ViewVC logotype

Annotation of /trunk/doc/user/execute.tex

Parent Directory Parent Directory | Revision Log Revision Log

Revision 2375 - (hide annotations)
Tue Apr 7 04:41:28 2009 UTC (10 years, 5 months ago) by gross
File MIME type: application/x-tex
File size: 10439 byte(s)
MPI hints added
1 gross 2316 \chapter{Execution of an {\it escript} Script}
2     \label{EXECUTION}
4     \section{Overview}
5 jfenwick 2331 A typical way of starting your {\it escript} script \file{myscript.py} is with the \program{escript} command\footnote{The \program{escript} launcher is not supported under \WINDOWS yet.}:
6 gross 2316 \begin{verbatim}
7     escript myscript.py
8     \end{verbatim}
9 jfenwick 2331 as already shown in section~\ref{FirstSteps}\footnote{For this discussion, it is assumed that \program{escript} is included in your \env{PATH} environment. See installation guide for details.}
10     . In some cases
11     it can be useful to work interactively e.g. when debugging a script, with the command
12 gross 2316 \begin{verbatim}
13     escript -i myscript.py
14     \end{verbatim}
15 jfenwick 2331 This will execute \var{myscript.py} and when it completes (or an error occurs), a \PYTHON prompt will be provided.
16     To leave the prompt press \kbd{Control-d}.
17 gross 2316
18     To start
19     \program{escript} using four threads (eg. if you use a multi-core processor) you can use
20     \begin{verbatim}
21     escript -t 4 myscript.py
22     \end{verbatim}
23 jfenwick 2331 This will require {\it escript} to be compiled for \OPENMP\cite{OPENMP}.
24 gross 2316
25 jfenwick 2331 To start \program{escript} using \MPI\cite{MPI} with $8$ processes you use
26 gross 2316 \begin{verbatim}
27 jfenwick 2331 escript -p 8 myscript.py
28 gross 2316 \end{verbatim}
29 gross 2375 If the processors which are used are multi--core processors or multi--processor shared memory architectures you can use threading in addition to \MPI. For instance to run $8$ \MPI processes with using $4$ threads each, you use the command
30 gross 2316 \begin{verbatim}
31 jfenwick 2331 escript -p 8 -t 4 myscript.py
32 gross 2316 \end{verbatim}
33 jfenwick 2331 In the case of a super computer or a cluster, you may wish to distribute the workload over a number of nodes\footnote{For simplicity, we will use the term node to refer to either a node in a super computer or an individual machine in a cluster}.
34     For example, to use $8$ nodes, with $4$ \MPI processes per node, write
35 gross 2316 \begin{verbatim}
36 jfenwick 2331 escript -n 8 -p 4 myscript.py
37 gross 2316 \end{verbatim}
38 jfenwick 2331 Since threading has some performance advantages over processes, you may specify a number of threads as well.
39     \begin{verbatim}
40     escript -n 8 -p 4 -t 2 myscript.py
41     \end{verbatim}
42     This runs the script on $8$ nodes, with $4$ processes per node and $2$ threads per process.
43 gross 2316
44     \section{Options}
45     The general form of the \program{escript} launcher is as follows:
47 jfenwick 2345 %%%%
48     % If you are thinking about changing this please remember to update the man page as well
49     %%%%
51 gross 2316 \program{escript}
52     \optional{\programopt{-n \var{nn}}}
53 jfenwick 2331 \optional{\programopt{-p \var{np}}}
54 gross 2316 \optional{\programopt{-t \var{nt}}}
55     \optional{\programopt{-f \var{hostfile}}}
56     \optional{\programopt{-x}}
57     \optional{\programopt{-V}}
58     \optional{\programopt{-e}}
59     \optional{\programopt{-h}}
60     \optional{\programopt{-v}}
61     \optional{\programopt{-o}}
62     \optional{\programopt{-c}}
63     \optional{\programopt{-i}}
64 jfenwick 2343 \optional{\programopt{-b}}
65 gross 2316 \optional{\var{file}}
66     \optional{\var{ARGS}}
68 jfenwick 2331 where \var{file} is the name of a script, \var{ARGS} are arguments for the script.
