Memory usage

[David Pruitt]

Here are two functions that I am using to sort through a database prior to analysis.  I call these function many times through a particular run.  My issue is the consumption of memory through the run.  If you recall my previous message, a long run will consume >750MB of memory while I can account for only about 10MB.  John Eaton suggested that looping in some functions might lead to virtual memory leaks.  I would appreciate any comment on my programming in two example functions. 

 

Thanks for the help.

 

Example1.m:

function [ out, weight, nmax, idx ] = S_out(code, mat1, mat2, S1, S2 )

#

# Input:

#       code - the three-column vector of three-letter reference codes

#       mat1 - matrix of values: if mat1 is a matrix of

#         a n-length time series of values for m observations, the m values

#         are averaged over time.

#       mat2 - matrix of identifying codes

#

# Output:

#        out - m-vector of weighted values

#        weight - weights

#        n - largest number of observations for any identifying code

#        idx - index of codes found in mat2

#

  m = size(idx,1);

  n = size(cap_mat,1);

  if n > 1

    mat1_mean = mean(mat1);

  else

    mat1_mean = mat1;

  endif

  for i = 1:m

    ix = [];

    ix = find( idx(i,:) != 0 );

    n(i) = size(ix,2);

    out(i) = sum( mat1_mean( idx(i, ix) ) );

  endfor;

  nmax = max(n);

 

  weight = zeros(m,nmax);

  imx = find(mat != 0);

  for i = imx'

    ix = [];

    ix = find( idx(i,:) != 0 );

    weight(i,ix) = (mat1_mean( idx(i, ix) ))/out(i);

  endfor;

 

endfunction;

 

Example 2:

function  [ Rtn_val, idx ] = S_Index(mat, code)

#

#  Return the ID-sorted indicies of elements in the input code

#

#  Input:

#           mat = database values (including several character strings - positions known)

#           code = the three-letter from the database

#           S1, S2 = reference indicies

#

#  Output:

#          Rtn_val = Vector of identification numbers

#          idx = the set of indicies from the ID-sorted list

#

#

  ID = mat(:,1:5);

  matrix = toascii(mat(:,S1:S2));

  c_code = toascii(code);

 

  m = size(code,1);

  n = size(sec_mat,2);

  idx = zeros(m,n);

  Rtn_val = zeros(m,1);

  for i = 1:m

 

    iz = [];

    iz1 = [];

    iz2 = [];

    iz3 = [];

    iz1 = find( matrix(:,1) == c_code(i,1) );

    iz2 = find( matrix(:,2) == c_code(i,2) );

    iz3 = find( matrix(:,3) == c_code(i,3) );

    iz = intersection(iz1,iz2);

    iz = intersection(iz,iz3);

    n = size(iz,2);

    idx(i,1:n) = iz;

    for j = 1:n

      Rtn_val = str2num(ID(j,1:5));

    endfor;

 

  endfor;

 

 

endfunction;

 

 

David Pruitt, PhD

Ashdon Investment Analysis and Reseach, LLC

Phone: 931.695.3220

Mobile: 931.639.2477

Fax: 931.695.5545

e-mail: address@hidden