xalan_copy.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "States and functions instantiated\n"
     ]
    }
   ],
   "source": [
    "from random import *\n",
    "import numpy as np\n",
    "from scipy.stats import truncnorm\n",
    "from operator import add\n",
    "\n",
    "seed (100)\n",
    "\n",
    "## Global variables\n",
    "global N, NUM_OF_ITERATION\n",
    "global STATE0, STATE1, STATE2, STATE3, STATE4, STATE5, STATE6, STATE7, STATE8, STATE9, STATE10, STATE11\n",
    "global BASE, OPT\n",
    "global TDM, GREEDY, GENERIC\n",
    "global iCache, dCache\n",
    "global PROB\n",
    "\n",
    "global TDM_OPT, GREEDY_OPT, GENERIC_OPT\n",
    "global OPT_LIST\n",
    "global METHOD\n",
    "global STATE_COUNT\n",
    "\n",
    "## Instantiating global variables\n",
    "## N is the total number of cycles the process runs for\n",
    "N = 100000000\n",
    "\n",
    "## Number of iterations for the complete run\n",
    "## 10000 takes around 2 min to run\n",
    "NUMBER_OF_ITERATION = 2\n",
    "\n",
    "## BASE list has all the base cycle values that were generated\n",
    "## OPT list has the optimized values for that cycle post optimization\n",
    "BASE = []\n",
    "OPT = []\n",
    "\n",
    "## States\n",
"STATE0 = {'meanCycle': 90000, 'stdDev': 1000, 'min': 85000, 'max': 95000, 'base': BASE, 'opt': OPT}\n",
"STATE1 = {'meanCycle': 500000, 'stdDev': 10000, 'min': 450000, 'max': 550000, 'base': BASE, 'opt': OPT}\n",
"STATE2 = {'meanCycle': 200000, 'stdDev': 10000, 'min': 150000, 'max': 250000, 'base': BASE, 'opt': OPT}\n",
"STATE3 = {'meanCycle': 100000, 'stdDev': 1000, 'min': 95000, 'max': 105000, 'base': BASE, 'opt': OPT}\n",
"STATE4 = {'meanCycle': 200000, 'stdDev': 10000, 'min': 150000, 'max': 250000, 'base': BASE, 'opt': OPT}\n",
"STATE5 = {'meanCycle': 500000, 'stdDev': 10000, 'min': 450000, 'max': 550000, 'base': BASE, 'opt': OPT}\n",
"STATE6 = {'meanCycle': 150000, 'stdDev': 10000, 'min': 100000, 'max': 200000, 'base': BASE, 'opt': OPT}\n",
"STATE7 = {'meanCycle': 200000, 'stdDev': 10000, 'min': 150000, 'max': 250000, 'base': BASE, 'opt': OPT}\n",
"STATE8 = {'meanCycle': 60000, 'stdDev': 1000, 'min': 55000, 'max': 65000, 'base': BASE, 'opt': OPT}\n",
"STATE9 = {'meanCycle': 200000, 'stdDev': 10000, 'min': 150000, 'max': 250000, 'base': BASE, 'opt': OPT}\n",
"STATE10 = {'meanCycle': 200000, 'stdDev': 10000, 'min': 150000, 'max': 250000, 'base': BASE, 'opt': OPT}\n",
"STATE11 = {'meanCycle': 80000, 'stdDev': 1000, 'min': 75000, 'max': 85000, 'base': BASE, 'opt': OPT}\n",
    "\n",
    "## name key added to ##M print the name\n",
    "TDM = {'name': 'TDM', 'freqSampling': 2000, 'lat': 500, 'stdDev': 20, 'quality': 0.7}\n",
    "GREEDY = {'name': 'GREEDY', 'freqSampling': 25000, 'lat': 1000, 'stdDev': 100, 'quality': 0.9}\n",
    "GENERIC = {'name': 'GENERIC', 'freqSampling': 50000, 'lat': 2000, 'stdDev': 500, 'quality': 1}\n",
    "\n",
    "## CycleRatio ==> is the percentage of cycles that the technique uses from the total cycles\n",
    "iCache = {'name': 'iCahche', 'lat': 250, 'stdDev': 10, 'quality': 0.2, 'cycleRatio': 0.67}\n",
    "dCache = {'name': 'dCache', 'lat': 400, 'stdDev': 50, 'quality': 0.3, 'cycleRatio': 0.11}\n",
    "ROB = {'name': 'ROB', 'lat': 300, 'stdDev': 50, 'quality': 0.1, 'cycleRatio': 0.11}\n",
    "IQ = {'name': 'IQ', 'lat': 400, 'stdDev': 50, 'quality': 0.1, 'cycleRatio': 0.11}\n",
    "\n",
    "## techList is the list containing all the techniques \n",
    "## it is fed into the getOptCycles to get the configuration time for the max of the techniques\n",
    "techList = [iCache, dCache, ROB, IQ]\n",
    "\n",
    "## Clearing the global opt lists\n",
    "TDM_OPT = []\n",
    "GREEDY_OPT = []\n",
    "GENERIC_OPT = []\n",
    "\n",
    "## Create OPT_LIST contains the speedup for all the variations of techniques\n",
    "OPT_LIST = [TDM_OPT, GREEDY_OPT, GENERIC_OPT]\n",
    "\n",
    "## Create a list of methods, do not create strings 'TDM' XX\n",
    "METHOD = [TDM, GREEDY, GENERIC]\n",
    "\n",
    "## State Count \n",
    "STATE_COUNT = {'state0': 0, 'state1': 0, 'state2': 0, 'state3': 0, 'state4': 0, 'state5': 0, 'state6': 0,\\\n",
    "              'state7': 0, 'state8': 0, 'state9': 0, 'state10': 0, 'state11': 0}\n",
    "## This list is not need as we have encoded the techniques in \n",
    "## the getOptCycles using the getActiveTech function\n",
    "#TECH = [iCache, dCache]\n",
    "\n",
    "## logFile contains the information about the runs\n",
    "logFile = 'xalan_trial2_1.txt'\n",
    "\n",
    "## imgFile points to the image file name for the run\n",
    "imgFile = 'xalan_trial2_1.png'\n",
    "\n",
    "## --------------------------------------------------------------- ##\n",
    "\n",
    "def state_0():\n",
    "    ##Dprint \"State_0\"\n",
    "    STATE_COUNT['state0'] = STATE_COUNT['state0'] + 1\n",
    "    \n",
    "    ## Lets take that the mean values of cycles in state0 is 50000 and std Dev is 5000\n",
    "    global STATE0, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE0['meanCycle'], STATE0['stdDev'],\\\n",
    "                                         STATE0['min'], STATE0['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State0 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "    ##Dprint \"Opt Cycle Satrt: 0\\n\"\n",
    "    ## For the first iteration we need to start with time=0\n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         ## startCycles for next iteration is the baseCycle from previous state\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    ## \n",
    "    #for i in range(len(BASE)):\n",
    "        ###M print \"BASE[%d] = %f\" %(i, BASE[i])\n",
    "    \n",
