Guide for using the python linux process memory profiler

Py_Linux_Memory_Profiler.md

Measure memory of a Linux process in Python

The module (found here)

from memorymeasure import get_memory

provides a single function get_memory which returns the current and peak, real and virtual memories (in bytes, accurate to ~kilobytes) used by your Python code's Linux process, in a dictionary. The same module but in C can be found here.

Calling

get_memory()

returns a dictionary with keys VmHWM, VmPeak, VmRSS and VmSize, e.g.

{
    'VmHWM': 6436000, 
    'VmPeak': 122260000, 
    'VmRSS': 6436000, 
    'VmSize': 122244000
}

though will raise a FileNotFoundError if ran on a non-linux machine.

---

Real memory (resident set size) is the amount of physical RAM your process is using, and virtual memory is the size of the memory address space your process is using. Linux chooses what in your virtual memory gets to reside in RAM. Note that in addition to your program data, these memories include the space taken up by your code itself, and any libraries your code is using (which may be shared by other running processes, skewing your usage). A good explanation can be found here.

Peak memory is the maximum amount of memory your process has used over its lifetime so far.

---

This module works by reading /proc/self/status which linux populates with info about your running process. It uses fields

Output	status Field	Description
currRealMem	VmRSS	Resident set size
peakRealMem	VmHWM	Peak resident set size ("high water mark")
currVirtMem	VmSize	Virtual memory size
peakVirtMem	VmPeak	Peak virtual memory size

taken from under /proc/[pid]/status on the proc man page.