Benchmarking some code from the numba mailing list

numba_bench.ipynb

{
 "metadata": {
  "name": ""
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Comparison of pure python, autojit and jit versions of the `legal` method"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numba\n",
      "from numba import bool_, list_\n",
      "\n",
      "dim = 3\n",
      "d = 3\n",
      "o = [[[False]*d for _ in range(d)] for _ in range(d)]\n",
      "\n",
      "def legal(i,j,k,mat):\n",
      "    return 0<=i and 0<=j and 0<=k and i<dim and j<dim and k<dim and not(mat[i][j][k])\n",
      "\n",
      "aj_legal = numba.autojit()(legal)\n",
      "j_legal = numba.jit(numba.bool_(numba.i8, numba.i8, numba.i8, list_(list_(list_(bool_, 3), 3), 3)))(legal)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 11
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%timeit legal(0,1,0,o)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "1000000 loops, best of 3: 594 ns per loop\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%timeit aj_legal(0,1,0,o)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "1 loops, best of 3: 8.11 \u00b5s per loop\n"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%timeit j_legal(0,1,0,o)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "1000000 loops, best of 3: 510 ns per loop\n"
       ]
      }
     ],
     "prompt_number": 15
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print legal(0,1,0,o)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "True\n"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print aj_legal(0,1,0,o)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "True\n"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print j_legal(0,1,0,o)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "True\n"
       ]
      }
     ],
     "prompt_number": 14
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Now let's consider a case where `o` is a numpy array"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numpy as np"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 17
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "o_arr = np.array(o, dtype=np.bool)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 18
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def legal_arr(i,j,k,mat):\n",
      "    return 0<=i and 0<=j and 0<=k and i<dim and j<dim and k<dim and not(mat[i,j,k])\n",
      "\n",
      "aj_legal_arr = numba.autojit()(legal_arr)\n",
      "j_legal_arr = numba.jit(numba.bool_(numba.i8, numba.i8, numba.i8, numba.bool_[:,:,:]))(legal_arr)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 21
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%timeit legal_arr(0,1,0,o_arr)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "1000000 loops, best of 3: 695 ns per loop\n"
       ]
      }
     ],
     "prompt_number": 22
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%timeit aj_legal_arr(0,1,0,o_arr)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "1 loops, best of 3: 6.91 \u00b5s per loop\n"
       ]
      }
     ],
     "prompt_number": 23
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%timeit j_legal_arr(0,1,0,o_arr)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "1000000 loops, best of 3: 320 ns per loop\n"
       ]
      }
     ],
     "prompt_number": 24
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print legal_arr(0,1,0,o_arr)\n",
      "print aj_legal_arr(0,1,0,o_arr)\n",
      "print j_legal_arr(0,1,0,o_arr)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "True\n",
        "True\n",
        "True\n"
       ]
      }
     ],
     "prompt_number": 25
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}