    "    ## ------------- Wrong way to code ---------------##\n",
    "#     print \"Before Entering if loop: State0\\n\"\n",
    "#     if random() <= 0.5:\n",
    "#         print\"State0_1\\n\"\n",
    "#         return state_1\n",
    "#     elif random() > 0.5:\n",
    "#         print\"State0_0\\n\"\n",
    "#         return state_0\n",
    "#     else: ## else is very important, otherwise the code is breaking even if you consider all the scenarios\n",
    "#         print\"Default: 0\\n\"\n",
    "#         return state_0\n",
    "#     elif random() > 0.6:\n",
    "#         print\"State0_3\\n\"\n",
    "#         return state_3\n",
    "\n",
    "    ## ------------------- Correct Way ------------------##\n",
    "    rand = random()\n",
    "#     if rand <= 0.4:\n",
    "#         return state_0\n",
    "#     elif rand > 0.4 and rand <= 0.8:\n",
    "#         return state_1\n",
    "#     elif rand > 0.8 and rand <= 0.9:\n",
    "#         return state_2\n",
    "#     elif rand > 0.9:\n",
    "#         return state_3\n",
    "    \n",
    "    return returnState(rand)\n",
    "    \n",
    "def state_1():\n",
    "    ##Dprint \"State_1\"\n",
    "    STATE_COUNT['state1'] = STATE_COUNT['state1'] + 1\n",
    "    ##\n",
    "    global STATE1, BASE, OPT\n",
    "    \n",
    "    ## baseCycles for the state at the current execution time\n",
    "    ## random number from the normal distribution with mean=STATE1['meanCycle']\n",
    "    ## and stdDev = STATE1['stdDev']\n",
    "    #baseCycles = round(np.random.normal(STATE1['meanCycle'], STATE1['stdDev']))\n",
    "    baseCycles = round(getTruncatedNormal(STATE1['meanCycle'], STATE1['stdDev'],\\\n",
    "                                         STATE1['min'], STATE1['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State1 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    ##Dprint \"Opt Cycle Satrt: 1\\n\"\n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    #for i in range(len(BASE)):\n",
    "        ###M print \"BASE[%d]= %f\" %(i, BASE[i])\n",
    "    \n",
    "    ##Dprint \"Before Entering if loop: State1\\n\"    \n",
    "    ## Do not create random number in the comparison \n",
    "    ## create a variable with random() and then check for the condition\n",
    "    ## if you do if random() elif random() => these are both generating different random numbers\n",
    "    ## and can break your code even if you consider all the possibilities, since each number get generated \n",
    "    ## independently of each other.\n",
    "    ## Wrong way to code\n",
    "#     if random() <= 0.5:\n",
    "#         print\"State1_0\\n\"\n",
    "#         return state_0\n",
    "#     elif random() > 0.5:\n",
    "#         print\"State1_1\\n\"\n",
    "#         return state_1\n",
    "#     else:\n",
    "#         print\"Default: 1\\n\"\n",
    "#         return state_1\n",
    "    \n",
    "    rand = random()\n",
    "#     if rand <= 0.4:\n",
    "#         return state_0\n",
    "#     elif rand > 0.4 and rand <= 0.8:\n",
    "#         return state_1\n",
    "#     elif rand > 0.8 and rand <= 0.9:\n",
    "#         return state_2\n",
    "#     elif rand > 0.9:\n",
    "#         return state_3\n",
    "\n",
    "    return returnState(rand)\n",
    "    \n",
    "def state_2():\n",
    "    ##M print \"State_2\"\n",
    "    ##\n",
    "    STATE_COUNT['state2'] = STATE_COUNT['state2'] + 1\n",
    "    global STATE2, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE2['meanCycle'], STATE2['stdDev'],\\\n",
    "                                         STATE2['min'], STATE2['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State2 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_3():\n",
    "    ##M print \"State_3\"\n",
    "    STATE_COUNT['state3'] = STATE_COUNT['state3'] + 1\n",
    "    ##\n",
    "    global STATE3, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE3['meanCycle'], STATE3['stdDev'],\\\n",
    "                                         STATE3['min'], STATE3['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_4():\n",
    "    ##M print \"State_4\"\n",
    "    STATE_COUNT['state4'] = STATE_COUNT['state4'] + 1\n",
    "    ##\n",
    "    global STATE4, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE4['meanCycle'], STATE4['stdDev'],\\\n",
    "                                         STATE4['min'], STATE4['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_5():\n",
    "    ##M print \"State_5\"\n",
    "    STATE_COUNT['state5'] = STATE_COUNT['state5'] + 1\n",
    "    ##\n",
    "    global STATE5, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE5['meanCycle'], STATE5['stdDev'],\\\n",
    "                                         STATE5['min'], STATE5['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_6():\n",
    "    ##M print \"State_6\"\n",
    "    STATE_COUNT['state6'] = STATE_COUNT['state6'] + 1\n",
    "    ##\n",
    "    global STATE6, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE6['meanCycle'], STATE6['stdDev'],\\\n",
    "                                         STATE6['min'], STATE6['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_7():\n",
    "    ##M print \"State_7\"\n",
    "    STATE_COUNT['state7'] = STATE_COUNT['state7'] + 1\n",
    "    ##\n",
    "    global STATE7, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE7['meanCycle'], STATE7['stdDev'],\\\n",
    "                                         STATE7['min'], STATE7['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_8():\n",
    "    ##M print \"State_8\"\n",
    "    STATE_COUNT['state8'] = STATE_COUNT['state8'] + 1\n",
    "    ##\n",