69     The \program{escript} program will import your current environment variables.
70     If no \var{file} is given, then you will be given a \PYTHON prompt (see \programopt{-i} for restrictions).
72 gross 2316 The options are used as follows:
73     \begin{itemize}
75     \item[\programopt{-n} \var{nn}] the number of compute nodes \var{nn} to be used. The total number of process being used is
76 gross 2363 $\var{nn} \cdot \var{ns}$. This option overwrites the value of the \env{ESCRIPT_NUM_NODES} environment variable.
77     If a hostfile is given, the number of nodes needs to match the number hosts given in the host file.
78     If $\var{nn}>1$ but {\it escript} is not compiled for \MPI a warning is printed but execution is continued with $\var{nn}=1$. If \programopt{-n} is not set the
79 gross 2350 number of hosts in the host file is used. The default value is 1.
80 gross 2316
81 jfenwick 2345 \item[\programopt{-p} \var{np}] the number of MPI processes per node. The total number of processes to be used is
82 gross 2370 $\var{nn} \cdot \var{np}$. This option overwrites the value of the \env{ESCRIPT_NUM_PROCS} environment variable. If $\var{np}>1$ but {\it escript} is not compiled for \MPI a warning is printed but execution is continued with $\var{np}=1$. The default value is 1.
83 gross 2316
84 jfenwick 2331 \item[\programopt{-t} \var{nt}] the number of threads used per processes.
85     The option overwrites the value of the \env{ESCRIPT_NUM_THREADS} environment variable.
86     If $\var{nt}>1$ but {\it escript} is not compiled for \OPENMP a warning is printed but execution is continued with $\var{nt}=1$. The default value is 1.
87 gross 2316
88 gross 2350 \item[\programopt{-f} \var{hostfile}] the name of a file with a list of host names. Some systems require to specify the addresses or names of the compute nodes where \MPI process should be spawned. The list of addresses or names of the compute nodes is listed in the file with the name \var{hostfile}. If \programopt{-n} is set the
89     the number of different
90 gross 2375 hosts defined in \var{hostfile} must be equal to the number of requested compute nodes \var{nn}. The option overwrites the value of the \env{ESCRIPT_HOSTFILE} environment variable. By default value no host file is used.
91 gross 2316 \item[\programopt{-c}] prints the information about the settings used to compile {\it escript} and stops execution..
92     \item[\programopt{-V}] prints the version of {\it escript} and stops execution.
93     \item[\programopt{-h}] prints a help message and stops execution.
94 gross 2375 \item[\programopt{-i}] executes the script \var{file} and switches to interactive mode after the execution is finished or an exception has occurred. This option is useful for debugging a script. The option cannot be used if more then one process ($\var{nn} \cdot \var{np}>1$) is used.
95 jfenwick 2343 \item[\programopt{-b}] do not invoke python. This is used to run non-python programs.
96 gross 2316
97 gross 2375 \item[\programopt{-e}] shows additional environment variables and commands used during \program{escript} execution. This option is useful if users wish to execute scripts without using the \program{escript} command.
98 gross 2316
99 gross 2355 \item[\programopt{-o}] switches on the redirection of output of processors with \MPI rank greater than zero to the files \file{stdout_\var{r}.out} and \file{stderr_\var{r}.out} where \var{r} is the rank of the processor. The option overwrites the value of the \env{ESCRIPT_STDFILES} environment variable
100 gross 2316
101 jfenwick 2331 % \item[\programopt{-x}] interpret \var{file} as an \esysxml \footnote{{\it esysxml} has not been released yet.} task.
102 gross 2375 % This option is still experimental.
103 gross 2316
104 gross 2375 \item[\programopt{-v}] prints some diagnostic information.
