A USB flash drive can be made into a substitute for a DVD in just three easy steps without using DISKPART commands.

1. Attach the USB drive and format it as NTFS, it must be NTFS for it to work

2. You have to mount your ISO one of the many tools available.

3. At a command prompt with Administrator privileges, use the following switches: xcopy d:\*.* /s /e /h /f g:\

where:
         D is the drive letter for your mounted ISO file or CD/DVD.  
         G is the drive letter for your USB flash drive.

    If your drive letter mappings are different, make the appropriate substitutions.

    XCOPY switches are described here.

4. Attach your USB drive to the machine and boot. You may have to go into the BIOS to force the machine to boot from the USB device (Most modern machines are capable of this).

And you are done!

MATLAB like most ISVs listed on Microsoft's HPC website, under supported applications, do already have some work done on integration. Although the links are there, users need to just take an additional step of performing the configuration and then they are ready to leverage on the power of parallel computing!

So what I am going to do, is I am go to walk through the steps required to configure MATLAB 7.8.0 (R2009a) to use Windows HPC Server 2008 for parallel processing.

So the very first step would be to launch the application. After launching the application, this is what most of you will see.

You will notice that at the top, on the menu bar, there is a "Parallel" option. By clicking it on it and then selecting "Configurations Manager", you should see the following screen. You should see only one option for now, we are going to add another.

On that screen, you need to click the "File" option. You should see the following.

You will see lots of other options but for now we will just focus on the "ccs" option. By selecting the "ccs" option, it will bring you to another screen shown below.

For now, we can leave the "Configuration Name" as it is, it just allows you to identify your Windows HPC configuration, situations where it might come in useful would be if you have various Windows HPC clusters, then you might have different names for different clusters, something meaningful that you can identify with.

Now let's explore the various options that are available to you. I will try and explain what each of the options mean as we go along.

First option, "Root Directory of MATLAB installation for workers (ClusterMatlabRoot)", this option simply points to where (the directory) MATLAB is installed on the compute nodes. I have to stress that in order for this work properly, MATLAB should be installed on all the nodes with a common path. i.e. if you install to C:\MATLAB on one node, MATLAB on all the other nodes should be installed to C:\MATLAB as well. Otherwise, the job/task/simulation will FAIL when you start. This is where your Node Templates will come in really useful to maintain a consistent image across the cluster.

Second option, "Number of workers available to scheduler (ClusterSize)", simply, the number of nodes that you might have or available to you to run MATLAB. Bearing in mind, the number of licenses, etc...

Third option, "Directory where job data is stored (DataLocation)", this is where your input and output files will go. Bear in mind that, I will choose a shared folder for this, if not, this is the possible scenario.

You will have to copy ALL your input data to a consistent location on ALL the compute nodes, i.e. if you choose C:\data, then C:\data MUST be available on al the compute nodes whether or not that node will be used because you will never know which node gets selected to run the simulation. Secondly, ALL your output data will go to that location. Which simply means that you will have to log onto each compute to check if your output is in that node... And if, this is not consistent, then your job will fail. Personally, I do not want to do that if I have a cluster of 128 nodes. So, do yourself a favour and choose a shared folder.

Fourth option, "CCS scheduler hostname (SchedulerHostname)", for this option, you just need to enter the hostname of your head node, and please do ensure that it is resolvable by DNS if you choose hostname.

Last option, "Workers run in SOA mode for distributed jobs (UseSOAJobSubmission)", setting this option will tell MATLAB to leverage on the latest feature available to Windows HPC Server 2008 which is the SOA mode jobs. Personally, I have not tried this option, and if you do, and you do find that it helps you to speed up the calculations, please drop me a comment, I would love to hear about it.

So to sum up, after you are done with the configuration, it should look something like this!

For now, these are the options that you need to set in order to get started. Click "OK" and you will be taken back to the "Configurations Manager" screen. Now on that screen, you should see two options now, the original, "Local" and your recently added "ccsconfig1" as shown below.

Now, I have to say that MATLAB provides an excellent function to validate the cluster configurations, it helps you to check if the cluster if running properly. So right now, you can click "Start Validation" and let MATLAB verify for you if the configurations are all in order.

So once you click "Start Validation", you should see the following.

From this point on, basically, your configuration is and you are ready to leverage your cluster. However, there is still some things that you need to do, you need to "parallelise" your scripts. For more information, you might want to consult the MATLAB Parallel Computing Toolbox. The site is here. You can even download the PDFs for offline viewing.

Some points to note though, ACUSIM recommends that only AcuSolve 1.7f  works on Windows HPC Server 2008 in a distributed mode, and since this is the version that I got my hands on, I can only ensure that this version works, and it did.

Now that we have gotten that out of the way, I'm going to elaborate a little on the environment set up that I used.

1) I have a cluster of 3 compute nodes running Windows HPC Server 2008. Excluding the head node.

2) ActivePerl for 64-bit has been installed on  the nodes, individually. If you plan on running a single instance of Perl installation, please do get in touch, I’ll try and see if I can get it to work. (This will come into play later when I outline the steps to get things to work.)