    "    global STATE8, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE8['meanCycle'], STATE8['stdDev'],\\\n",
    "                                         STATE8['min'], STATE8['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_9():\n",
    "    ##M print \"State_9\"\n",
    "    STATE_COUNT['state9'] = STATE_COUNT['state9'] + 1\n",
    "    ##\n",
    "    global STATE9, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE9['meanCycle'], STATE9['stdDev'],\\\n",
    "                                         STATE9['min'], STATE9['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_10():\n",
    "    ##M print \"State_10\"\n",
    "    STATE_COUNT['state10'] = STATE_COUNT['state10'] + 1\n",
    "    ##\n",
    "    global STATE6, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE10['meanCycle'], STATE10['stdDev'],\\\n",
    "                                         STATE10['min'], STATE10['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "def state_11():\n",
    "    ##M print \"State_11\"\n",
    "    STATE_COUNT['state11'] = STATE_COUNT['state11'] + 1\n",
    "    ##\n",
    "    global STATE11, BASE, OPT\n",
    "    \n",
    "    baseCycles = round(getTruncatedNormal(STATE11['meanCycle'], STATE11['stdDev'],\\\n",
    "                                         STATE11['min'], STATE11['max']))\n",
    "    BASE.append(baseCycles)\n",
    "    \n",
    "    ##MwriteStr = \"State3 \" + str(baseCycles) + \"\\n\"\n",
    "    ##Mf.write(writeStr)\n",
    "    \n",
    "#     if (len(BASE) == 1):\n",
    "#         optStateCycles = getOptCycles(baseCycles, 0, TDM, techList, getActiveTech(10))\n",
    "#     else:\n",
    "#         optStateCycles = getOptCycles(baseCycles, getTotalCycles(BASE, len(BASE)-1), TDM, techList, getActiveTech(10))\n",
    "    \n",
    "#     OPT.append(optStateCycles)\n",
    "    \n",
    "    rand = random()\n",
    "    return returnState(rand)\n",
    "\n",
    "\n",
    "## Define the state function return for each state beased on the pobabilities\n",
    "def returnState(randVar):\n",
    "    if randVar <= 0.06:\n",
    "        return state_0\n",
    "    elif randVar > 0.06 and randVar <= 0.13:\n",
    "        return state_1\n",
    "    elif randVar > 0.13 and randVar <= 0.22:\n",
    "        return state_2\n",
    "    elif randVar > 0.22 and randVar <= 0.35:\n",
    "        return state_3\n",
    "    elif randVar > 0.35 and randVar <= 0.41:\n",
    "        return state_4\n",
    "    elif randVar > 0.41 and randVar <= 0.54:\n",
    "        return state_5\n",
    "    elif randVar > 0.54 and randVar <= 0.57:\n",
    "        return state_6\n",
    "    elif randVar > 0.57 and randVar <= 0.70:\n",
    "        return state_7\n",
    "    elif randVar > 0.70 and randVar <= 0.80:\n",
    "        return state_8\n",
    "    elif randVar > 0.80 and randVar <= 0.84:\n",
    "        return state_9\n",
    "    elif randVar > 0.84 and randVar <= 0.98:\n",
    "        return state_10\n",
    "    elif randVar > 0.98 and randVar <= 1:\n",
    "        return state_11\n",
    "\n",
    "    \n",
    "## Other functions\n",
    "## get the techniques that are active for a given simulation\n",
    "def getActiveTech(x):\n",
    "    if (x == 0):\n",
    "        return (0, 0, 0, 0)\n",
    "    elif (x==1):\n",
    "        return (0, 0, 0, 1)\n",
    "    elif (x==2):\n",
    "        return (0, 0, 1, 0)\n",
    "    elif (x==3):\n",
    "        return (0, 0, 1, 1)\n",
    "    elif (x==4):\n",
    "        return (0, 1, 0, 0)\n",
    "    elif (x==5):\n",
    "        return (0, 1, 0, 1)\n",
    "    elif (x==6):\n",
    "        return (0, 1, 1, 0)\n",
    "    elif (x==7):\n",
    "        return (0, 1, 1, 1)\n",
    "    elif (x==8):\n",
    "        return (1, 0, 0, 0)\n",
    "    elif (x==9):\n",
    "        return (1, 0, 0, 1)\n",
    "    elif (x==10):\n",
    "        return (1, 0, 1, 0)\n",
    "    elif (x==11):\n",
    "        return (1, 0, 1, 1)\n",
    "    elif (x==12):\n",
    "        return (1, 1, 0, 0)\n",
    "    elif (x==13):\n",
    "        return (1, 1, 0, 1)\n",
    "    elif (x==14):\n",
    "        return (1, 1, 1, 0)\n",
    "    elif (x==15):\n",
    "        return (1, 1, 1, 1)\n",
    "\n",
    "###############################################\n",
    "\n",
    "## sum() => internal function can also be used\n",
    "## Usage sum(list)\n",
    "def getTotalCycles(list1, index):\n",
    "    tot = 0\n",
    "    for i in range(index):\n",
    "        tot = tot + list1[i]\n",
    "    return tot\n",
    "\n",
    "###############################################\n",
    "\n",
    "## getOptCycles takes in input basecycle, time at which the state started,\n",
    "## method and technique applied and returns the optimized time for that state, that method \n",
    "## and technique\n",
    "\n",
    "## baseCycles => is generated randomly based on mean and variance for a state\n",
    "## startCycles => is the time at which the previous state finished. Start of this state in globale\n",
    "## time\n",
    "## method => method, TDM, GREEDY, GENERIC\n",
    "## method['lat'] => gives the latency time for the method\n",
    "## techList => technique applied, iCache, dCache, ROB etc. Used to get config time for \n",
    "##          all the techniques and max will be used for process\n",
    "## activeTech ==> is the tuple with the active techniques \n",
    "\n",
    "def getOptCycles(baseCycles, startCycles, method, techList, activeTech):\n",
    "    ###M print \"method= %s\\n%s['freqSampling'] = %d\" %(method, method, method['freqSampling'])\n",
    "    ###M print \"Type of Method = %s\" %(type(method))\n",
    "    samplingRate = method['freqSampling']\n",