105 gross 2316 \end{itemize}
106 gross 2370 \subsection{Notes}
107     \begin{itemize}
108     \item Make sure that \program{mpiexec} is in your \env{PATH}.
109     \item For MPICH and INTELMPI and for the case a hostfile is present
110     \program{escript} will start the \program{mpd} demon before execution.
111     \end{itemize}
112 gross 2316
113     \section{Input and Output}
114 jfenwick 2331 When \MPI is used on more than one process ($\var{nn} \cdot \var{np} >1$) no input from the standard input is accepted. Standard output on any process other the the master process (\var{rank}=0) will not be available.
115 gross 2375 Error output from any processor will be redirected to the node where \program{escript} has been invoked.
116 gross 2355 If the \programopt{-o} or \env{ESCRIPT_STDFILES} is set\footnote{That is, it has a non-empty value.}, then the standard and error output from any process other than the master process will be written to files of the names \file{stdout_\var{r}.out} and \file{stderr_\var{r}.out} (where
117 jfenwick 2331 \var{r} is the rank of the process).
118 gross 2316
119 jfenwick 2331 If files are created or read by individual \MPI processes with information local to the process (e.g in the \function{dump} function) and more than one process is used ($\var{nn} \cdot \var{np} >1$), the \MPI process rank is appended to the file names.
120     This will avoid problems if processes are using a shared file system.
121     Files which collect data which are global for all \MPI processors will created by the process with \MPI rank 0 only.
122     Users should keep in mind that if the file system is not shared, then a file containing global information
123 gross 2316 which is read by all processors needs to be copied to the local file system before \program{escript} is invoked.
126 gross 2375 \section{Hints for MPI Programming}
127     In general a script based on the \escript module does not require modifications when running under \MPI. However, one needs to be careful if other modules are used.
129     When \MPI is used on more than one process ($\var{nn} \cdot \var{np} >1$) the user needs to keep in mind that several copies of his script are executed at the same time
130     \footnote{In case of OpenMP only one copy is running but \escript temporarily spawns threads.} while data exchange is performed through the \escript module. At any time,
131     \escript assumes that an argument of the type \var{int}, \var{float}, \var{str}
132     and \numarray has an identical value across all processors. All
133     values of these types returned by \escript have the same value on all processors.
134     If values produced by other modules are used as argument the user has to make sure that
135     the value is identical on all processors. For instance, the usage of a random number
136     generator to create a value for argument bears the risk that the value may depend on
137     the processor.
139     An other case which needs special attention is the usage of files. When reading data
140     from a file it advisable to use the \var{'r'} for readable when opened. Keep in mind that
141     several scripts will simultaneously access the file. If data are written to a file
142     only one processor must open the file for writing. The function \function{getMPIRankWorld}
143     which returns the processor id between $0$ and the number of processors
144     helps to achieve this. The following script writes to the file
145     \var{'test.txt'} on the processor with id $0$ only:
146     \begin{python}
147     from esys.escript import *
148     if getMPIRankWorld() == 0 :
149     f = open('test.txt', 'w')
150     f.write('test message')
151     f.close()
152     \end{python}
153     Another technique is to extend the file name by the processor id to avoid conflicts while
154     writing into a shared file system:
155     \begin{python}
156     from esys.escript import *
157     f = open('test.txt.%s'%getMPIRankWorld(), 'w')
158     f.write('test message')
159     f.close()
160     \end{python}
161     creating files with names
162     \var{test.txt.0},
163     \var{test.txt.1},
164     \var{test.txt.2},
165     $\ldots$.
167     If there is the situation that if execution on one of the processors is throwing an exception,
168     for instance as opening a file for writing fails, are not made aware of this as \MPI
169     is not handling exceptions. However, \MPI will terminate the other processes but
170     may not inform the user of the reason in the obvious way. The user needs to inspect the
171     error output files to identify the exception.

  ViewVC Help
Powered by ViewVC 1.1.26