==ISIS LET control machine==

The PC came prepared for Libisis and Horace. [http://download.nexusformat.org/kits/ NeXus], [http://www.mathworks.co.uk/products/matlab/index.html Matlab] and [http://software.intel.com/en-us/articles/intel-compilers/ Intel Fortran] compiler are installed. The task is to identify the locations of all parts, necessary for instillation and run the installation script resolving all dependencies.

'''1)''' Verify if nexus is installed
$ locate libNeXus
...
/usr/lib/libNeXus.so
...

'''2)''' Check if ifort works form command line:
$ifort --version
ifort (IFORT) 11.1 20100414
Copyright (C) 1985-2010 Intel Corporation. All rights reserved.

'''3)''' Identify where Matlab is really installed and if it has all files necessary to use it.
$what matlab:
/usr/local/bin/matlab
$
$ls -l /usr/local/bin
...
lrwxrwxrwx 1 root root 38 2011-03-18 11:48 matlab -> /usr/local/MATLAB/Matlab2010bSP1/bin/matlab
...

'''/usr/local/MATLAB/Matlab2010bSP1 is in our case or $MATLAB_ROOT folder which we will use everywhere, referring to matlab'''

'''4)''' Verify if ''mwdebug.c'' file is present:
$
$ls $MATLAB_ROOT/extern
'''examples include lib'''

'''4a)''' ''etc'' folder is not there, so we have to make it and place ''mwdebug.c'' there.
$
$sudo mkdir $MATLAB_ROOT/extern/etc
$sudo mv ~/Downdloads/mwdebug.c $MATLAB_ROOT/extern/etc/

'''5)''' Download libisis from the repository
$svn co https://svn.libisis.org/code/trunk Libisis

'''6)''' Use autotools to build libisis:
$cd Libisis
$ sh autogen.sh
...
A lot of messages, no errors
...

$./configure --with-matlabroot=/usr/local/MATLAB/Matlab2010bSP1
...
A Lot of messages, no errors
...

$make
...
A lot of messages no errors
...

'''6)''' Check if libisis has been build successfully.
$ls -l ~/Libisis/Libisis/DLL/_glnxa64
...
-rwxr-xr-x 1 User Group 36434653 2011-09-14 19:09 libisisexc.mexa64
...

I assume 64 bit linux here. If it is 32-bit linux, check ''~/Libisis/Libisis/DLL/_glnxa'' folder;

'''7)''' Building binary part of libisis is now completed. Start Matlab and check if it was successful.
$matlab&
in matlab:
>>cd ~/Libisis/Libisis/DLL/_glnxa64
>> aa=libisisexc();
>> aa =
$Revision:: 1808 $ ($Date:: 2011-03-29 21:43:21 +0000 (Tue, 29 Mar 2011) $)

Follow Windows Maltab installation steps, described on [http://www.libisis.org/Download_and_setup_LIBISIS#Download_LIBISIS Libisis page].

==University of Colorado Group==

Special thanks to '''Dan Parshall'''; dan.parshall_AT_colorado.edu, who shared his experience with us.

This took a while to get back to, but eventually we got this sorted out.
It was a combination of legacy files, compilers, permissions issues, and
typos. I'm posting the solution here in order to help future Ubuntu
users.

'''1) >>> Get legacy file mwdebug.c in place <<<'''

The file mwdebug.c (which one can get from Alex or Freddie) must be placed into matlabroot/extern/src/ (one needs to make the src
directory). Do the standard:

chown -R root matlabroot/*
chmod -R 755 matlabroot/*

'''2) >>> set up IFORT <<<'''

A 30-day trial is free, academic licenses fairly cheap (I've mucked about a bit, and gfortran seems to also be functional, but as a rule it's 20% slower). In order to invoke IFORT on the command line, one has to use the "source" command and link to the "compilervars.sh" script.

Because one can't use "source" in combination with "sudo", one must change to the
root account, then complete the remaining steps:
cd /path/to/libisis
sudo su
source /path/to/ifort/compilervars.sh intel64

'''3) >>> run autogen.sh / configure / make (as root) <<<'''

One must set matlabroot directly- simlinks aren't followed properly
bash autogen.sh
./configure --with-f77=ifort --with-f90=ifort --with-matlabroot=/path
make

'''4) >>> set ownership and permissions of files made during compilation <<'''
chown -R <username> *
chmod -R 777 *

==Jülich Centre for Neutron Science==

Special thanks to '''Johannes Reim''', j.reim_AT_fz-juelich.de; who shared his experience with us;

And afterwards I have to say, actually it is not that complex, if you
know a bit about matlab, linux and how horace and libisis works but
well, neither of that applied to me *g*.

