{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## How much faster vectorized computation with Numpy is comparing to loops in python?\n",
    "\n",
    "### The notebook / example from:\n",
    "https://python.plainenglish.io/vectorized-and-non-vectorized-mathematical-computations-a9f7a7846a01\n",
    "\n",
    "### Copyright belongs to the originial author of this post.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import time\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "x1 = np.random.randint(low=11, high=99, size=100000)\n",
    "x2 = np.random.randint(low=11, high=99, size=100000)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Comparison about he dot product of two vectors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Computation time =  66.12992286682129\n",
      "Computation time =  0.1919269561767578\n"
     ]
    }
   ],
   "source": [
    "# CLASSIC DOT VECTORS IMPLEMENTATION\n",
    "tic = time.time()\n",
    "dot = 0\n",
    "for i in range(len(x1)):\n",
    "    dot+= x1[i]*x2[i]\n",
    "toc = time.time()\n",
    "print(\"Computation time = \",1000*(toc - tic))\n",
    "\n",
    "# VECTORIZED DOT IMPLEMENTATION\n",
    "tic = time.time()\n",
    "dot = np.dot(x1,x2)\n",
    "toc = time.time()\n",
    "print(\"Computation time = \",1000*(toc - tic))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### How about matrix-vector multiplication？"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Computation time =  286.9250774383545\n",
      "Computation time =  0.9779999999999234\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# CLASSIC GENERAL DOT IMPLEMENTATION\n",
    "W = np.random.rand(3,len(x1)) # Random 3*len(x1) numpy array\n",
    "tic = time.time()\n",
    "gdot = np.zeros(W.shape[0])\n",
    "for i in range(W.shape[0]):\n",
    "    for j in range(len(x1)):\n",
    "        gdot[i] += W[i,j]*x1[j]\n",
    "toc = time.time()\n",
    "print(\"Computation time = \",1000*(toc - tic))\n",
    "\n",
    "# VECTORIZED GENERAL DOT IMPLEMENTATION\n",
    "tic = time.process_time()\n",
    "dot = np.dot(W,x1)\n",
    "toc = time.process_time()\n",
    "print(\"Computation time = \",1000*(toc - tic))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}