure of model examples} |
|
|
|
|
|
Each example directory has the following subdirectories: |
|
|
|
|
|
\begin{itemize} |
|
|
\item \texttt{code}: contains the code particular to the example. At a |
|
|
minimum, this directory includes the following files: |
|
|
|
|
|
\begin{itemize} |
|
|
\item \texttt{code/packages.conf}: declares the list of packages or |
|
|
package groups to be used. If not included, the default version |
|
|
is located in \texttt{pkg/pkg\_default}. Package groups are |
|
|
simply convenient collections of commonly used packages which are |
|
|
defined in \texttt{pkg/pkg\_default}. Some packages may require |
|
|
other packages or may require their absence (that is, they are |
|
|
incompatible) and these package dependencies are listed in |
|
|
\texttt{pkg/pkg\_depend}. |
|
|
|
|
|
\item \texttt{code/CPP\_EEOPTIONS.h}: declares CPP keys relative to |
|
|
the ``execution environment'' part of the code. The default |
|
|
version is located in \texttt{eesupp/inc}. |
|
|
|
|
|
\item \texttt{code/CPP\_OPTIONS.h}: declares CPP keys relative to |
|
|
the ``numerical model'' part of the code. The default version is |
|
|
located in \texttt{model/inc}. |
|
|
|
|
|
\item \texttt{code/SIZE.h}: declares size of underlying |
|
|
computational grid. The default version is located in |
|
|
\texttt{model/inc}. |
|
|
\end{itemize} |
|
307 |
|
|
308 |
In addition, other include files and subroutines might be present in |
\item \texttt{lsopt}: Line search code used for optimization. |
|
\texttt{code} depending on the particular experiment. See Section 2 |
|
|
for more details. |
|
|
|
|
|
\item \texttt{input}: contains the input data files required to run |
|
|
the example. At a minimum, the \texttt{input} directory contains the |
|
|
following files: |
|
|
|
|
|
\begin{itemize} |
|
|
\item \texttt{input/data}: this file, written as a namelist, |
|
|
specifies the main parameters for the experiment. |
|
309 |
|
|
310 |
\item \texttt{input/data.pkg}: contains parameters relative to the |
\item \texttt{optim}: Interface between MITgcm and line search code. |
|
packages used in the experiment. |
|
311 |
|
|
|
\item \texttt{input/eedata}: this file contains ``execution |
|
|
environment'' data. At present, this consists of a specification |
|
|
of the number of threads to use in $X$ and $Y$ under multithreaded |
|
|
execution. |
|
|
\end{itemize} |
|
|
|
|
|
In addition, you will also find in this directory the forcing and |
|
|
topography files as well as the files describing the initial state |
|
|
of the experiment. This varies from experiment to experiment. See |
|
|
section 2 for more details. |
|
|
|
|
|
\item \texttt{results}: this directory contains the output file |
|
|
\texttt{output.txt} produced by the simulation example. This file is |
|
|
useful for comparison with your own output when you run the |
|
|
experiment. |
|
312 |
\end{itemize} |
\end{itemize} |
313 |
|
|
|
Once you have chosen the example you want to run, you are ready to |
|
|
compile the code. |
|
|
|
|
314 |
\section[Building MITgcm]{Building the code} |
\section[Building MITgcm]{Building the code} |
315 |
\label{sect:buildingCode} |
\label{sect:buildingCode} |
316 |
\begin{rawhtml} |
\begin{rawhtml} |
408 |
where we are re-directing the stream of text output to the file |
where we are re-directing the stream of text output to the file |