Okay what did I do for the installation:

1. Installation of libisis, I actually did not do that myself, but the
IT did it, because it should be installed for all users within our
institute, but I sent them the instructions found on the subversion
server of Libisis:
http://svn.libisis.org/viewvc/LIBISISCode/trunk/Linux/README.linux?revision=1577

I will call the libisis directory $1 (think of $1as the variable for the
full path to the directory)

2. check if there is a matlab folder within the libisis folder (we will
need that later on)

3. download the up-to-date version of horace (not the one which is
specified for linux)

4. Just extract it to your directory of choice, I will call the
directory $2 (like $1)

5. download the homer-standalone from:
http://download.libisis.org/kits/homer_standalone.zip

6. create a folder called InstrumentFiles, my suggestion would be
creating it next to the Horace directory.

7. extract the content of the instrument_setup folder within into the
created directory, I will call the directory $3 (like $1)

8. if not existing, create a lib directory within your home directory,
and into that a matlab directory, access the matlab directory

9. link the Horace installation to the current directory: "ln -s $2 ."

10. go to the parent folder (should be "lib")

11. link the InstrumentFiles folder to this directory: "ln -s $3 ."

12. create a file called startup.m within this directory

13. edit it and insert the following:
try
libisis_off
catch
end
addpath('$1/matlab') %if there is a matlabfolder within the libsis
directory use that, if all the matlab routines are directly within the
libisis folder, then omit the "/matlab"
libisis_init
try
horace_off
catch
end
addpath('$2')
horace_init

14. Now you can start matlab, go to the lib folder and run the startup
routine -> Libisis and Horace should be initialised.

==If you need Libisis for data reduction==

Additional steps have to be considered if you need to use Libisis for data reduction.
Libisis uses NeXus to access instrument description and NeXus[http://www.nexusformat.org] is HDF-based[http://www.hdfgroup.org/HDF5/] file format. Matlab also supports HDF5 and is supplied with its own HDF5 version. Unless you are very lucky, this would be not the version, the NeXus is linked with. Different run-time versions of HDF5 are incompatible with each other. (e.g. Matlab 2010-2011a uses HDF5 1.8.3 and Matlab 2011b uses HDF 1.8.6, when NeXus in Lucid 10.4.3 repository comes linked with HDF5-1.8.4 version). This mean that your Matlab application will crash as soon as you try to use NeXus unless you link NeXus against HDF5-libraries statically.

To do this you have to download static library version of HDF-5 or build this version from the source code and build NeXus linked against this library.
The example of how to do it is as below:

cd hdf5-1.8.3/
ls
env CFLAGS="-fPIC -DPIC" ./configure --prefix=/usr/local/hdf5 --disable-fortran --with-pic --disable-shared
make
make check
ls /usr/local/
ls /usr/local/hdf5/
sudo make install
ls /usr/local/hdf5/
ls /usr/local/hdf5/lib/

cd nexus-4.2.1/
sudo apt-get mxml

./configure --with-hdf5=/usr/local/hdf5/
ls /usr/local/lib/
make
make check
sudo make install

User experience

2011-10-04T12:47:21Z

Alex Buts: /* If you need Libisis for data reduction */

User experience

2011-10-04T12:40:17Z

Alex Buts: /* Jülich Centre for Neutron Science */

2011-09-19T11:03:40Z

Alex Buts: /* Jülich Centre for Neutron Science */

== ISIS LET control machine ==
The PC came prepared for Libisis and Horace. NeXus, Matlab and Intel Fortran compilers are installed. The task is to identify the locations of all parts, necessary for instillation and run the installation script resolving all dependencies.

'''1)''' Verify if nexus is installed
$ locate libNeXus
...
/usr/lib/libNeXus.so
...

'''2)''' Check if ifort works form command line:
$ifort --version
ifort (IFORT) 11.1 20100414
Copyright (C) 1985-2010 Intel Corporation. All rights reserved.

'''3)''' Identify where Matlab is really installed and if it has all files necessary to use it.
$what matlab:
/usr/local/bin/matlab
$
$ls -l /usr/local/bin
...
lrwxrwxrwx 1 root root 38 2011-03-18 11:48 matlab -> /usr/local/MATLAB/Matlab2010bSP1/bin/matlab
...