    "    \n",
    "    ## samplingTime is the time in the state where the technique will sample for optimisation\n",
    "    ## If the startCyles is 0 => Starting the simulation\n",
    "    ## startCycles is divisible by sampling rate => We discussed to go on to the next \n",
    "    ## sampling cycle\n",
    "    if (startCycles % samplingRate == 0 or startCycles == 0):\n",
    "        samplingTime = samplingRate\n",
    "    elif (startCycles % samplingRate != 0):\n",
    "        ## Since we know the starting time, we divide it by samplingRate to see how much \n",
    "        ## it is above the samplingRate. Now we subtract it from the samplingRate to get \n",
    "        ## how further down will the next sampling be done.\n",
    "        ## We multiply 2*samplingRate since the program will take at least freqSampling\n",
    "        ## to optimize.\n",
    "        #samplingTime = startCycles + (samplingRate - np.remainder(startCycles, samplingRate))\n",
    "        ## For this run we assume the sampling point can be the first sampling point and need \n",
    "        ## not wait for the next sampling point. One more assumption is that the latency \n",
    "        ## should account for the time it needs to configure the state variables for optimization\n",
    "        ## Because of thois we use 1*samplingRate - ....\n",
    "        samplingTime = (1*samplingRate) - np.remainder(startCycles, samplingRate)\n",
    "    \n",
    "    if (samplingTime > baseCycles):\n",
    "        ##M print \"Sampling Time greater than baseCycles ==> No optimisation possible\\n\"\n",
    "        ##M print \"baseCycles= %d\\nsamplingTime= %d\\n\" %(baseCycles, samplingTime)\n",
    "        return baseCycles\n",
    "    else:\n",
    "        ## globalSamplingTime => Gives the sampling time with respect to global time\n",
    "        globalSamplingTime = startCycles + samplingTime\n",
    "\n",
    "        ## optTime => Represents the time taken by method to optimize\n",
    "        optTime = round(np.random.normal(method['lat'], method['stdDev']))\n",
    "\n",
    "        ## configTimeList => Represents the list of time taken by technique to configure\n",
    "        #configTime = round(np.random.normal(tech['lat'], tech['stdDev']))\n",
    "        configTimeList = []\n",
    "        for ii in range(len(techList)):\n",
    "            configTimeList.append(round(np.random.normal(techList[ii]['lat'], techList[ii]['stdDev'])))\n",
    "        \n",
    "        ## configTime is the max of all the cpnfiguration times required, since they can happen in parallel\n",
    "        configTime = max(configTimeList)\n",
    "\n",
    "        ## timeleft => Represents the time left after sampling and optimisation and configuration time\n",
    "        timeLeft = baseCycles - (samplingTime + optTime + configTime)\n",
    "        if (timeLeft < 0):\n",
    "            ##M print \"Cannot optimize because there is no time left for optimisation\\n\"\n",
    "            ##M print \"timeLeft < 0\\n\"\n",
    "            ##M print \"baseCycles= %d\\nstartCycles= %d\\nsamplingTime= %d\\nglobalSamplingTime=%d\\noptTime= %d\\\n",
    "        ##M\\nconfigTime= %d\\ntimeLeft= %d\\n\"\\\n",
    "        ##M% (baseCycles, startCycles, samplingTime, globalSamplingTime, optTime, \\\n",
    "        ##M   configTime, timeLeft)\n",
    "            return baseCycles\n",
    "        \n",
    "        else:\n",
    "            ## this time is now optimized according to the method\n",
    "            ##opt_timeLeft = round(method['quality']*timeLeft)\n",
    "            a1 = activeTech[0]\n",
    "            a2 = activeTech[1]\n",
    "            a3 = activeTech[2]\n",
    "            a4 = activeTech[3]\n",
    "#             opt_timeLeft = round(timeLeft*method['quality']*(a1*iCache['cycleRatio']*(1-iCache['quality']) + \\\n",
    "#                                           a2*dCache['cycleRatio']*(1-dCache['quality']) + \\\n",
    "#                                           a3*ROB['cycleRatio']*(1-ROB['quality']) + \\\n",
    "#                                           a4*IQ['cycleRatio']*(1-IQ['quality']) \\\n",
    "#                                           ) + \\\n",
    "#                                 timeLeft*(1-(a1*iCache['cycleRatio'] +\\\n",
    "#                                            a2*dCache['cycleRatio'] +\\\n",
    "#                                            a3*ROB['cycleRatio'] +\\\n",
    "#                                            a4*IQ['cycleRatio'])\\\n",
    "#                                         )\\\n",
    "#                                 )\n",
    "            \n",
    "            opt_timeLeft = round(timeLeft*(a1*iCache['cycleRatio']*((1-iCache['quality']) + iCache['quality']*(1-method['quality'])) +\\\n",
    "                                          a2*dCache['cycleRatio']*((1-dCache['quality']) + dCache['quality']*(1-method['quality']) ) +\\\n",
    "                                          a3*ROB['cycleRatio']*((1-ROB['quality']) + ROB['quality']*(1-method['quality']) ) +\\\n",
    "                                          a4*IQ['cycleRatio']*((1-IQ['quality']) + IQ['quality']*(1-method['quality']) )  \n",
    "                                          ) + \\\n",
    "                                timeLeft*(1-(a1*iCache['cycleRatio'] +\\\n",
    "                                           a2*dCache['cycleRatio'] +\\\n",
    "                                           a3*ROB['cycleRatio'] +\\\n",
    "                                           a4*IQ['cycleRatio'])\\\n",
    "                                        )\\\n",