409 |
\texttt{output.txt}. |
\texttt{output.txt}. |
410 |
|
|
411 |
|
\subsection{Building/compiling the code elsewhere} |
412 |
|
|
413 |
|
In the example above (section \ref{sect:buildingCode}) we built the |
414 |
|
executable in the {\em input} directory of the experiment for |
415 |
|
convenience. You can also configure and compile the code in other |
416 |
|
locations, for example on a scratch disk with out having to copy the |
417 |
|
entire source tree. The only requirement to do so is you have {\tt |
418 |
|
genmake2} in your path or you know the absolute path to {\tt |
419 |
|
genmake2}. |
420 |
|
|
421 |
|
The following sections outline some possible methods of organizing |
422 |
|
your source and data. |
423 |
|
|
424 |
|
\subsubsection{Building from the {\em ../code directory}} |
425 |
|
|
426 |
|
This is just as simple as building in the {\em input/} directory: |
427 |
|
\begin{verbatim} |
428 |
|
% cd verification/exp2/code |
429 |
|
% ../../../tools/genmake2 |
430 |
|
% make depend |
431 |
|
% make |
432 |
|
\end{verbatim} |
433 |
|
However, to run the model the executable ({\em mitgcmuv}) and input |
434 |
|
files must be in the same place. If you only have one calculation to make: |
435 |
|
\begin{verbatim} |
436 |
|
% cd ../input |
437 |
|
% cp ../code/mitgcmuv ./ |
438 |
|
% ./mitgcmuv > output.txt |
439 |
|
\end{verbatim} |
440 |
|
or if you will be making multiple runs with the same executable: |
441 |
|
\begin{verbatim} |
442 |
|
% cd ../ |
443 |
|
% cp -r input run1 |
444 |
|
% cp code/mitgcmuv run1 |
445 |
|
% cd run1 |
446 |
|
% ./mitgcmuv > output.txt |
447 |
|
\end{verbatim} |
448 |
|
|
449 |
|
\subsubsection{Building from a new directory} |
450 |
|
|
451 |
|
Since the {\em input} directory contains input files it is often more |
452 |
|
useful to keep {\em input} pristine and build in a new directory |
453 |
|
within {\em verification/exp2/}: |
454 |
|
\begin{verbatim} |
455 |
|
% cd verification/exp2 |
456 |
|
% mkdir build |
457 |
|
% cd build |
458 |
|
% ../../../tools/genmake2 -mods=../code |
459 |
|
% make depend |
460 |
|
% make |
461 |
|
\end{verbatim} |
462 |
|
This builds the code exactly as before but this time you need to copy |
463 |
|
either the executable or the input files or both in order to run the |
464 |
|
model. For example, |
465 |
|
\begin{verbatim} |
466 |
|
% cp ../input/* ./ |
467 |
|
% ./mitgcmuv > output.txt |
468 |
|
\end{verbatim} |
469 |
|
or if you tend to make multiple runs with the same executable then |
470 |
|
running in a new directory each time might be more appropriate: |
471 |
|
\begin{verbatim} |
472 |
|
% cd ../ |
473 |
|
% mkdir run1 |
474 |
|
% cp build/mitgcmuv run1/ |
475 |
|
% cp input/* run1/ |
476 |
|
% cd run1 |
477 |
|
% ./mitgcmuv > output.txt |
478 |
|
\end{verbatim} |
479 |
|
|
480 |
|
\subsubsection{Building on a scratch disk} |
481 |
|
|
482 |
|
Model object files and output data can use up large amounts of disk |
483 |
|
space so it is often the case that you will be operating on a large |
484 |
|
scratch disk. Assuming the model source is in {\em ~/MITgcm} then the |
485 |
|
following commands will build the model in {\em /scratch/exp2-run1}: |
486 |
|
\begin{verbatim} |
487 |
|
% cd /scratch/exp2-run1 |
488 |
|
% ~/MITgcm/tools/genmake2 -rootdir=~/MITgcm \ |
489 |
|
-mods=~/MITgcm/verification/exp2/code |
