Running MATLAB script that uses parallel processing from a windows batch file

I'm stumped on how to run a MATLAB script that uses parallel processing from a windows batch file without having the overhead of starting a new parpool each time the code (or Matlab session) executes. Note. I am running this all on my 'local' computer.

Specifically, when I execute the following windows batch file to run a Matlab sript, i.e.'mywave.m'; each time I execute the batch file, the new Matlab session starts a new parpool on my 'local' machine which incurs about 25 seconds of overhead time. Since I had already created a parpool, how can I utilize the existing parpool without starting/creating an entirely new parpool each time I execute my windows batch file?

For example, after creating the initial parpool, I execute the the following windows batch file to execute my Matlab script file:

"D:\Program Files\MATLAB\R2016b\bin\matlab.exe" -nodisplay -nosplash -nodesktop -r "try, run('D:\Matlab\Work\mywave.m'), catch, exit, end; exit;"

Code for ('mywave.m')

parfor i = 1:1024
A(i) = sin(i*2*pi/1024);
end

Therefore, is it possible to run Matlab from the windows command line and utilize an existing parpool? And if so, how?

Thanks in advance.

Answer 1

[Jonathan Kwang]

Unfortunately, if you exit MATLAB while the parallel pool is running, the parallel pool shuts down along with MATLAB so new parallel pool would need to be started when another instance of MATLAB is started.

[Walter Roberson]

Is this something that MATLAB Production Server could help overcome? Or perhaps the Distributed Computing Server might allow the workers to be "kept warm" ?

[Darrell]

Thanks for the response; unfortunately I was afraid that was going to be the answer. However, it would be nice if you could add an argument to call an existing Matlab session from the command line. That would save a lot of time from having to load Matlab and creating a new pool of workers.