    "                                ) \n",
    "\n",
    "\n",
    "            newStateCycles = samplingTime + optTime + configTime + opt_timeLeft\n",
    "\n",
    "            ##M print \"\\nbaseCycles= %d\\nstartCycles= %d\\nsamplingTime= %d\\nglobalSamplingTime=%d\\noptTime= %d\\\n",
    "            ##M\\nconfigTime= %d\\ntimeLeft= %d\\nopt_timeLeft=%d\\nnewStateCycles= %d\\n\"\\\n",
    "            ##M% (baseCycles, startCycles, samplingTime, globalSamplingTime, optTime, \\\n",
    "            ##M   configTime, timeLeft, opt_timeLeft, newStateCycles)\n",
    "            \n",
    "            ##M print \"\\tiCache= %d, %d\\n\\tdCAche= %d, %d\\n\\tROB= %d, %d\\n\\tIQ= %d, %d\\n\" \\\n",
    "            ##M%(round(method['quality']*timeLeft*a1*iCache['cycleRatio']*(1-iCache['quality'])), round(timeLeft*iCache['cycleRatio']),\\\n",
    "            ##M  round(method['quality']*timeLeft*a2*dCache['cycleRatio']*(1-dCache['quality'])), round(timeLeft*dCache['cycleRatio']),\\\n",
    "            ##M  round(method['quality']*timeLeft*a3*ROB['cycleRatio']*(1-ROB['quality'])), round(timeLeft*ROB['cycleRatio']),\\\n",
    "            ##M  round(method['quality']*timeLeft*a4*IQ['cycleRatio']*(1-IQ['quality'])), round(timeLeft*IQ['cycleRatio'])\\\n",
    "            ##M )\n",
    "\n",
    "            ## returning the optimized time the state takes because of the technique and method applied\n",
    "            return newStateCycles\n",
    "\n",
    "#######################################################    \n",
    "\n",
    "## getTruncatedNormal gets random number by normal distribution within 'low' and 'up' given\n",
    "## the mean and standard deviation of the numbers.\n",
    "## rvs ==> is random variable continuous \n",
    "## Refer ==> https://stackoverflow.com/questions/36894191/how-to-get-a-normal-distribution-within-a-range-in-numpy\n",
    "\n",
    "def getTruncatedNormal (mean, stdDev, low, up):\n",
    "    return truncnorm ((low-mean)/stdDev, (up-mean)/stdDev, loc=mean, scale=stdDev).rvs()\n",
    "\n",
    "###M print \"truncated Normal number between %d and %d with mean=%d and stdDev = %d is %d\"\\\n",
    "#% (numMin, numMax, numMean, numSd, getTruncatedNormal(numMean, numSd, numMin, numMax).rvs() )    \n",
    "################\n",
    "\n",
    "print \"States and functions instantiated\"\n",
    "## ------------------------------------------------------------- ##\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "TOTAL_CYCLES= 100000000\n",
      "NUMBER_OF_ITERATION= 2\n",
      "TDM_OPT= [1.0, 1.0074133683141262, 1.0074134138986315, 1.0225741888815538, 1.0984684861394802, 1.0149378801260869, 1.0303272275968061, 1.030328681548103, 1.1074193732430728, 1.1074229074866588, 1.1257725417526956, 1.0381978414542341, 1.1165216033081475, 1.135171266914301, 1.1351757920534111, 1.1447386878469312]\n",
      "GREEDY_OPT= [1.0, 1.0070095015281528, 1.0070090356749399, 1.0213267179839998, 1.0926413553447638, 1.0141163771623585, 1.0286380244256221, 1.0286386376357499, 1.1010171862029954, 1.1010203383226016, 1.118156039744193, 1.0360564659879175, 1.1095216489820201, 1.1269239524348824, 1.1269254299453968, 1.1358326305147086]\n",
      "GENERIC_OPT= [1.0, 1.0055016120133082, 1.005502329310646, 1.0166935925992984, 1.0713808060265455, 1.0110597389793514, 1.0223742826997779, 1.0223723198656836, 1.0777265399395897, 1.0777227943000289, 1.0905616147075801, 1.0281213501799402, 1.0840949147625905, 1.097101535000311, 1.0970983522881026, 1.1038177741911563]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# # global variables\n",
    "# # create OPT list for each Method\n",
    "# from operator import add\n",
    "\n",
    "# global TDM_OPT, GREEDY_OPT, GENERIC_OPT\n",
    "# global OPT_LIST\n",
    "# global METHOD\n",
    "# global NUM_OF_ITERATION\n",
    "\n",
    "# TDM_OPT = []\n",
    "# GREEDY_OPT = []\n",
    "# GENERIC_OPT = []\n",
    "\n",
    "# ## Create OPT_LIST contains the speedup for all the variations of techniques\n",
    "# OPT_LIST = [TDM_OPT, GREEDY_OPT, GENERIC_OPT]\n",
    "\n",
    "# ## name key added to print the name\n",
    "# TDM = {'name': 'TDM', 'freqSampling': 2000, 'lat': 500, 'stdDev': 20, 'quality': 0.8}\n",
    "# GREEDY = {'name': 'GREEDY', 'freqSampling': 4000, 'lat': 1500, 'stdDev': 100, 'quality': 0.9}\n",
    "# GENERIC = {'name': 'GENERIC', 'freqSampling': 20000, 'lat': 5000, 'stdDev': 500, 'quality': 1}\n",
    "\n",
    "# ## CycleRatio ==> is the percentage of cycles that the technique uses from the total cycles\n",
    "# iCache = {'name': 'iCahche', 'lat': 250, 'stdDev': 10, 'quality': 0.2, 'cycleRatio': 0.6}\n",
    "# dCache = {'name': 'dCache', 'lat': 400, 'stdDev': 50, 'quality': 0.3, 'cycleRatio': 0.1}\n",
    "# ROB = {'name': 'ROB', 'lat': 300, 'stdDev': 50, 'quality': 0.1, 'cycleRatio': 0.1}\n",
    "# IQ = {'name': 'IQ', 'lat': 400, 'stdDev': 50, 'quality': 0.1, 'cycleRatio': 0.2}\n",
    "\n",
    "# ## techList is the list containing all the techniques \n",
    "# ## it is fed into the getOptCycles to get the configuration time for the max of the techniques\n",
    "# techList = [iCache, dCache, ROB, IQ]\n",
    "\n",
    "# ## Create a list of methods, do not create strings 'TDM' XX\n",
    "# METHOD = [TDM, GREEDY, GENERIC]\n",
    "\n",
    "# ## This list is not need as we have encoded the techniques in \n",
    "# ## the getOptCycles using the getActiveTech function\n",
    "# #TECH = [iCache, dCache]\n",
    "\n",
    "# ## Number of iterations for the complete run\n",
    "# ## 10000 takes around 2 min to run\n",
    "# NUMBER_OF_ITERATION = 10  \n",
    "\n",
    "###############################################################################\n",