490 |
|
% make depend |
491 |
|
% make |
492 |
|
\end{verbatim} |
493 |
|
To run the model here, you'll need the input files: |
494 |
|
\begin{verbatim} |
495 |
|
% cp ~/MITgcm/verification/exp2/input/* ./ |
496 |
|
% ./mitgcmuv > output.txt |
497 |
|
\end{verbatim} |
498 |
|
|
499 |
|
As before, you could build in one directory and make multiple runs of |
500 |
|
the one experiment: |
501 |
|
\begin{verbatim} |
502 |
|
% cd /scratch/exp2 |
503 |
|
% mkdir build |
504 |
|
% cd build |
505 |
|
% ~/MITgcm/tools/genmake2 -rootdir=~/MITgcm \ |
506 |
|
-mods=~/MITgcm/verification/exp2/code |
507 |
|
% make depend |
508 |
|
% make |
509 |
|
% cd ../ |
510 |
|
% cp -r ~/MITgcm/verification/exp2/input run2 |
511 |
|
% cd run2 |
512 |
|
% ./mitgcmuv > output.txt |
513 |
|
\end{verbatim} |
514 |
|
|
515 |
|
|
516 |
|
\subsection{Using \texttt{genmake2}} |
517 |
|
\label{sect:genmake} |
518 |
|
|
519 |
|
To compile the code, first use the program \texttt{genmake2} (located |
520 |
|
in the \texttt{tools} directory) to generate a Makefile. |
521 |
|
\texttt{genmake2} is a shell script written to work with all |
522 |
|
``sh''--compatible shells including bash v1, bash v2, and Bourne. |
523 |
|
%Internally, \texttt{genmake2} determines the locations of needed |
524 |
|
%files, the compiler, compiler options, libraries, and Unix tools. It |
525 |
|
%relies upon a number of ``optfiles'' located in the |
526 |
|
%\texttt{tools/build\_options} directory. |
527 |
|
\texttt{genmake2} parses information from the following sources: |
528 |
|
\begin{description} |
529 |
|
\item[-] a {\em gemake\_local} file if one is found in the current |
530 |
|
directory |
531 |
|
\item[-] command-line options |
532 |
|
\item[-] an "options file" as specified by the command-line option |
533 |
|
\texttt{--optfile=/PATH/FILENAME} |
534 |
|
\item[-] a {\em packages.conf} file (if one is found) with the |
535 |
|
specific list of packages to compile. The search path for |
536 |
|
file {\em packages.conf} is, first, the current directory and |
537 |
|
then each of the "MODS" directories in the given order (see below). |
538 |
|
\end{description} |
539 |
|
|
540 |
|
\subsubsection{Optfiles in \texttt{tools/build\_options} directory:} |
541 |
|
|
542 |
|
The purpose of the optfiles is to provide all the compilation options |
543 |
|
for particular ``platforms'' (where ``platform'' roughly means the |
544 |
|
combination of the hardware and the compiler) and code configurations. |
545 |
|
Given the combinations of possible compilers and library dependencies |
546 |
|
({\it eg.} MPI and NetCDF) there may be numerous optfiles available |
547 |
|
for a single machine. The naming scheme for the majority of the |
548 |
|
optfiles shipped with the code is |
549 |
|
\begin{center} |
550 |
|
{\bf OS\_HARDWARE\_COMPILER } |
551 |
|
\end{center} |
552 |
|
where |
553 |
|
\begin{description} |
554 |
|
\item[OS] is the name of the operating system (generally the |
555 |
|
lower-case output of the {\tt 'uname'} command) |
556 |
|
\item[HARDWARE] is a string that describes the CPU type and |
557 |
|
corresponds to output from the {\tt 'uname -m'} command: |
558 |
|
\begin{description} |
559 |
|
\item[ia32] is for ``x86'' machines such as i386, i486, i586, i686, |
560 |
|
and athlon |
561 |
|
\item[ia64] is for Intel IA64 systems (eg. Itanium, Itanium2) |
