1-Memory-Profilers.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#Introduction to Memory Profiling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Objectives:\n",
    "> * Be introduced to memory profiling using different tools\n",
    "> * Some small introduction to time profiling in IPython too\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##ipython_memwatcher"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our recommended way to profile memory consumption for this tutorial will be [ipython_memwatcher](https://pypi.python.org/pypi/ipython_memwatcher):\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [1] used 0.035 MiB RAM in 0.002s, peaked 0.000 MiB above current, total RAM usage 28.773 MiB\n"
     ]
    }
   ],
   "source": [
    "from ipython_memwatcher import MemWatcher\n",
    "mw = MemWatcher()\n",
    "mw.start_watching_memory()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [2] used 33.234 MiB RAM in 0.142s, peaked 7.324 MiB above current, total RAM usage 62.008 MiB\n"
     ]
    }
   ],
   "source": [
    "# Let's create a big object\n",
    "a = [i for i in range(1000*1000)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Measurements(memory_delta=33.234375, time_delta=0.1421198844909668, memory_peak=7.32421875, memory_usage=62.0078125)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [3] used -1.949 MiB RAM in 0.021s, peaked 0.000 MiB above current, total RAM usage 60.059 MiB\n"
     ]
    }
   ],
   "source": [
    "# Get some measurements from the last executed cell:\n",
    "meas = mw.measurements\n",
    "meas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "33.234375"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [4] used 0.016 MiB RAM in 0.006s, peaked 0.000 MiB above current, total RAM usage 60.074 MiB\n"
     ]
    }
   ],
   "source": [
    "# MemWatcher.measurements is a named tuple.  We can easily get info out of it:\n",
    "meas.memory_delta"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "34.848768"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [5] used 0.004 MiB RAM in 0.005s, peaked 0.000 MiB above current, total RAM usage 60.078 MiB\n"
     ]
    }
   ],
   "source": [
    "# This takes betweed 32 ~ 35 bytes per element:\n",
    "meas.memory_delta * (2**20) / 1e6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "int"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [6] used 0.000 MiB RAM in 0.006s, peaked 0.000 MiB above current, total RAM usage 60.078 MiB\n"
     ]
    }
   ],
   "source": [
    "# What are these elements made from?\n",
    "type(a[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "24"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [7] used 0.004 MiB RAM in 0.006s, peaked 0.000 MiB above current, total RAM usage 60.082 MiB\n"
     ]
    }
   ],
   "source": [
    "# How much memory take an int?\n",
    "# Beware: the size below will depend on whether you are using a 32-bit or 64-bit Python\n",
    "import sys\n",
    "sys.getsizeof(a[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ok.  On 64-bits, that means that the int object allocates 8 bytes for the integer value, and 16 bytes for other metadata (Python object).  But 24 is quite less than 32~35.  Where this overhead comes from?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##objgraph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Graph written to /tmp/objgraph-gLJZaX.dot (4 nodes)\n",
      "Image generated as simple-list.png\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARMAAACgCAYAAAAvgIjXAAAABmJLR0QA/wD/AP+gvaeTAAAR5UlE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\nAvncZ0JEqmCZEJEqWCZEpIr/AWZ2VGRkWctBAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [8] used 0.066 MiB RAM in 0.094s, peaked 0.000 MiB above current, total RAM usage 60.148 MiB\n"
     ]
    }
   ],
   "source": [
    "# Let's introduce the objgraph package and see\n",
    "b = [1,2,3]\n",
    "import objgraph\n",
    "objgraph.show_refs([b], filename='simple-list.png')\n",
    "from IPython.core.display import Image \n",
    "Image('simple-list.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, the list is an structure that takes a pointer (8 bytes in 64-bit platforms) per every element in the list.  If we add this to the 24 bytes per int, then we have 32 bytes per element, which is close to the computed 32~35 bytes above.  The remaining difference is probably due to how Python handles memory internally (over-allocation). "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##memory_profiler"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[memory_profiler](https://pypi.python.org/pypi/memory_profiler) is a basic module for memory profiling that many others use (like the `ipython_memwatcher` above) and it interacts well with ipython, so it is worth to see how it works:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [9] used 0.113 MiB RAM in 0.003s, peaked 0.000 MiB above current, total RAM usage 60.262 MiB\n"