    "## BASE is empty and and is will contain the base numbers of the states\n",
    "## optList has the empty list to which the optCycles will be appended\n",
    "## optList = [pdmList, greedyList, genericList]\n",
    "def runMCF(BASE, optList):\n",
    "    state = state_0\n",
    "\n",
    "    count = 0\n",
    "\n",
    "    while state:\n",
    "        state = state()\n",
    "        ## when the sum of base goes above N we stop\n",
    "        ## Stops close to the maximum number of cycles that we want to execute\n",
    "        if sum(BASE) > N:\n",
    "            ## state => None. while loop stops\n",
    "            ##Dprint\"sum(BASE)= %d, N= %d\\nBASE= %s\\n\" %(sum(BASE), N, BASE)\n",
    "            state = None\n",
    "    \n",
    "    ##Dprint\"\\nBASE2 = %s, Sum= %d\\n\" %(BASE, sum(BASE))\n",
    "    wrStr = \"Base Length: \" + str(len(BASE)) + \" , Base Sum: \" + str(sum(BASE)) + \"\\n\"\n",
    "    f.write(wrStr)\n",
    "    # ##M => Commented last to refrain from seeing all the debuggin outputs\n",
    "    ##Mprint \"\\ntype of BASE = %s\\n\" %(type(BASE))\n",
    "    #totalBaseCycles = getTotalCycles(BASE, len(BASE))\n",
    "\n",
    "    ##Mfor i in range(len(BASE)):\n",
    "    ##M    print \"BASE[%d]= %d, OPT[%d]=%d\" %(i, BASE[i], i, OPT[i])\n",
    "\n",
    "    #print \"\\nTotal Base Cycles = %d\" %(totalBaseCycles)\n",
    "\n",
    "    ### \n",
    "    #getOptCycles(BASE[1], BASE[0], TDM, iCache)\n",
    "\n",
    "    ##Mprint \"\\nTotal Base = %d, Total Opt = %d\\nTotal Speed Up = %f\" \\\n",
    "    ##M% (sum(BASE), sum(OPT), sum(BASE)/sum(OPT))\n",
    "\n",
    "    ############################################################\n",
    "\n",
    "    ## Once BASE is created, we can use it to get speedup for \n",
    "    ## various METHODS and TECHNIQUES\n",
    "\n",
    "    ## interList is the list to store intermediate optimisation cycles \n",
    "    ## for calculating the speedup \n",
    "    interList = []\n",
    "\n",
    "    ## groupList goes over the Methods in the order of activation of techniques\n",
    "    groupList = [0,1,2,4,8,3,5,6,9,10,12,7,11,13,14,15]\n",
    "\n",
    "    for i in range(len(METHOD)):\n",
    "        ## When we have a ist of list , the only way to makeit zero is by calling it by its name\n",
    "        ## pdmList = [] does not work even though pdmList is a list within optList\n",
    "        optList[i]=[]\n",
    "        ## This loop iterates over all combination of techniques\n",
    "        for j in groupList:\n",
    "            iterList = []\n",
    "            for b in range(len(BASE)):\n",
    "                #print \"Ooptimisation Method = %s , Tech = %s\" %(METHOD[j], TECH[i])\n",
    "                #print \"Type of method = %s type of TDM variable=%s\\nb=%d type of b=%s\" \\\n",
    "                #%(type(METHOD[j]), type(TDM), b, type(b))\n",
    "                if (b==0):\n",
    "                    optCycle = getOptCycles(BASE[b], getTotalCycles(BASE, 0), METHOD[i], techList, getActiveTech(j))\n",
    "                else:\n",
    "                    optCycle = getOptCycles(BASE[b], getTotalCycles(BASE, b), METHOD[i], techList, getActiveTech(j))\n",
    "\n",
    "                #optList = iterList\n",
    "                #print \"\\nname= %s, optList len = %d\\n\" %(METHOD_LIST[j], len(optList))\n",
    "                iterList.append(optCycle)\n",
    "\n",
    "\n",
    "            ##Mprint \"\\n############################\\n\"\n",
    "            ##Mprint \"Ooptimisation Method = %s , Tech = %d\" %(METHOD[i]['name'], j)\n",
    "            ##Mfor k in range(len(BASE)):\n",
    "            ##M    print \"BASE[%d]= %d, iterList[%d]=%d\" %(k, BASE[k], k, iterList[k])\n",
    "            ##Mprint \"\\nTotal Base = %d, Total Opt = %d\\nTotal Speed Up = %f\" \\\n",
    "            ##M% (sum(BASE), sum(iterList), sum(BASE)/sum(iterList))\n",
    "            ##Mprint \"\\n############################\\n\"\n",
    "\n",
    "            speedUp = sum(BASE)/sum(iterList)\n",
    "            #optList = OPT_LIST[i]\n",
    "            optList[i].append(speedUp)\n",
    "            ##Mprint \"optList[%d] = %s\" %(i, optList[i])\n",
    "\n",
    "    ## TDM_OPT and GREEDY_OPT will have the speedup's according to the activated techniques 0000, 0001, 0010..1111\n",
    "    ## Format the list to print only upto 4 decimal places\n",
    "    formatted_TDM_OPT = ['%.4f' % elem for elem in optList[0]]\n",
    "    formatted_GREEDY_OPT = ['%.4f' % elem for elem in optList[1]]\n",
    "    formatted_GENERIC_OPT = ['%.4f' % elm for elm in optList[2]]\n",
    "    ##Pprint \"TDM= %s\\nGREEDY= %s\\nGENERIC= %s\\n\" %(formatted_TDM_OPT, formatted_GREEDY_OPT, formatted_GENERIC_OPT)\n",
    "    \n",
    "    ## return optList with all the optimized lists for an iteration\n",
    "    return optList\n",
    "\n",
    "######## End of function runMCF\n",
    "## --------------------------------------------------------------- ##\n",
    "\n",
    "## These list contain the optCycle values for each method for each run\n",
    "pdmList = []\n",
    "greedyList = []\n",
    "genericList = []\n",
    "BASE = []\n",
    "TDM_OPT=[]\n",
    "GREEDY_OPT=[]\n",
    "GENERIC_OPT=[]\n",
    "\n",
    "##Mprint \"TDM_OPT= %s\\nGREEDY_OPT= %s\\nGENERIC_OPT= %s\\n\" %(TDM_OPT, GREEDY_OPT, GENERIC_OPT)\n",
    "\n",
    "optList = [pdmList, greedyList, genericList]\n",
    "##Mprint \"optList = %s\" %(optList)\n",
    "\n",
    "## ---------------------------------------------------------------- ##\n",
    "## opening file to write\n",
    "## Always close the file handle\n",
    "f = open(logFile, 'w')\n",