562 |
|
\item[amd64] is AMD x86\_64 systems |
563 |
|
\item[ppc] is for Mac PowerPC systems |
564 |
|
\end{description} |
565 |
|
\item[COMPILER] is the compiler name (generally, the name of the |
566 |
|
FORTRAN executable) |
567 |
|
\end{description} |
568 |
|
|
569 |
|
In many cases, the default optfiles are sufficient and will result in |
570 |
|
usable Makefiles. However, for some machines or code configurations, |
571 |
|
new ``optfiles'' must be written. To create a new optfile, it is |
572 |
|
generally best to start with one of the defaults and modify it to suit |
573 |
|
your needs. Like \texttt{genmake2}, the optfiles are all written |
574 |
|
using a simple ``sh''--compatible syntax. While nearly all variables |
575 |
|
used within \texttt{genmake2} may be specified in the optfiles, the |
576 |
|
critical ones that should be defined are: |
577 |
|
|
578 |
|
\begin{description} |
579 |
|
\item[FC] the FORTRAN compiler (executable) to use |
580 |
|
\item[DEFINES] the command-line DEFINE options passed to the compiler |
581 |
|
\item[CPP] the C pre-processor to use |
582 |
|
\item[NOOPTFLAGS] options flags for special files that should not be |
583 |
|
optimized |
584 |
|
\end{description} |
585 |
|
|
586 |
|
For example, the optfile for a typical Red Hat Linux machine (``ia32'' |
587 |
|
architecture) using the GCC (g77) compiler is |
588 |
|
\begin{verbatim} |
589 |
|
FC=g77 |
590 |
|
DEFINES='-D_BYTESWAPIO -DWORDLENGTH=4' |
591 |
|
CPP='cpp -traditional -P' |
592 |
|
NOOPTFLAGS='-O0' |
593 |
|
# For IEEE, use the "-ffloat-store" option |
594 |
|
if test "x$IEEE" = x ; then |
595 |
|
FFLAGS='-Wimplicit -Wunused -Wuninitialized' |
596 |
|
FOPTIM='-O3 -malign-double -funroll-loops' |
597 |
|
else |
598 |
|
FFLAGS='-Wimplicit -Wunused -ffloat-store' |
599 |
|
FOPTIM='-O0 -malign-double' |
600 |
|
fi |
601 |
|
\end{verbatim} |
602 |
|
|
603 |
|
If you write an optfile for an unrepresented machine or compiler, you |
604 |
|
are strongly encouraged to submit the optfile to the MITgcm project |
605 |
|
for inclusion. Please send the file to the |
606 |
|
\begin{rawhtml} <A href="mail-to:MITgcm-support@mitgcm.org"> \end{rawhtml} |
607 |
|
\begin{center} |
608 |
|
MITgcm-support@mitgcm.org |
609 |
|
\end{center} |
610 |
|
\begin{rawhtml} </A> \end{rawhtml} |
611 |
|
mailing list. |
612 |
|
|
613 |
|
\subsubsection{Command-line options:} |
614 |
|
|
615 |
|
In addition to the optfiles, \texttt{genmake2} supports a number of |
616 |
|
helpful command-line options. A complete list of these options can be |
617 |
|
obtained from: |
618 |
|
\begin{verbatim} |
619 |
|
% genmake2 -h |
620 |
|
\end{verbatim} |
621 |
|
|
622 |
|
The most important command-line options are: |
623 |
|
\begin{description} |
624 |
|
|
625 |
|
\item[\texttt{--optfile=/PATH/FILENAME}] specifies the optfile that |
626 |
|
should be used for a particular build. |
627 |
|
|
628 |
|
If no "optfile" is specified (either through the command line or the |
629 |
|
MITGCM\_OPTFILE environment variable), genmake2 will try to make a |
630 |
|
reasonable guess from the list provided in {\em |
631 |
|
tools/build\_options}. The method used for making this guess is |
632 |
|
to first determine the combination of operating system and hardware |
633 |
|
(eg. "linux\_ia32") and then find a working FORTRAN compiler within |
634 |
|