     ]
    }
   ],
   "source": [
    "%load_ext memory_profiler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "peak memory: 132.43 MiB, increment: 72.16 MiB\n",
      "In [10] used 72.164 MiB RAM in 0.435s, peaked 0.000 MiB above current, total RAM usage 132.426 MiB\n"
     ]
    }
   ],
   "source": [
    "# Use %memit magic command exposed by memory_profiler\n",
    "%memit b = [i for i in range(1000*1000)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Please note that the `peak_memory` in this case is different than the `peaked_memory` reported by ipython_memwatcher package."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##Guppy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Guppy is nice for having an overview of how different structures are using our memory:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Partition of a set of 2128439 objects. Total size = 81725384 bytes.\n",
       " Index  Count   %     Size   % Cumulative  % Kind (class / dict of class)\n",
       "     0 2004742  94 48113808  59  48113808  59 int\n",
       "     1   1180   0 16434528  20  64548336  79 list\n",
       "     2  57604   3  5178288   6  69726624  85 str\n",
       "     3  30885   1  2677600   3  72404224  89 tuple\n",
       "     4   1745   0  1713944   2  74118168  91 dict (no owner)\n",
       "     5    436   0  1256416   2  75374584  92 dict of module\n",
       "     6   7828   0  1001984   1  76376568  93 types.CodeType\n",
       "     7   7722   0   926640   1  77303208  95 function\n",
       "     8    943   0   849784   1  78152992  96 type\n",
       "     9    943   0   779752   1  78932744  97 dict of type\n",
       "<562 more rows. Type e.g. '_.more' to view.>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [11] used 31.668 MiB RAM in 3.135s, peaked 0.000 MiB above current, total RAM usage 164.094 MiB\n"
     ]
    }
   ],
   "source": [
    "from guppy import hpy; hp=hpy()\n",
    "hp.heap()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Partition of a set of 1128716 objects. Total size = 49609424 bytes.\n",
       " Index  Count   %     Size   % Cumulative  % Kind (class / dict of class)\n",
       "     0 1005000  89 24120000  49  24120000  49 int\n",
       "     1   1177   0  8307848  17  32427848  65 list\n",
       "     2  57609   5  5178584  10  37606432  76 str\n",
       "     3  30885   3  2677624   5  40284056  81 tuple\n",
       "     4   1751   0  1715624   3  41999680  85 dict (no owner)\n",
       "     5    436   0  1256416   3  43256096  87 dict of module\n",
       "     6   7828   1  1001984   2  44258080  89 types.CodeType\n",
       "     7   7721   1   926520   2  45184600  91 function\n",
       "     8    943   0   849784   2  46034384  93 type\n",
       "     9    943   0   779752   2  46814136  94 dict of type\n",
       "<558 more rows. Type e.g. '_.more' to view.>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [12] used 37.398 MiB RAM in 0.766s, peaked 0.000 MiB above current, total RAM usage 201.492 MiB\n"
     ]
    }
   ],
   "source": [
    "del b\n",
    "hp.heap()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Partition of a set of 1 object. Total size = 8126536 bytes.\n",
       " Index  Count   %     Size   % Cumulative  % Kind (class / dict of class)\n",
       "     0      1 100  8126536 100   8126536 100 list"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [13] used 0.023 MiB RAM in 0.009s, peaked 0.000 MiB above current, total RAM usage 201.516 MiB\n"
     ]
    }
   ],
   "source": [
    "# Size of the list (beware, this does not include the contents!)\n",
    "hp.iso(a)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "##%time and %timeit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 18.5 ms, sys: 264 µs, total: 18.8 ms\n",
      "Wall time: 22 ms\n",
      "In [14] used 0.016 MiB RAM in 0.027s, peaked 0.000 MiB above current, total RAM usage 201.531 MiB\n"
     ]
    }
   ],
   "source": [
    "# IPython provides a magic command to see how much time a command takes to run\n",
    "%time asum = sum(a)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that `%time` offers quite detailed statistics on the time spent.\n",
    "\n",
    "Also, the time reported by MemoryWatcher has a typical overhead of 3~10 ms over the time reported by %time, so when the times to measure are about this order then it is better to rely on the %time (or %timeit below) values.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100 loops, best of 3: 6.95 ms per loop\n",
      "In [15] used 0.020 MiB RAM in 2.970s, peaked 0.000 MiB above current, total RAM usage 201.551 MiB\n"
     ]
    }
   ],
   "source": [
    "# We have another way to measure timings doing several loops and getting the mean\n",