    "f.write(\"This is the first line\\n\")\n",
    "\n",
    "## ---------------------------------------------------------------- ##\n",
    "\n",
    "## ---------- Iterating over the complete process NUMBER_OF_ITERATION times ------------ ##\n",
    "for p in range(NUMBER_OF_ITERATION):\n",
    "    ##Mprint \"\\nBefore runFFT\\n\"\n",
    "    optListReturn = runMCF(BASE, optList)\n",
    "    ##Mprint \"\\nFinished runFFT\\n\"\n",
    "    ##Mprint\"\\nLoop Num= %d\" %(p)\n",
    "    \n",
    "    ## Create 3 list of pdm, greedy and generic from the returned list\n",
    "    ## Do not use the optReturn[0] to do anyoperation as it currupts the list\n",
    "    ## Always create separate lists fromt the list of list and then operate on it\n",
    "    list1 = optList[0][:]\n",
    "    list2 = optList[1][:]\n",
    "    list3 = optList[2][:]\n",
    "    \n",
    "    ##Mprint \"optList= %s\\n\\nlist1= %s\\n\\nlist2= %s\\n\\nlist3= %s\\n\" \\\n",
    "    ##M%(optList, list1, list2, list3)\n",
    "    ##Mprint \"optList= %s\\nlist1= %s\\nlist2= %s\\nlist3= %s\\n\" \\\n",
    "    ##M%(type(optList), type(list1), type(list2), type(list3))\n",
    "    \n",
    "#     list1 = [x*2 for x in list1]\n",
    "#     optList[0] = []\n",
    "#     print \"\\nlist1= %s\\n \\nTDM_OPT= %s\\n\\noptList= %s\" %(list1, TDM_OPT, optList)\n",
    "    # Creates a list of all state_count*meancycle values for each state\n",
    "    # np.sort sorts the the keys in ascending order, normally dict.keys() iterates randomly over keys\n",
    "    # Since sort uses 1st index, so putting state10 and state11 after state1\n",
    "    stateDictList = [STATE0, STATE1, STATE10, STATE11, STATE2, STATE3, STATE4, STATE5, STATE6, STATE7, STATE8, STATE9]\n",
    "    avgProductList = [stateDict['meanCycle']*STATE_COUNT[keyStateCount] \\\n",
    "                          for keyStateCount, stateDict in zip(np.sort(STATE_COUNT.keys()), stateDictList)]\n",
    "    # reduce uses the lambda function x+y, which iterates one by one over the entire \n",
    "    # stateCountList to give the sum.\n",
    "    stateProductTotal = reduce(lambda x,y: x+y, avgProductList)\n",
    "        \n",
    "    # The type becomes str when we use format\n",
    "    state0_percentage = \"{0:4.4f}\".format(avgProductList[0]*1.0/stateProductTotal)\n",
    "    state1_percentage = \"{0:4.4f}\".format(avgProductList[1]*1.0/stateProductTotal)\n",
    "    state10_percentage = \"{0:4.4f}\".format(avgProductList[2]*1.0/stateProductTotal)\n",
    "    state11_percentage = \"{0:4.4f}\".format(avgProductList[3]*1.0/stateProductTotal)\n",
    "    state2_percentage = \"{0:4.4f}\".format(avgProductList[4]*1.0/stateProductTotal)\n",
    "    state3_percentage = \"{0:4.4f}\".format(avgProductList[5]*1.0/stateProductTotal)\n",
    "    state4_percentage = \"{0:4.4f}\".format(avgProductList[6]*1.0/stateProductTotal)\n",
    "    state5_percentage = \"{0:4.4f}\".format(avgProductList[7]*1.0/stateProductTotal)\n",
    "    state6_percentage = \"{0:4.4f}\".format(avgProductList[8]*1.0/stateProductTotal)\n",
    "    state7_percentage = \"{0:4.4f}\".format(avgProductList[9]*1.0/stateProductTotal)\n",
    "    state8_percentage = \"{0:4.4f}\".format(avgProductList[10]*1.0/stateProductTotal)\n",
    "    state9_percentage = \"{0:4.4f}\".format(avgProductList[11]*1.0/stateProductTotal)\n",
    "    \n",
    "\n",
    "    \n",
    "    if (p==0):\n",
    "        ##Mprint \"loop0\"\n",
    "        TDM_OPT = list1[:]\n",
    "        GREEDY_OPT = list2[:]\n",
    "        GENERIC_OPT = list3[:]\n",
    "        \n",
    "        \n",
    "        f.write(\"State0: \" + str(STATE_COUNT['state0']) + \"(\" + state0_percentage + \")\" + \\\n",
    "                \" , State1: \" + str(STATE_COUNT['state1']) + \"(\" + state1_percentage + \")\" + \\\n",
    "               \" , State2: \" + str(STATE_COUNT['state2']) + \"(\" + state2_percentage + \")\" + \\\n",
    "                \" , State3: \" + str(STATE_COUNT['state3']) + \"(\" + state3_percentage + \")\" + \\\n",
    "                \", State4: \" + str(STATE_COUNT['state4']) + \"(\" + state4_percentage + \")\" + \\\n",
    "                \" , State5: \" + str(STATE_COUNT['state5']) + \"(\" + state5_percentage + \")\" + \\\n",
    "                \", State6: \" + str(STATE_COUNT['state6']) + \"(\" + state6_percentage + \")\" + \\\n",
    "                \" , State7: \" + str(STATE_COUNT['state7']) + \"(\" + state7_percentage + \")\" + \\\n",
    "                \" , State8: \" + str(STATE_COUNT['state8']) + \"(\" + state8_percentage + \")\" + \\\n",
    "                \", State9: \" + str(STATE_COUNT['state9']) + \"(\" + state9_percentage + \")\" + \\\n",
    "                \" , State10: \" + str(STATE_COUNT['state10']) + \"(\" + state10_percentage + \")\" + \\\n",
    "                \", State11: \" + str(STATE_COUNT['state11']) + \"(\" + state11_percentage + \")\" + \"\\n\")\n",
    "        f.write(\"TDM: \" + str(list1[len(list1)-1]) + \" , GREEDY: \" + str(list2[len(list2)-1]) + \\\n",
    "                \" , GENERIC: \" + str(list3[len(list3)-1]) + \"\\n\")\n",
    "        f.write(\"Iteration 1 complete\\n\\n\")\n",
    "    \n",
    "        ##Mprint \"\\nTDM_OPT= %s\\n\\nGREEDY_OPT= %s\\n\\nGENERIC_OPT= %s\" %(TDM_OPT, GREEDY_OPT, GENERIC_OPT)\n",
    "\n",
    "    else:\n",
    "       \n",
    "        ## Adding contents of the each iteration to its coressponding old value\n",
    "        ## This will be divided by the total number of iterations to get an average value\n",
    "        TDM_OPT = map(add, list1, TDM_OPT)\n",
    "        GREEDY_OPT = map(add, list2, GREEDY_OPT)\n",