the user's path. When these three items have been identified, |
635 |
|
genmake2 will try to find an optfile that has a matching name. |
636 |
|
|
637 |
|
\item[\texttt{--mods='DIR1 DIR2 DIR3 ...'}] specifies a list of |
638 |
|
directories containing ``modifications''. These directories contain |
639 |
|
files with names that may (or may not) exist in the main MITgcm |
640 |
|
source tree but will be overridden by any identically-named sources |
641 |
|
within the ``MODS'' directories. |
642 |
|
|
643 |
|
The order of precedence for this "name-hiding" is as follows: |
644 |
|
\begin{itemize} |
645 |
|
\item ``MODS'' directories (in the order given) |
646 |
|
\item Packages either explicitly specified or provided by default |
647 |
|
(in the order given) |
648 |
|
\item Packages included due to package dependencies (in the order |
649 |
|
that that package dependencies are parsed) |
650 |
|
\item The "standard dirs" (which may have been specified by the |
651 |
|
``-standarddirs'' option) |
652 |
|
\end{itemize} |
653 |
|
|
654 |
|
\item[\texttt{--pgroups=/PATH/FILENAME}] specifies the file |
655 |
|
where package groups are defined. If not set, the package-groups |
656 |
|
definition will be read from {\em pkg/pkg\_groups}. |
657 |
|
It also contains the default list of packages (defined |
658 |
|
as the group ``{\it default\_pkg\_list}'' which is used |
659 |
|
when no specific package list ({\em packages.conf}) |
660 |
|
is found in current directory or in any "MODS" directory. |
661 |
|
|
662 |
|
\item[\texttt{--pdepend=/PATH/FILENAME}] specifies the dependency file |
663 |
|
used for packages. |
664 |
|
|
665 |
|
If not specified, the default dependency file {\em pkg/pkg\_depend} |
666 |
|
is used. The syntax for this file is parsed on a line-by-line basis |
667 |
|
where each line containes either a comment ("\#") or a simple |
668 |
|
"PKGNAME1 (+|-)PKGNAME2" pairwise rule where the "+" or "-" symbol |
669 |
|
specifies a "must be used with" or a "must not be used with" |
670 |
|
relationship, respectively. If no rule is specified, then it is |
671 |
|
assumed that the two packages are compatible and will function |
672 |
|
either with or without each other. |
673 |
|
|
674 |
|
\item[\texttt{--adof=/path/to/file}] specifies the "adjoint" or |
675 |
|
automatic differentiation options file to be used. The file is |
676 |
|
analogous to the ``optfile'' defined above but it specifies |
677 |
|
information for the AD build process. |
678 |
|
|
679 |
|
The default file is located in {\em |
680 |
|
tools/adjoint\_options/adjoint\_default} and it defines the "TAF" |
681 |
|
and "TAMC" compilers. An alternate version is also available at |
682 |
|
{\em tools/adjoint\_options/adjoint\_staf} that selects the newer |
683 |
|
"STAF" compiler. As with any compilers, it is helpful to have their |
684 |
|
directories listed in your {\tt \$PATH} environment variable. |
685 |
|
|
686 |
|
\item[\texttt{--mpi}] This option enables certain MPI features (using |
687 |
|
CPP \texttt{\#define}s) within the code and is necessary for MPI |
688 |
|
builds (see Section \ref{sect:mpi-build}). |
689 |
|
|
690 |
|
\item[\texttt{--make=/path/to/gmake}] Due to the poor handling of |
691 |
|
soft-links and other bugs common with the \texttt{make} versions |
692 |
|
provided by commercial Unix vendors, GNU \texttt{make} (sometimes |
