In this recipe, we will work with the following example Fortran code (example.f90):

program example

  implicit none

  real(8) :: array(20000000)
  real(8) :: r
  integer :: i

  do i = 1, size(array)
    call random_number(r)
    array(i) = r
  end do

  print *, sum(array)

end program

The fact that this is Fortran code does not matter much for the discussion that will follow, but we have chosen Fortran since there is a lot of legacy Fortran code out there where static size allocations are an issue.

In this code, we define an array holding 20,000,000 double precision floats, and we expect this array to occupy 160 MB of memory. What we have done here is not recommended programming practice, since in general this memory will be consumed independently of whether it is used in the code. A much better approach would have been to allocate the array dynamically only when it is needed and deallocate it right afterwards.

The example code fills the array with random numbers and computes their sum - this was done to make sure that the array is really used and the compiler does not optimize the allocation away. We will measure the size of static allocation of the example binary with a Python script (static-size.py) wrapping around the size command:

import subprocess
import sys

# for simplicity we do not check number of
# arguments and whether the file really exists
file_path = sys.argv[-1]

try:
    output = subprocess.check_output(['size', file_path]).decode('utf-8')
except FileNotFoundError:
    print('command "size" is not available on this platform')
    sys.exit(0)

size = 0.0
for line in output.split('\n'):
    if file_path in line:
        # we are interested in the 4th number on this line
        size = int(line.split()[3])

print('{0:.3f} MB'.format(size/1.0e6))

To print the link line, we will use a second Python helper script (echo-file.py) to print the contents of a file:

import sys
  
# for simplicity we do not verify the number and
# type of arguments
file_path = sys.argv[-1]

try:
    with open(file_path, 'r') as f:
        print(f.read())
except FileNotFoundError:
    print('ERROR: file {0} not found'.format(file_path))