    "        GENERIC_OPT = map(add, list3, GENERIC_OPT)\n",
    "        \n",
    "        f.write(\"State0: \" + str(STATE_COUNT['state0']) + \"(\" + state0_percentage + \")\" + \\\n",
    "                \" , State1: \" + str(STATE_COUNT['state1']) + \"(\" + state1_percentage + \")\" + \\\n",
    "               \" , State2: \" + str(STATE_COUNT['state2']) + \"(\" + state2_percentage + \")\" + \\\n",
    "                \" , State3: \" + str(STATE_COUNT['state3']) + \"(\" + state3_percentage + \")\" + \\\n",
    "                \", State4: \" + str(STATE_COUNT['state4']) + \"(\" + state4_percentage + \")\" + \\\n",
    "                \" , State5: \" + str(STATE_COUNT['state5']) + \"(\" + state5_percentage + \")\" + \\\n",
    "                \", State6: \" + str(STATE_COUNT['state6']) + \"(\" + state6_percentage + \")\" + \\\n",
    "                \" , State7: \" + str(STATE_COUNT['state7']) + \"(\" + state7_percentage + \")\" + \\\n",
    "                \" , State8: \" + str(STATE_COUNT['state8']) + \"(\" + state8_percentage + \")\" + \\\n",
    "                \", State9: \" + str(STATE_COUNT['state9']) + \"(\" + state9_percentage + \")\" + \\\n",
    "                \" , State10: \" + str(STATE_COUNT['state10']) + \"(\" + state10_percentage + \")\" + \\\n",
    "                \", State11: \" + str(STATE_COUNT['state11']) + \"(\" + state11_percentage + \")\" + \"\\n\")\n",
    "        f.write(\"TDM: \" + str(list1[len(list1)-1]) + \" , GREEDY: \" + str(list2[len(list2)-1]) + \\\n",
    "                \" , GENERIC: \" + str(list3[len(list3)-1]) + \"\\n\")\n",
    "        iterationStr = \"Iteration \" + str(p+1) + \" completed\\n\\n\"\n",
    "        f.write(iterationStr)\n",
    "        \n",
    "        ##Mprint \"\\nTDM_OPT= %s\\n\\nGREEDY_OPT= %s\\n\\nGENERIC_OPT= %s\" %(TDM_OPT, GREEDY_OPT, GENERIC_OPT)\n",
    "\n",
    "    ## Clearing the contents of the lists for next iteration\n",
    "    BASE = []\n",
    "    ## Clearing the STATE_COUNT Variable\n",
    "    STATE_COUNT['state0'] = 0\n",
    "    STATE_COUNT['state1'] = 0\n",
    "    STATE_COUNT['state2'] = 0\n",
    "    STATE_COUNT['state3'] = 0\n",
    "    STATE_COUNT['state4'] = 0\n",
    "    STATE_COUNT['state5'] = 0\n",
    "    STATE_COUNT['state6'] = 0\n",
    "    STATE_COUNT['state7'] = 0\n",
    "    STATE_COUNT['state8'] = 0\n",
    "    STATE_COUNT['state9'] = 0\n",
    "    STATE_COUNT['state10'] = 0\n",
    "    STATE_COUNT['state11'] = 0\n",
    "    ## this doe not work when we have to clear contents of a list within a list\n",
    "    #pdmList = []\n",
    "    #greedyList = []\n",
    "    #genericList = []\n",
    "    \n",
    "\n",
    "##Mprint\"optListReturn= %d, %s\" %(len(optListReturn), optListReturn)\n",
    "\n",
    "## Averaging the values of all the iterations\n",
    "TDM_OPT = [x/NUMBER_OF_ITERATION for x in TDM_OPT]\n",
    "GREEDY_OPT = [x/NUMBER_OF_ITERATION for x in GREEDY_OPT]\n",
    "GENERIC_OPT = [x/NUMBER_OF_ITERATION for x in GENERIC_OPT]\n",
    "\n",
    "print \"\\nTOTAL_CYCLES= %d\\nNUMBER_OF_ITERATION= %d\\nTDM_OPT= %s\\nGREEDY_OPT= %s\\nGENERIC_OPT= %s\\n\" \\\n",
    "%(N,NUMBER_OF_ITERATION, TDM_OPT, GREEDY_OPT, GENERIC_OPT)\n",
    "\n",
    "## closing the file handle\n",
    "f.close()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABnIAAANnCAYAAAAFm8aLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7b9d830>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## taken from http://matplotlib.org/examples/api/barchart_demo.html\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "numBars = 16\n",
    "## making the multiplier bigger reduces the size of bar's\n",
    "index = 1.5*np.arange(numBars) # location of groups\n",
    "width = 0.35 #width of each bar\n",
    "\n",
    "#fig, ax = plt.subplots()\n",
    "#rect1 = ax.bar(index, TDM_OPT, width, color='b')\n",
    "#rect2 = ax.bar(index+width, GREEDY_OPT, width, color='g')\n",
    "#rect3 = ax.bar(index+2*width, GENERIC_OPT, width, color='r')\n",
    "## Adding labels and titles\n",
    "#ax.set_ylabel('Speed Up')\n",
    "#ax.set_title('Speed up for various Methods')\n",
    "#ax.set_xticks(index+1.5*width)\n",
    "#ax.set_xticklabels(('0000', '0001', '0010', '0011',\\\n",
    "#                    '0100', '0101', '0110', '0111',\\\n",
    "#                    '1000', '1001', '1010', '1011',\\\n",
    "#                    '1100', '1101', '1110', '1111'))\n",
    "#ax.legend((rect1[0], rect2[0], rect3[0]), ('TDM', 'GREEDY', 'GENERIC'))\n",
    "\n",
    "fig = plt.figure(figsize=(20,10))\n",
    "rect1 = plt.bar(index, TDM_OPT, width, color='b')\n",
    "rect2 = plt.bar(index+width, GREEDY_OPT, width, color='g')\n",
    "rect3 = plt.bar(index+2*width, GENERIC_OPT, width, color='orange')\n",
    "\n",
    "tickLabel = (('0000', '0001', '0010', '0100', '1000',\\\n",
    "            '0011', '0101', '0110', '1001', '1010', '1100',\\\n",
    "            '0111', '1011', '1101', '1110',\\\n",
    "            '1111'))\n",
    "\n",
    "## Adding labels and titles\n",
    "plt.ylabel('Speed-up', fontsize=20)\n",
    "plt.xlabel('Optimized Components  [iCache, dCache, RUU, IQ]', fontsize=20)\n",
    "\n",
    "##title = 'Speed up of MCF,   Total iterations=' + str(N) + '   simulated for=' + str(NUMBER_OF_ITERATION) + ' times'\n",
    "##plt.title(title)\n",
    "plt.xticks(index+1.5*width, tickLabel, fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "\n",
    "plt.legend((rect1[0], rect2[0], rect3[0]), ('TDM', 'GREEDY', 'GENETIC'), fontsize=16, loc='lower right')\n",
    "plt.grid()\n",
    "\n",
    "plt.show()\n",
    "fig.savefig(imgFile)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}