'''/usr/local/MATLAB/Matlab2010bSP1 is in our case or $MATLAB_ROOT folder which we will use everywhere, referring to matlab'''

'''4)''' Verify if ''mwdebug.c'' file is present:
$
$ls $MATLAB_ROOT/extern
'''examples include lib'''

'''4a)''' ''etc'' folder is not there, so we have to make it and place ''mwdebug.c'' there.
$
$sudo mkdir $MATLAB_ROOT/extern/etc
$sudo mv ~/Downdloads/mwdebug.c $MATLAB_ROOT/extern/etc/

'''5)''' Download libisis from the repository
$svn co https://svn.libisis.org/code/trunk Libisis

'''6)''' Use autotools to build libisis:
$cd Libisis
$ sh autogen.sh
...
A lot of messages, no errors
...

$./configure --with-matlabroot=/usr/local/MATLAB/Matlab2010bSP1
...
A Lot of messages, no errors
...

$make
...
A lot of messages no errors
...

'''6)''' Check if libisis has been build successfully.
$ls -l ~/Libisis/Libisis/DLL/_glnxa64
...
-rwxr-xr-x 1 User Group 36434653 2011-09-14 19:09 libisisexc.mexa64
...

I assume 64 bit linux here. If it is 32-bit linux, check ''~/Libisis/Libisis/DLL/_glnxa'' folder;

'''7)''' Building binary part of libisis is now completed. Start Matlab and check if it was successful.
$matlab&
in matlab:
>>cd ~/Libisis/Libisis/DLL/_glnxa64
>> aa=libisisexc();
>> aa =
$Revision:: 1808 $ ($Date:: 2011-03-29 21:43:21 +0000 (Tue, 29 Mar 2011) $)

Follow Windows Maltab installation steps, described on [http://www.libisis.org/Download_and_setup_LIBISIS#Download_LIBISIS Libisis page].

== University of Colorado Group ==

Special thanks to '''Dan Parshall'''; dan.parshall_AT_colorado.edu, who shared his experience with us.

This took a while to get back to, but eventually we got this sorted out.
It was a combination of legacy files, compilers, permissions issues, and
typos. I'm posting the solution here in order to help future Ubuntu
users.

1) >>> get legacy file mwdebug.c in place <<<
The file mwdebug.c (which one can get from Alex or Freddie) must be
placed into matlabroot/extern/src/ (one needs to make the src
directory). Do the standard:
chown -R root matlabroot/*
chmod -R 755 matlabroot/*

2) >>> set up IFORT <<<
A 30-day trial is free, academic licenses fairly cheap (I've mucked about a bit, and gfortran seems to also be functional, but as a rule it's 20% slower).
In order to invoke IFORT on the command line, one has to use the "source" command and link to the "compilervars.sh" script.

Because one can't use "source" in combination with "sudo", one must change to the
root account, then complete the remaining steps:
cd /path/to/libisis
sudo su
source /path/to/ifort/compilervars.sh intel64

3) >>> run autogen.sh / configure / make (as root) <<<
One must set matlabroot directly- simlinks aren't followed properly
bash autogen.sh
./configure --with-f77=ifort --with-f90=ifort --with-matlabroot=/path
make

4) >>> set ownership and permissions of files made during compilation <<
chown -R <username> *
chmod -R 777 *

== Jülich Centre for Neutron Science ==

Special thanks to '''Johannes Reim''', j.reim_AT_fz-juelich.de; who shared his experience with us;

And afterwards I have to say, actually it is not that complex, if you
know a bit about matlab, linux and how horace and libisis works but
well, neither of that applied to me *g*.

Okay what did I do for the installation:

1. Installation of libisis, I actually did not do that myself, but the
IT did it, because it should be installed for all users within our
institute, but I sent them the instructions found on the subversion
server of Libisis:
http://svn.libisis.org/viewvc/LIBISISCode/trunk/Linux/README.linux?revision=1577

I will call the libisis directory $1 (think of $1as the variable for the
full path to the directory)

2. check if there is a matlab folder within the libisis folder (we will
need that later on)

3. download the up-to-date version of horace (not the one which is
specified for linux)

4. Just extract it to your directory of choice, I will call the
directory $2 (like $1)

5. download the homer-standalone from:
http://download.libisis.org/kits/homer_standalone.zip

6. create a folder called InstrumentFiles, my suggestion would be
creating it next to the Horace directory.

