multiple instances of function call in paralle

hi,currently this is is my code structure

arg4=10;

open a txt file

for arg1=20:10:60 for arg2= 20:10:60 for arg3= 20:10:60 [ans1 ans2]=myfunt([arg1 arg2],[arg3,arg4]); write in txt file end end end close txt file

note: 1. myfunt.m further calls to different files. 2. each myfunt()function call is Independent from each other.

total calls 50*50*50= 125000 times

Q1:Now i do consider code like a distributed job as multiple instances of same function call with different data parameters.is it right?

Q2:how can i run it on local machine with dual2core or quadcore processor in a more faster manner. like i want maximum instances to run parallel.i dont want to use parfor etc because its putting many limitations on my code.

Q3: how can i run this code at some/any server with 16 or more core?

thank you

Answer 1

[Friedrich]

Hi,

Iam not sure if this will speed up things but I would try the following

%generate all input args
all_args ={};
i = 1;
arg4=10;

for arg1 = 20:10:60 
    for arg2 = 20:10:60 
        for arg3= 20:10:60 
            all_args{i} = [arg1,arg2,arg3,arg4];
            i = i +1;
        end
    end
end

%call cellfun, should work more or less in parallel,
%NOTE: you have to modify the myfunt input structure to 1x4 
%array instead of two arrays of size 1x2

[out1,out2] = cellfun(@myfunt,all_args,'UniformOutput',false);

%write the stuff to file