Compiling MATLAB with Release 14 SP2

Compiling MATLAB under MATLAB R14 SP2

Here are the steps for MATLAB compilation
Why create MATLAB executables?
MATLAB Compiler Setup at PSU
MATLAB Compiler Setup at LSC grids
Testing your MATLAB compiler setup
After this, you will want to find out how to deploy MATLAB executables on grid computers.

Why compile MATLAB into executables?

MATLAB has become quite popular for interactively analyzing data and doing time-series analysis within the LSC. It would be even better if one could then use the same programs and functions to do large-scale batch processing, instead of having to translate into a different language! (It would certainly eliminate an extra code walkthrough review)

But you can! MATLAB now has a compiler which can be used to make stand-alone executables which have been run on LSC Grid computers for the S2, S3 and S4 data analysis (Burst Group r-statistic, BlockNormal, Stochastic Group).

While you do need to purchase a MATLAB Compiler license to build the executables, you can then freely distribute the executable and the MATLAB Component Runtime (MCR) to as many computers as you want! No licenses, no limits.

Recently, the 'matapps' software directory has been created in the DASWG 'lscsoft' CVS repository to house MATLAB applications developed by the LSC community.

MATLAB Compiler 4.2

The new release of MATLAB (called both MATLAB 7.2 and Release 14 SP2) comes with a new Compiler 4.2. The documentation for the compiler is available online from Mathworks.

This uses the MATLAB Component Runtime (MCR) which allows the use of nearly every function in the MATLAB libraries in a stand-alone executable. This means that compilable variants of MATLAB functions (required for MATLAB 6) are no longer needed! In particular, MATLAB classes and objects can be used.

Unfortunately, because the MCR runs M-files instead of compiling an executable, compiled code is now no faster than interactive code.

The MCR requires a different method for distributing executables. This is most important when exporting stand-alone programs to the LSC Grid computers

Getting the MATLAB Compiler

It is important to compile MATLAB programs on the same operating system type where the executables will be run. As the LSC Grid utilities all use Linux at present, you need to run the compiler on a Linux system.
The license can be purchased for an existing MATLAB installation from MathWorks. The cost for an academic license for the compiler is ~ $200. If you are using the Penn State or LSC clusters for development, the license and installation are already provided.
In addition, you should also generate the executable on the same Linux/gcc version as you will run it on. The PSU Gravity cluster (gravity.psu.edu) and Pleiades grid computer (ligo-grid.aset.psu.edu) presently use RedHat Enterprise Linux 3 (termed RHEL3). RHEL3 uses Linux kernel 2.4 and gcc 3.2.3. The LSC grid computers at Caltech, Hanford and Livingston (ldas-grid.ligo.caltech.edu, ldas-grid.ligo-wa.caltech.edu and ldas-grid.ligo-la.caltech.edu) use Fedora Core 3 (termed FC3). FC3 uses Linux kernel 2.6 and gcc 3.4.3. As MATLAB Release 14 was built on the older gcc 3.2.3, some care has to be taken to build and run executables under the newer gcc 3.4.3.
Note that Compiler 4.2 has corrected problems with the initial compiler release. This has simplified the installation and setup from Compiler 4.0.

Compiling MATLAB on PSU Gravity Cluster

MATLAB Release 14 is installed on the gravity cluster at /usr/local/matlab-r14sp2. It is also now the default installation, so entering

matlab

at the command prompt will start that version in interactive mode.

1) Create build options file

As detailed in the Mathworks documentation, after installing the compiler each user should make a build options script (mbuildopts.sh). On PSU Gravity cluster, do the following at the command line:

/usr/local/matlab-r14sp2/bin/matlab -nodisplay

(NOTE: I have used the -nodisplay option to allow non-X-Windows operation, but that is not necessary)
At the MATLAB prompt, enter the following:

>> mbuild -setup

Choose the default option (if asked) and allow it to override the existing mbuildopts.sh.
This creates a file mbuildopts.sh in a .matlab/R14 sub-directory of your home directory. Here is a log file of the mbuild process.