    "%timeit bsum = sum(a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'bsum' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-16-24d8d3d27b48>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;31m# However, one must notice that %timeit does not return the result of expressions\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m \u001b[0mbsum\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'bsum' is not defined"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [16] used 0.742 MiB RAM in 0.089s, peaked 0.000 MiB above current, total RAM usage 202.293 MiB\n"
     ]
    }
   ],
   "source": [
    "# However, one must notice that %timeit does not return the result of expressions\n",
    "bsum"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Interestingly, %timeit allows to retrieve the measured times in loops with the -o flag:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100 loops, best of 3: 6.94 ms per loop\n",
      "[0.7002749443054199, 0.6935679912567139, 0.6935369968414307]\n",
      "0.00693536996841\n",
      "In [17] used 0.000 MiB RAM in 2.948s, peaked 0.000 MiB above current, total RAM usage 202.293 MiB\n"
     ]
    }
   ],
   "source": [
    "t = %timeit -o sum(a)\n",
    "print(t.all_runs)\n",
    "print(t.best)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And one can specify the number of loops (-n) and the number of repetitions (-r):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 loops, best of 1: 17.6 ms per loop\n",
      "[0.017561912536621094]\n",
      "0.0175619125366\n",
      "In [18] used 0.000 MiB RAM in 0.024s, peaked 0.000 MiB above current, total RAM usage 202.293 MiB\n"
     ]
    }
   ],
   "source": [
    "t = %timeit -r1 -n1 -o sum(a)\n",
    "print(t.all_runs)\n",
    "print(t.best)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise 1\n",
    "\n",
    "Provided a dictionary like:\n",
    "\n",
    "```\n",
    "d = dict((\"key: %i\"%i, i*2) for i in a)\n",
    "```\n",
    "\n",
    "Try to guess how much RAM it uses using the techniques introduced above.\n",
    "\n",
    "Why do you think it takes more space than a list?\n",
    "\n",
    "*Hint*: Use the `objgraph` package on a short dictionary so as to better see the data structure.  In case you cannot get `objgraph` to work, every entry in a dictionary has pointers to two objects: key and value. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "##Solution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [19] used 94.137 MiB RAM in 0.967s, peaked 0.000 MiB above current, total RAM usage 296.430 MiB\n"
     ]
    }
   ],
   "source": [
    "d = dict((\"key: %i\"%i, i*2) for i in a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Graph written to /tmp/objgraph-gnyjwx.dot (11 nodes)\n",
      "Image generated as dict.png\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABWcAAADICAIAAAALPsTXAAAABmJLR0QA/wD/AP+gvaeTAAAgAElE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\nsv0fAMg+0jlI5RAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [22] used 0.000 MiB RAM in 1.473s, peaked 0.004 MiB above current, total RAM usage 296.828 MiB\n"
     ]
    }
   ],
   "source": [
    "objgraph.show_refs([d], filename='dict.png')\n",
    "from IPython.core.display import Image \n",
    "Image('dict.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "72"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [36] used 0.000 MiB RAM in 0.006s, peaked 0.000 MiB above current, total RAM usage 296.836 MiB\n"
     ]
    }
   ],
   "source": [
    "# Compute the size of key + value\n",
    "sys.getsizeof(\"key: 100000\") + sys.getsizeof(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "83.92333984375"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [38] used 0.000 MiB RAM in 0.008s, peaked 0.000 MiB above current, total RAM usage 296.836 MiB\n"
     ]
    }
   ],
   "source": [
    "# Compute the size of a key + value + pointers for all elements in MB\n",
    "(sys.getsizeof(\"key: 100000\") + sys.getsizeof(1) + 16) * 1000*1000 / 2**20."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Partition of a set of 1 object. Total size = 50331928 bytes.\n",
       " Index  Count   %     Size   % Cumulative  % Kind (class / dict of class)\n",
       "     0      1 100 50331928 100  50331928 100 dict (no owner)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In [33] used 0.000 MiB RAM in 0.689s, peaked 0.000 MiB above current, total RAM usage 296.836 MiB\n"
     ]
    }
   ],
   "source": [
    "# Using guppy\n",
    "hp.iso(d)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, guppy is telling us that just the dictionary structure is taking ~50 MB, whereas the contents alone are taking ~70MB, so we should have expected the dictionary to consume ~120 MB.  However, our `MemWatcher` instance is reporting just 94 MB.  Take away lesson: measuring memory consumption in Python is tricky but using proper tools we can still have valuable hints!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}