693 |
|
called \texttt{gmake}) should be preferred. This option provides a |
694 |
|
means for specifying the make executable to be used. |
695 |
|
|
696 |
|
\item[\texttt{--bash=/path/to/sh}] On some (usually older UNIX) |
697 |
|
machines, the ``bash'' shell is unavailable. To run on these |
698 |
|
systems, \texttt{genmake2} can be invoked using an ``sh'' (that is, |
699 |
|
a Bourne, POSIX, or compatible) shell. The syntax in these |
700 |
|
circumstances is: |
701 |
|
\begin{center} |
702 |
|
\texttt{\% /bin/sh genmake2 -bash=/bin/sh [...options...]} |
703 |
|
\end{center} |
704 |
|
where \texttt{/bin/sh} can be replaced with the full path and name |
705 |
|
of the desired shell. |
706 |
|
|
707 |
|
\end{description} |
708 |
|
|
709 |
|
|
710 |
|
\subsection{Building with MPI} |
711 |
|
\label{sect:mpi-build} |
712 |
|
|
713 |
|
Building MITgcm to use MPI libraries can be complicated due to the |
714 |
|
variety of different MPI implementations available, their dependencies |
715 |
|
or interactions with different compilers, and their often ad-hoc |
716 |
|
locations within file systems. For these reasons, its generally a |
717 |
|
good idea to start by finding and reading the documentation for your |
718 |
|
machine(s) and, if necessary, seeking help from your local systems |
719 |
|
administrator. |
720 |
|
|
721 |
|
The steps for building MITgcm with MPI support are: |
722 |
|
\begin{enumerate} |
723 |
|
|
724 |
|
\item Determine the locations of your MPI-enabled compiler and/or MPI |
725 |
|
libraries and put them into an options file as described in Section |
726 |
|
\ref{sect:genmake}. One can start with one of the examples in: |
727 |
|
\begin{rawhtml} <A |
728 |
|
href="http://mitgcm.org/viewvc/MITgcm/MITgcm/tools/build_options/"> |
729 |
|
\end{rawhtml} |
730 |
|
\begin{center} |
731 |
|
\texttt{MITgcm/tools/build\_options/} |
732 |
|
\end{center} |
733 |
|
\begin{rawhtml} </A> \end{rawhtml} |
734 |
|
such as \texttt{linux\_ia32\_g77+mpi\_cg01} or |
735 |
|
\texttt{linux\_ia64\_efc+mpi} and then edit it to suit the machine at |
736 |
|
hand. You may need help from your user guide or local systems |
737 |
|
administrator to determine the exact location of the MPI libraries. |
738 |
|
If libraries are not installed, MPI implementations and related |
739 |
|
tools are available including: |
740 |
|
\begin{itemize} |
741 |
|
\item \begin{rawhtml} <A |
742 |
|
href="http://www-unix.mcs.anl.gov/mpi/mpich/"> |
743 |
|
\end{rawhtml} |
744 |
|
MPICH |
745 |
|
\begin{rawhtml} </A> \end{rawhtml} |
746 |
|
|
747 |
|
\item \begin{rawhtml} <A |
748 |
|
href="http://www.lam-mpi.org/"> |
749 |
|
\end{rawhtml} |
750 |
|
LAM/MPI |
751 |
|
\begin{rawhtml} </A> \end{rawhtml} |
752 |
|
|
753 |
|
\item \begin{rawhtml} <A |
754 |
|
href="http://www.osc.edu/~pw/mpiexec/"> |
755 |
|
\end{rawhtml} |
756 |
|
MPIexec |
757 |
|
\begin{rawhtml} </A> \end{rawhtml} |
758 |
|
\end{itemize} |
759 |
|
|
760 |
|
\item Build the code with the \texttt{genmake2} \texttt{-mpi} option |
761 |
|
(see Section \ref{sect:genmake}) using commands such as: |
762 |
|
{\footnotesize \begin{verbatim} |
763 |
|
% ../../../tools/genmake2 -mods=../code -mpi -of=YOUR_OPTFILE |
764 |
|
% make depend |
765 |
|
% make |
766 |
|
\end{verbatim} } |
767 |
|
|
768 |
|
\item Run the code with the appropriate MPI ``run'' or ``exec'' |
769 |