2) Modify the build options file

There is still an error in the options script which needs to be fixed. It involves adding gcc_s as a library.
1) Open mbuildopts.sh in an editor
2) On or about line 84, add -lgcc_s to the CLIBS line for 'gcc' under 'glnx86' as below:

#----------------------------------------------------------------------------
            ;;
        glnx86)
#----------------------------------------------------------------------------
            RPATH="-Wl,--rpath-link,$TMW_ROOT/bin/$Arch"
#           gcc -v
#           gcc version 3.2.3
            CC='gcc'
            CFLAGS="$MFLAGS -ansi -D_GNU_SOURCE -pthread -fexceptions"
            CLIBS="$RPATH $MLIBS -lm -lstdc++ -lgcc_s"

3) Save the modified mbuildopts.sh

Here is a reference mbuildopts.sh file with the gcc_s change .

3) Fix your startup.m file to running executables

When testing a stand-alone build, you need to make sure it sees the local encrypted m-files instead of ones from your CVS sandbox. We wrap any addpath commands (which let us see the CVS sandbox files) with a conditional to prevent their use when running an executable. The syntax of the startup.m file should be:

if((str2num(version('-release')) < 14) || ~isdeployed)
   addpath commands
   ...
end

Here is an example startup.m for the Gravity cluster and the psurg CVS.

The startup.m file should be in your ~/matlab directory.

4) Add MATLAB R14SP2 run-time to your path

To test the stand-alone executables on the machine you have MATLAB installed on, you need to have the MATLAB Component Runtime in your dynamic library path. This is detailed in on page 5-4 of the Compiler manual. It is similar to Compiler 4.2, with an updated Java run-time

On the Gravity cluster add the following to your .cshrc

setenv MATLAB_ROOT /usr/local/matlab-r14sp2
if ( $?LD_LIBRARY_PATH ) then
    setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/java/jre/glnx86/jre1.5.0/lib/i386:${LD_LIBRARY_PATH}
else
    setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/java/jre/glnx86/jre1.5.0/lib/i386
endif
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/java/jre/glnx86/jre1.5.0/lib/i386/client:${LD_LIBRARY_PATH}
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/java/jre/glnx86/jre1.5.0/lib/i386/native_threads:${LD_LIBRARY_PATH}
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/opengl/lib/glnx86:${LD_LIBRARY_PATH}
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/bin/glnx86:${LD_LIBRARY_PATH}
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/os/glnx86:${LD_LIBRARY_PATH}
setenv XAPPLRESDIR ${MATLAB_ROOT}/X11/app-defaults
unsetenv MATLAB

Note that this is in 'reverse' order from the Compiler list to ensure correct order in LD_LIBRARY_PATH. Also, I have added a library (/sys/opengl/lib/glnx86) which previous experience with Compiler 4.2 showed was often necessary. I have provided an example csh script which handles a non-existent LD_LIBRARY_PATH and checks if MATLAB was already added to it.

** NOTE: Do NOT define an environment variable named "MATLAB" under MATLAB R14. This will screw up the compiler in quite subtle ways.

Compiling MATLAB on LSC Development Nodes

The LSC grid clusters have nodes specifically set aside for development work, as opposed to running jobs on the grid nodes. These have the ldas-pcdev1 name (i.e. ldas-pcdev1.ligo.caltech.edu). At present, the CIT and LLO nodes are at Fedora Core 3. As of May 2nd, 2005, the LHO development node was still at Redhat 9.
MATLAB Release 14 is installed in the same place (/ldcg/matlab_r14_sp2) on these machines. Also note that these development computers see the same drives as the grid computers at that site, which simplifies deployment of MATLAB executables.

1) Create build options file

As detailed in the Mathworks documentation, after installing the compiler each user should make a build options script (mbuildopts.sh). On PSU Gravity cluster, do the following at the command line:

/ldcg/matlab_r14_sp2/bin/matlab -nodisplay

(NOTE: I have used the -nodisplay option to allow non-X-Windows operation, but that is not necessary)
At the MATLAB prompt, enter the following:

>> mbuild -setup

2) Check the build options file

The LSC grid administrators have already modified the default MATLAB build options file to make it compatible with Fedora Core 3 and gcc 3.4.3. If you wish, you can verify that the build options file has the correct change. The change specifies the libstdc++ library used to build MATLAB
1) Open mbuildopts.sh in an editor
2) On or about line 84, verify that -lstdc++ has been modified to point within the MATLAB release files as below:

