Function memory_usage

memory_profiler.py:269–480

Return the memory usage of a process or piece of code Parameters ---------- proc : {int, string, tuple, subprocess.Popen}, optional The process to monitor. Can be given by an integer/string representing a PID, by a Popen object or by a tuple representing a P

(proc=-1, interval=.1, timeout=None, timestamps=False,
                 include_children=False, multiprocess=False, max_usage=False,
                 retval=False, stream=None, backend=None, max_iterations=None)

def memory_usage(proc=-1, interval=.1, timeout=None, timestamps=False,
                 include_children=False, multiprocess=False, max_usage=False,
                 retval=False, stream=None, backend=None, max_iterations=None):
    """
    Return the memory usage of a process or piece of code

    Parameters
    ----------
    proc : {int, string, tuple, subprocess.Popen}, optional
        The process to monitor. Can be given by an integer/string
        representing a PID, by a Popen object or by a tuple
        representing a Python function. The tuple contains three
        values (f, args, kw) and specifies to run the function
        f(*args, **kw).
        Set to -1 (default) for current process.

    interval : float, optional
        Interval at which measurements are collected.

    timeout : float, optional
        Maximum amount of time (in seconds) to wait before returning.

    max_usage : bool, optional
        Only return the maximum memory usage (default False)

    retval : bool, optional
        For profiling python functions. Save the return value of the profiled
        function. Return value of memory_usage becomes a tuple:
        (mem_usage, retval)

    timestamps : bool, optional
        if True, timestamps of memory usage measurement are collected as well.

    include_children : bool, optional
        if True, sum the memory of all forked processes as well

    multiprocess : bool, optional
        if True, track the memory usage of all forked processes.

    stream : File
        if stream is a File opened with write access, then results are written
        to this file instead of stored in memory and returned at the end of
        the subprocess. Useful for long-running processes.
        Implies timestamps=True.

    backend : str, optional
        Current supported backends: 'psutil', 'psutil_pss', 'psutil_uss', 'posix', 'tracemalloc'
        If `backend=None` the default is "psutil" which measures RSS aka "Resident Set Size". 
        For more information on "psutil_pss" (measuring PSS) and "psutil_uss" please refer to:
        https://psutil.readthedocs.io/en/latest/index.html?highlight=memory_info#psutil.Process.memory_full_info 

    max_iterations : int
        Limits the number of iterations (calls to the process being monitored). Relevant
        when the process is a python function.

    Returns
    -------
    mem_usage : list of floating-point values