7. extract the content of the instrument_setup folder within into the
created directory, I will call the directory $3 (like $1)

8. if not existing, create a lib directory within your home directory,
and into that a matlab directory, access the matlab directory

9. link the Horace installation to the current directory: "ln -s $2 ."

10. go to the parent folder (should be "lib")

11. link the InstrumentFiles folder to this directory: "ln -s $3 ."

12. create a file called startup.m within this directory

13. edit it and insert the following:
try
libisis_off
catch
end
addpath('$1/matlab') %if there is a matlabfolder within the libsis
directory use that, if all the matlab routines are directly within the
libisis folder, then omit the "/matlab"
libisis_init
try
horace_off
catch
end
addpath('$2')
horace_init

14. Now you can start matlab, go to the lib folder and run the startup
routine -> Libisis and Horace should be initialised.

2011-09-19T09:22:36Z

Alex Buts: Created page with "== ISIS LET control machine == The PC came prepared for Libisis and Horace. NeXus, Matlab and Intel Fortran compilers are installed. == University of Colorado Group == Spacial..."

Download and setup LIBISIS

2011-09-19T08:36:48Z

Alex Buts: /* Download LIBISIS */

By downloading any of the files below, you accept the terms of the Matlab MCR Library License included with the distribution. Read the COPYING.rtf file in the License folder for information about which components may or may not be changed or redistributed.

If you have any problems, email the [http://lists.libisis.org/mailman/listinfo/libisis-users libisis-users] mailing list.

===Requirements===

* Matlab 2007a (or later)
* 1GB RAM memory
* 40MB free hard drive space

== Download LIBISIS ==
* Download the latest libisis build for your version of Matlab

{| {{Tablestyle}}
|-
! {{Headcellstyle}} | Windows
! {{Headcellstyle}} | Linux&MacOS
|-
|
*'''Latest distributive for Matlab version 7.4(2007a) and higher (32 and 64 bit Matlab versions):''' (''this is what you probably need'')

[http://download.libisis.org/kits/libisis_distribution_kit.zip Matlab 2007a and higher]

* Old (compatibility) distributives for older versions of Matlab (32bit) (''use this only if your Matlab is really old or you used Libisis before and want to process your previous data with your old libisis''):
[http://download.libisis.org/kits/libisis1.zip Matlab Ver 7.1]
[http://download.libisis.org/kits/libisis1.zip Matlab 2006a]
[http://download.libisis.org/kits/libisis2.zip Matlab 2006b]
[http://download.libisis.org/kits/libisis2.zip Matlab 2007a]
[http://download.libisis.org/kits/libisis2.zip Matlab 2007b]

|You need to install [http://download.nexusformat.org/kits/ NeXus] and build the code from source using GNU autotools, see [[SubversionServer]] for instructions
You should be well familiar with Unix installation process to do it smoothly. There is [[user experience]] page, which summarises the experience of some users who have already done this.
|}

===Windows===
* '''All versions : '''
* Unpack downloaded file into <tt>'''C:\mprogs'''</tt> or a directory of your choice eg <tt>'''X:\ISIS'''</tt>
* Open your startup.m file (if it does not exist, then startup.m in the folder that matlab starts)
* Add the following lines to the startup.m file

<pre>addpath('c:\mprogs\Libisis') % where c:\mprogs\Libisis is the folder you unzipped Libisis into
libisis_init</pre>

* '''for the latest versions : '''
* Instead of above, you can unpack downloaded file into a folder of your choice. (ISIS by default so ''''C:\ISIS\'''' and this folder have two sub-folders namely '''Libisis''' and '''InstrumentFiles''' in it)
* Start MATLAB and navigate it to the folder where the Libisis has been unpacked. (e.g. C:\ISIS)
* Find the MATLAB function '''install_isis''' in this folder and run this function specifying the path (short or full) to the folder where you have Libisis unpacked. (e.g. if your Libisis is in C:\ISIS\Libisis, navigate MATLAB to the folder C:\ISIS and run '''install_isis('Libisis')''' or '''install_isis('C:\ISIS\Libisis')'''
* Type '''libisis_on''' in MATLAB command prompt to access Libisis functions
This is actually all you need having an administrative rights to the system. install_isis creates an folder '''ISIS''' in the toolbox folder of the MATLAB root directory, places files necessary to initiate Libisis there and modifies MATLAB search path to look for files in this folder. If you use startup file, the script also adds the rows, described above and necessary to initiate Libisis during MATLAB startup. These rows are added to the end of your startup file.