|
program provided with your particular implementation of MPI. |
770 |
|
Typical MPI packages such as MPICH will use something like: |
771 |
|
\begin{verbatim} |
772 |
|
% mpirun -np 4 -machinefile mf ./mitgcmuv |
773 |
|
\end{verbatim} |
774 |
|
Sightly more complicated scripts may be needed for many machines |
775 |
|
since execution of the code may be controlled by both the MPI |
776 |
|
library and a job scheduling and queueing system such as PBS, |
777 |
|
LoadLeveller, Condor, or any of a number of similar tools. A few |
778 |
|
example scripts (those used for our \begin{rawhtml} <A |
779 |
|
href="http://mitgcm.org/public/testing.html"> \end{rawhtml}regular |
780 |
|
verification runs\begin{rawhtml} </A> \end{rawhtml}) are available |
781 |
|
at: |
782 |
|
\begin{rawhtml} <A |
783 |
|
href="http://mitgcm.org/viewvc/MITgcm/MITgcm/tools/example_scripts/"> |
784 |
|
\end{rawhtml} |
785 |
|
{\footnotesize \tt |
786 |
|
http://mitgcm.org/viewvc/MITgcm/MITgcm/tools/example\_scripts/ } |
787 |
|
\begin{rawhtml} </A> \end{rawhtml} |
788 |
|
or at: |
789 |
|
\begin{rawhtml} <A |
790 |
|
href="http://mitgcm.org/viewvc/MITgcm/MITgcm_contrib/test_scripts/"> |
791 |
|
\end{rawhtml} |
792 |
|
{\footnotesize \tt |
793 |
|
http://mitgcm.org/viewvc/MITgcm/MITgcm\_contrib/test\_scripts/ } |
794 |
|
\begin{rawhtml} </A> \end{rawhtml} |
795 |
|
|
796 |
|
\end{enumerate} |
797 |
|
|
798 |
|
An example of the above process on the MITgcm cluster (``cg01'') using |
799 |
|
the GNU g77 compiler and the mpich MPI library is: |
800 |
|
|
801 |
|
{\footnotesize \begin{verbatim} |
802 |
|
% cd MITgcm/verification/exp5 |
803 |
|
% mkdir build |
804 |
|
% cd build |
805 |
|
% ../../../tools/genmake2 -mpi -mods=../code \ |
806 |
|
-of=../../../tools/build_options/linux_ia32_g77+mpi_cg01 |
807 |
|
% make depend |
808 |
|
% make |
809 |
|
% cd ../input |
810 |
|
% /usr/local/pkg/mpi/mpi-1.2.4..8a-gm-1.5/g77/bin/mpirun.ch_gm \ |
811 |
|
-machinefile mf --gm-kill 5 -v -np 2 ../build/mitgcmuv |
812 |
|
\end{verbatim} } |
813 |
|
|
814 |
\section[Running MITgcm]{Running the model in prognostic mode} |
\section[Running MITgcm]{Running the model in prognostic mode} |
815 |
\label{sect:runModel} |
\label{sect:runModel} |
909 |
used to restart the model but are overwritten every other time they are |
used to restart the model but are overwritten every other time they are |
910 |
output to save disk space during long integrations. |
output to save disk space during long integrations. |
911 |
|
|
|
|
|
|
|
|
912 |
\subsubsection{MNC output files} |
\subsubsection{MNC output files} |
913 |
|
|
914 |
Unlike the \texttt{mdsio} output, the \texttt{mnc}--generated output |
Unlike the \texttt{mdsio} output, the \texttt{mnc}--generated output |
915 |
is usually (though not necessarily) placed within a subdirectory with |
is usually (though not necessarily) placed within a subdirectory with |
916 |
a name such as \texttt{mnc\_test\_\${DATE}\_\${SEQ}}. The files |
a name such as \texttt{mnc\_test\_\${DATE}\_\${SEQ}}. |
|
within this subdirectory are all in the ``self-describing'' netCDF |
|
|
format and can thus be browsed and/or plotted using tools such as: |
|
|
\begin{itemize} |
|
|
\item \texttt{ncdump} is a utility which is typically included |
|
|
with every netCDF install: |
|
|
\begin{rawhtml} <A href="http://www.unidata.ucar.edu/packages/netcdf/"> \end{rawhtml} |