3) Things to take note, the batch scripts provided by ACUSIM has typos in them. Note: I performed the installation to a D: and the scripts were seemingly hardcoded to a C: install, and even then, I’m not too sure if they are right because they point to a C:\Acusim\classic, but my installation went into a D:\Acusim only. I have since fed this back to ACUSIM. :)

Steps to get it to work:

1) First, we need to fix the paths, from this point onwards, I am assuming that you have performed a first-cut installation to your head node and shared the Acusim installation directory out already. Now, you need to go to your Acusim installation “bin” directory and locate this perl script file called acuFixPath.pl. Edit line 190 to read “$outfile= "$home\@TMP_$file" ;” Reason is this, the script was coded for *nix based systems and the original syntax didn’t work. Also this syntax will assume that you will specify a $home for ACUSIM home, which we will.

2) Second, you need to run acuFixPath.bat with the following options, in my case it would be “acuFixPath.bat –files _all –home \\hpchead\Apps2\Acusim -machine WIN64 -version V1.7f –v 1” This would update all the paths to now point to the head node shared directory as the home installation for ACUSIM. This would fix all the path typos.

3) Now that the paths are fixed, the next step would be to get the compute nodes to accept UNC paths in its command processor. Now, log into each of the compute nodes individually, and execute “\\hpchead\Apps2\Acusim\WIN64\V1.7f\bin\acuFixUnc.bat”. Unfortunately, this has to be done individually on each of the nodes. Once the script has executed, the compute nodes are ready to accept UNC paths on its command processor.

4) Now we need to customize the acuSub.bat to point to the right paths. In my case, I referenced acuEnv.bat to get certain parameters into the acuSub.bat file. Once you open up your acuSub.bat file you need to paste the following lines above the line that reads, “perl %ACUSIM_HOME%\%ACUSIM_MACHINE%\%ACUSIM_VERSION%\bin\acuSub.pl %opt%”

rem -----------------------------------------------------------------------

rem Basic parameters

rem -----------------------------------------------------------------------

set ACUSIM_HOME=\\hpchead\Apps2\Acusim
set ACUSIM_VERSION=V1.7f
set ACUSIM_MACHINE=WIN64
set ACUSIM_ROOT=%ACUSIM_HOME%\%ACUSIM_MACHINE%\%ACUSIM_VERSION%
set ACUSIM_SYSTEM_CNF=%ACUSIM_ROOT%\script\Acusim.cnf
set ACUSIM_CNF_FILES=.\Acusim.cnf:~\Acusim.cnf:%ACUSIM_SYSTEM_CNF%

set PATH=%ACUSIM_ROOT%\base\lib\site-packages;%PATH%
set PATH=%ACUSIM_ROOT%\base\DLLs;%PATH%
set PATH=%ACUSIM_ROOT%\base\bin;%PATH%
set PATH=%ACUSIM_ROOT%\bin;%PATH%

set PATH=%ACUSIM_ROOT%\ODB_6.9\lib;%PATH%

set Visual3_CP=%ACUSIM_ROOT&\script
set PERL5OPT=-I%ACUSIM_ROOT%\script %PERL5OPT%
set PYTHONPATH=%ACUSIM_ROOT%\base\bin
set PYTHONPATH=%ACUSIM_ROOT%\script;%PYTHONPATH%
set P_SCHEMA=%ACUSIM_ROOT%\script\schema

set PV_INTERFACE_PATH=%ACUSIM_ROOT%\pv\packages
set ACUSIM_PARAVIEW_LIB=%ACUSIM_ROOT%\pv\bin\vtkAcuSolveReader.dll

This is to set up the ACUSIM environment when it runs in the compute nodes. Please do let me know if you want to know more and why.

5) We also need to edit the acuSub.pl file. At line 217 of the acuSub.pl file you will see, “$perlPath = "C:\\Perl\\bin" if $perlPath eq "_auto";” You need to change it to read “$perlPath = "C:\\Perl64\\bin" if $perlPath eq "_auto";” (This is because I’m using ActivePerl’s 64-bit installation, I guess if you are using the 32-bit one, you are fine.)

6) Because we are using the Windows HPC Server MPI layer, we need to submit it through the batch scheduler. But before you do that, you need to edit your system Acusim.cnf. You need to change the 2 lines, to read as follows:

message_passing_type=          msmpi
view_message_passing_type=     msmpi

You need to use msmpi. Original setting was using HPMPI.

7) So, once you are done, you can submit your job and start having fun! I used the following parameters. “D:\Jerald_Share\AS_door>acusub -np 12 -sched ccs -pdir \\hpchead\Jerald_Share\AS_Door -pb door”

With Windows Server 2008 and Windows Vista, Microsoft has introduced VA 2.0. Also, on top of that, quiet a few of our customers are using Volume Licensing (VL) and VL needs a local activation server to work.

So my best advice?
Refer to https://www.microsoft.com/licensing/resources/vol/default.mspx for an overview of Volume Licensing.

As they say, the devil is in the details.
https://technet.microsoft.com/en-us/library/bb892849.aspx