.. role:: python(code)
   :language: python


.. index::
   profiling; memory

..
   `profiling data size <ex:data_size>`_  
   `memory usage <ex:memory_usage>`_

.. _`sec:mem`:

Memory profiling
----------------

Memory profiling is much more cumbersome than time profiling. 

+ :python:`memory_profiler` (external dependency) allows a line by line memory usage analysis of decorated functions.
+ :python:`tracemalloc` (standard) for manual instrumentation


Using the :python:`memory_profiler.profile` decorator, a console output is produced, for example:

.. code-block:: python

   from memory_profiler import profile

   @profile
   def task_2(size: int) -> None:
      _ = [0] * size


.. code-block:: console
   
   Line #    Mem usage    Increment  Occurrences   Line Contents
   =============================================================
    24    150.0 MiB    150.0 MiB           1   @profile
    25                                         def task_2(size: int) -> None:
    26    211.1 MiB     61.0 MiB           1       _ = [0] * size


.. admonition:: Exercise (easy)

   What was the ``size`` value for the example above to make sense?

.. _`ex:data_size`:   

.. collapse:: Correction

   An int is 8B, a MB is :math:`1024^2 B`, :math:`61\times1024^2/8\approx8` million ints. 



To measure a specific memory usage variation :

.. code-block:: python
   :linenos:

   tracemalloc.start()
   _, peak_before = tracemalloc.get_traced_memory()
   # do something
   _, peak_after = tracemalloc.get_traced_memory()

.. admonition:: Exercise (easy)

   Verify the :python:`memory_profiler` values using :python:`tracemalloc` on ``task_2``.

.. _`ex:memory_usage`:  

.. collapse:: Correction

   .. code-block:: python
      :linenos:

      tracemalloc.start()
      _, peak_before = tracemalloc.get_traced_memory()
      task_2(8_000_000)
      _, peak_after = tracemalloc.get_traced_memory()
      print((peak_after - peak_before) / (1024 * 1024)) # 61 MB