|
|
\begin{verbatim} |
|
|
http://www.unidata.ucar.edu/packages/netcdf/ |
|
|
\end{verbatim} |
|
|
\begin{rawhtml} </A> \end{rawhtml} and it converts the netCDF |
|
|
binaries into formatted ASCII text files. |
|
|
|
|
|
\item \texttt{ncview} utility is a very convenient and quick way |
|
|
to plot netCDF data and it runs on most OSes: |
|
|
\begin{rawhtml} <A href="http://meteora.ucsd.edu/~pierce/ncview_home_page.html"> \end{rawhtml} |
|
|
\begin{verbatim} |
|
|
http://meteora.ucsd.edu/~pierce/ncview_home_page.html |
|
|
\end{verbatim} |
|
|
\begin{rawhtml} </A> \end{rawhtml} |
|
|
|
|
|
\item MatLAB(c) and other common post-processing environments provide |
|
|
various netCDF interfaces including: |
|
|
\begin{rawhtml} <A href="http://mexcdf.sourceforge.net/"> \end{rawhtml} |
|
|
\begin{verbatim} |
|
|
http://mexcdf.sourceforge.net/ |
|
|
\end{verbatim} |
|
|
\begin{rawhtml} </A> \end{rawhtml} |
|
|
\begin{rawhtml} <A href="http://woodshole.er.usgs.gov/staffpages/cdenham/public_html/MexCDF/nc4ml5.html"> \end{rawhtml} |
|
|
\begin{verbatim} |
|
|
http://woodshole.er.usgs.gov/staffpages/cdenham/public_html/MexCDF/nc4ml5.html |
|
|
\end{verbatim} |
|
|
\begin{rawhtml} </A> \end{rawhtml} |
|
|
\end{itemize} |
|
|
|
|
917 |
|
|
918 |
\subsection{Looking at the output} |
\subsection{Looking at the output} |
919 |
|
|
949 |
Similar scripts for netCDF output (\texttt{rdmnc.m}) are available and |
Similar scripts for netCDF output (\texttt{rdmnc.m}) are available and |
950 |
they are described in Section \ref{sec:pkg:mnc}. |
they are described in Section \ref{sec:pkg:mnc}. |
951 |
|
|
952 |
|
The MNC output files are all in the ``self-describing'' netCDF |
953 |
|
format and can thus be browsed and/or plotted using tools such as: |
954 |
|
\begin{itemize} |
955 |
|
\item \texttt{ncdump} is a utility which is typically included |
956 |
|
with every netCDF install: |
957 |
|
\begin{rawhtml} <A href="http://www.unidata.ucar.edu/packages/netcdf/"> \end{rawhtml} |
958 |
|
\begin{verbatim} |
959 |
|
http://www.unidata.ucar.edu/packages/netcdf/ |
960 |
|
\end{verbatim} |
961 |
|
\begin{rawhtml} </A> \end{rawhtml} and it converts the netCDF |
962 |
|
binaries into formatted ASCII text files. |
963 |
|
|
964 |
|
\item \texttt{ncview} utility is a very convenient and quick way |
965 |
|
to plot netCDF data and it runs on most OSes: |
966 |
|
\begin{rawhtml} <A href="http://meteora.ucsd.edu/~pierce/ncview_home_page.html"> \end{rawhtml} |
967 |
|
\begin{verbatim} |
968 |
|
http://meteora.ucsd.edu/~pierce/ncview_home_page.html |
969 |
|
\end{verbatim} |
970 |
|
\begin{rawhtml} </A> \end{rawhtml} |
971 |
|
|
972 |
|
\item MatLAB(c) and other common post-processing environments provide |
973 |
|
various netCDF interfaces including: |
974 |
|
\begin{rawhtml} <A href="http://mexcdf.sourceforge.net/"> \end{rawhtml} |
975 |
|
\begin{verbatim} |
976 |
|
http://mexcdf.sourceforge.net/ |
977 |
|
\end{verbatim} |
978 |
|
\begin{rawhtml} </A> \end{rawhtml} |
979 |
|
\begin{rawhtml} <A href="http://woodshole.er.usgs.gov/staffpages/cdenham/public_html/MexCDF/nc4ml5.html"> \end{rawhtml} |
980 |
|
\begin{verbatim} |
981 |
|
http://woodshole.er.usgs.gov/staffpages/cdenham/public_html/MexCDF/nc4ml5.html |
982 |
|
\end{verbatim} |
983 |
|
\begin{rawhtml} </A> \end{rawhtml} |
984 |
|
\end{itemize} |
985 |
|
|