'''You need manually modify you startup file if you used Libsisis before to remove references to previous Libisis installations and avoid multiple Libisis initialisations.'''

== Download Stand Alone Homer ==
* If you do not have matlab it is possible to download an executable which will run the homer GUI standalone
* Homer software created by Dr. [[User:Dickon Champion|Dickon Champion]], standalone installation developed by [[User:Dean Whittaker|Dean Whittaker]]. MCRInstaller and Matlab (c) [http://www.matlab.com The MathWorks, Inc.] 1984-2010

{| {{Tablestyle}}
|-
! {{Headcellstyle}} | Windows
!{{Headcellstyle}} | Linux
! {{Headcellstyle}} | MacOS
|-
| [http://download.libisis.org/kits/homer_standalone.zip Stand Alone Homer] or [http://download.libisis.org/kits/homer_standalone_kit_win64.zip Stand Alone Homer for win64 (with proper MCR installer)]- the standalone homer zip file

[http://download.libisis.org/kits/MCRInstaller.exe MCRInstaller] - required for the first run of homer.

| not available
| not available
|}

you must [[Homer Gui setup|set up]] Homer properly in order for it to work. Run the MCRInstaller once on your computer, if updating your homer installation, it is not required.

===Install Stand Alone Homer===

* The '''First''' time you download homer:
** Download the MCRInstaller file above
** Run the MCRInstaller.exe file and follow the on-screen instructions
* Download the homer zip file above
* Unzip to directory of your choice
* Run homer.exe to start homer.

===Requirements===

* 1GB RAM memory
* 40MB free hard drive space
* Administrative Rights for Installation.

== Setup Memory Usage ==

Sometimes when we are treating large datasets, such as on MAPS, we cannot load all the data into the memory of the computer at once, we need to 'chunk' the data into smaller pieces. We therefore need to set the default number of workspaces in each 'chunk' to treat at any one time during the population of an [[IXTrunfile]] and the total amount of memory to allow the computer to use.

If the memory required for the number of workspaces in each chunk is greater than that defined then the number of workspaces is reduced until it fits in the memory or a failure status is raised. These two values are defined in <tt>./matlab/homer/default_homer_nchunk.m</tt> which needs to be edited to suit the memory needs of your computer. Tests have shown that LIBISIS will populate data fastest when it is chunked in groups of 100 workspaces, so this is the default value in the file.

<pre>
function nchunk=default_homer_nchunk
% nchunk = default_homer_nchunk
% nchunk is a two membered array
% the first element is the default number of workspaces to be treated at
% one time, and the second element is the maximum amount of memory (in MB)
% that you want the computer to use
nchunk=int32([100,500]);
</pre>

Download and setup LIBISIS

2011-03-31T11:07:34Z

Alex Buts: /* Download LIBISIS */

By downloading any of the files below, you accept the terms of the Matlab MCR Library License included with the distribution. Read the COPYING.rtf file in the License folder for information about which components may or may not be changed or redistributed.

If you have any problems, email the [http://lists.libisis.org/mailman/listinfo/libisis-users libisis-users] mailing list.

===Requirements===

* Matlab 2007a (or later)
* 1GB RAM memory
* 40MB free hard drive space

== Download LIBISIS ==
* Download the latest libisis build for your version of Matlab

{| {{Tablestyle}}
|-
! {{Headcellstyle}} | Windows
! {{Headcellstyle}} | Linux&MacOS
|-
|
*'''Latest distributive for Matlab version 7.4(2007a) and higher (32 and 64 bit Matlab versions):''' (''this is what you probably need'')

[http://download.libisis.org/kits/libisis_distribution_kit.zip Matlab 2007a and higher]

* Old (compatibility) distributives for older versions of Matlab (32bit) (''use this only if your Matlab is really old or you used Libisis before and want to process your previous data with your old libisis''):
[http://download.libisis.org/kits/libisis1.zip Matlab Ver 7.1]
[http://download.libisis.org/kits/libisis1.zip Matlab 2006a]
[http://download.libisis.org/kits/libisis2.zip Matlab 2006b]
[http://download.libisis.org/kits/libisis2.zip Matlab 2007a]
[http://download.libisis.org/kits/libisis2.zip Matlab 2007b]

|You need to build the code from source using GNU autotools; see [[SubversionServer]] for instructions
|}