#----------------------------------------------------------------------------
            ;;
        glnx86)
#----------------------------------------------------------------------------
            RPATH="-Wl,--rpath-link,$TMW_ROOT/bin/$Arch"
#           gcc -v
#           gcc version 3.2.3
            CC='gcc'
            CFLAGS="$MFLAGS -ansi -D_GNU_SOURCE -pthread -fexceptions"
            CLIBS="$RPATH $MLIBS -lm $TMW_ROOT/sys/os/$Arch/libstdc++.so.5""

Here is a reference mbuildopts.sh file with the libstdc++ change.

3) Fix your startup.m file to running executables

if((str2num(version('-release')) < 14) || ~isdeployed)
   addpath commands
   ...
end

The startup.m file should be in your ~/matlab directory.

4) Add MATLAB R14SP2 run-time to your path

setenv MATLAB_ROOT /ldcg/matlab_r14_sp2
if ( $?LD_LIBRARY_PATH ) then
    setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/java/jre/glnx86/jre1.5.0/lib/i386:${LD_LIBRARY_PATH}
else
    setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/java/jre/glnx86/jre1.5.0/lib/i386
endif
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/java/jre/glnx86/jre1.5.0/lib/i386/client:${LD_LIBRARY_PATH}
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/java/jre/glnx86/jre1.5.0/lib/i386/native_threads:${LD_LIBRARY_PATH}
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/opengl/lib/glnx86:${LD_LIBRARY_PATH}
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/bin/glnx86:${LD_LIBRARY_PATH}
setenv LD_LIBRARY_PATH ${MATLAB_ROOT}/sys/os/glnx86:${LD_LIBRARY_PATH}
setenv XAPPLRESDIR ${MATLAB_ROOT}/X11/app-defaults
unsetenv MATLAB

Testing your compiler setup

We will now follow the stand-alone application example (page 5-3 of the Compiler manual).

1) Get the magicsquare.m sample code

This is located in <matlabroot>/extern/examples/compiler, but I have also made it available following this magicsquare.m link.

Create a directory (~/mat_test) and copy the file to it.

2) Build the application

Here we start up MATLAB and invoke the mcc compiler. Use the following at the command prompt (assumes 'matlab' points to the MATLAB R14SP2 executable)
$

cd ~/mat_test
matlab -nodisplay
>>mcc -mv magicsquare.m

This will create the following files
magicsquare --> the executable
magicsquare_main.c --> 'main' wrapper
magicsquare_mcc_component_data.c
magicsquare.ctf --> encoded matlab files
magicsquare_mcr --> directory of m-files for application

When the compilation is progressing, a long blow-by-blow listing is shown because we added the -v option (for verbose). Once things work, you can skip this option. Here is a copy of the compilation log.

Warning: when you first run the executable it will unpack the .ctf file into an _mcr directory. Having this (or any _mcr) directory in your matlab path may cause problems when trying to compile! This is probably your problem if you get an error:

Error: File: dirname Line: 1 Column: 25

in which case, you will want to check the verbose listing to see what _mcr directory is in your path.

3) Test the application

Here we exit MATLAB, and run the application from the command line:

>> quit
[thorne@auriga ~/mat_test]$ ./magicsquare 4
Extracting CTF archive. This may take a few seconds, depending on the
 size of your application. Please wait...
...CTF archive extraction complete.

ans =

    16     2     3    13
     5    11    10     8
     9     7     6    12
     4    14    15     1

[thorne@auriga ~/mat_test]$

Handling function inputs on the command line

Any function inputs are passed as command line arguments. One difficulty is that numbers are passed as doubles within MATLAB, but as text strings when entered at the command line. A way to handle this is shown in the example code

function magicsquare(n)
%MAGICSQUARE generates a magic square matrix of the size specified
% by the input parameter n.

% Copyright 2003 The MathWorks, Inc.
if (ischar(n))
  n=str2num(n);
end
magic(n)

The ischar(n) test checks that if text is encountered for a numeric input, it is converted from a string to a numeric equivalent. This will even work for MATLAB constructs like [1:5].

Keith Thorne

Last modified: Mon May 2 2005