===Windows===
* '''All versions : '''
* Unpack downloaded file into <tt>'''C:\mprogs'''</tt> or a directory of your choice eg <tt>'''X:\ISIS'''</tt>
* Open your startup.m file (if it does not exist, then startup.m in the folder that matlab starts)
* Add the following lines to the startup.m file

<pre>addpath('c:\mprogs\Libisis') % where c:\mprogs\Libisis is the folder you unzipped Libisis into
libisis_init</pre>

* '''for the latest versions : '''
* Instead of above, you can unpack downloaded file into a folder of your choice. (ISIS by default so ''''C:\ISIS\'''' and this folder have two sub-folders namely '''Libisis''' and '''InstrumentFiles''' in it)
* Start MATLAB and navigate it to the folder where the Libisis has been unpacked. (e.g. C:\ISIS)
* Find the MATLAB function '''install_isis''' in this folder and run this function specifying the path (short or full) to the folder where you have Libisis unpacked. (e.g. if your Libisis is in C:\ISIS\Libisis, navigate MATLAB to the folder C:\ISIS and run '''install_isis('Libisis')''' or '''install_isis('C:\ISIS\Libisis')'''
* Type '''libisis_on''' in MATLAB command prompt to access Libisis functions
This is actually all you need having an administrative rights to the system. install_isis creates an folder '''ISIS''' in the toolbox folder of the MATLAB root directory, places files necessary to initiate Libisis there and modifies MATLAB search path to look for files in this folder. If you use startup file, the script also adds the rows, described above and necessary to initiate Libisis during MATLAB startup. These rows are added to the end of your startup file.

'''You need manually modify you startup file if you used Libsisis before to remove references to previous Libisis installations and avoid multiple Libisis initialisations.'''

== Download Stand Alone Homer ==
* If you do not have matlab it is possible to download an executable which will run the homer GUI standalone
* Homer software created by Dr. [[User:Dickon Champion|Dickon Champion]], standalone installation developed by [[User:Dean Whittaker|Dean Whittaker]]. MCRInstaller and Matlab (c) [http://www.matlab.com The MathWorks, Inc.] 1984-2010

{| {{Tablestyle}}
|-
! {{Headcellstyle}} | Windows
!{{Headcellstyle}} | Linux
! {{Headcellstyle}} | MacOS
|-
| [http://download.libisis.org/kits/homer_standalone.zip Stand Alone Homer] or [http://download.libisis.org/kits/homer_standalone_kit_win64.zip Stand Alone Homer for win64 (with proper MCR installer)]- the standalone homer zip file

[http://download.libisis.org/kits/MCRInstaller.exe MCRInstaller] - required for the first run of homer.

| not available
| not available
|}

you must [[Homer Gui setup|set up]] Homer properly in order for it to work. Run the MCRInstaller once on your computer, if updating your homer installation, it is not required.

===Install Stand Alone Homer===

* The '''First''' time you download homer:
** Download the MCRInstaller file above
** Run the MCRInstaller.exe file and follow the on-screen instructions
* Download the homer zip file above
* Unzip to directory of your choice
* Run homer.exe to start homer.

===Requirements===

* 1GB RAM memory
* 40MB free hard drive space
* Administrative Rights for Installation.

== Setup Memory Usage ==

Sometimes when we are treating large datasets, such as on MAPS, we cannot load all the data into the memory of the computer at once, we need to 'chunk' the data into smaller pieces. We therefore need to set the default number of workspaces in each 'chunk' to treat at any one time during the population of an [[IXTrunfile]] and the total amount of memory to allow the computer to use.

If the memory required for the number of workspaces in each chunk is greater than that defined then the number of workspaces is reduced until it fits in the memory or a failure status is raised. These two values are defined in <tt>./matlab/homer/default_homer_nchunk.m</tt> which needs to be edited to suit the memory needs of your computer. Tests have shown that LIBISIS will populate data fastest when it is chunked in groups of 100 workspaces, so this is the default value in the file.

<pre>
function nchunk=default_homer_nchunk
% nchunk = default_homer_nchunk
% nchunk is a two membered array
% the first element is the default number of workspaces to be treated at
% one time, and the second element is the maximum amount of memory (in MB)
% that you want the computer to use
nchunk=int32([100,500]);
</pre>