RLNC_performance_study.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 128, 64, 78\n",
      " 128000, 64000, 78769\n",
      " 1280000, 640000, 787692\n",
      " 2560000, 1280000, 1575384\n",
      " 4096000, 2048000, 2520615\n",
      " 8192000, 4096000, 5041230\n",
      " 16384000, 8192000, 10082461\n",
      " 32768000, 16384000, 20164923\n",
      " 65536000, 32768000, 40329846\n",
      "\n"
     ]
    }
   ],
   "source": [
    "one_kb = 1024\n",
    "one_mb = one_kb * 1000\n",
    "ten_mb = 10 * one_mb\n",
    "twenty_mb = 2 * ten_mb\n",
    "thirdtwo_mb = 32 * one_mb\n",
    "six = 64 * one_mb\n",
    "oneh = 128 * one_mb\n",
    "twof = 256 * one_mb\n",
    "fivet = 512 * one_mb\n",
    "\n",
    "sizes = [one_kb, one_mb, ten_mb, twenty_mb, thirdtwo_mb, six, oneh, twof, fivet]\n",
    "generation_sizes = [8, 16, 13]\n",
    "\n",
    "def calc_symbol_size(ds, gs):\n",
    "    res = list()\n",
    "    \n",
    "    for d in ds:\n",
    "        t = list()\n",
    "        for g in gs:\n",
    "            t.append(d/g)\n",
    "        res.append(t)\n",
    "    return res\n",
    "\n",
    "ss = calc_symbol_size(sizes, generation_sizes)\n",
    "\n",
    "m = ''\n",
    "for s in ss:\n",
    "    \n",
    "    for x in s:\n",
    "        m = m + ' {},'.format(x)\n",
    "    m = m[:-1] + '\\n'\n",
    "print m\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Full Vector\n",
      "Diff: 2.06666666667 for g: 8 and gg: 16\n",
      "Diff: 2.03225806452 for g: 16 and gg: 32\n",
      "On-The-Fly\n",
      "Diff: 2.14285714286 for g: 8 and gg: 16\n",
      "Diff: 2.06666666667 for g: 16 and gg: 32\n",
      "llll\n",
      "g: 8 diff: 2.14285714286\n",
      "g: 16 diff: 2.06666666667\n",
      "g: 32 diff: 2.03225806452\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "one_kb = 1024\n",
    "one_mb = one_kb * 1000\n",
    "ten_mb = 10 * one_mb\n",
    "twenty_mb = 2 * ten_mb\n",
    "thirdtwo_mb = 32 * one_mb\n",
    "six = 64 * one_mb\n",
    "oneh = 128 * one_mb\n",
    "twof = 256 * one_mb\n",
    "fivet = 512 * one_mb\n",
    "\n",
    "#one_gb = 1024 * one_mb\n",
    "#two_gb = 2 * one_gb\n",
    "\n",
    "sizes = [one_kb, one_mb, ten_mb, twenty_mb, thirdtwo_mb, six, oneh, twof, fivet]\n",
    "\n",
    "\n",
    "def symbol_size(g, data_sizes):\n",
    "    symbol_sizes = list()\n",
    "    for ds in data_sizes:\n",
    "        symbol_sizes.append(ds/g)\n",
    "    return symbol_sizes\n",
    "\n",
    "def on_the_fly(g,k,r):\n",
    "    operations = 0\n",
    "    for i in range(0, g):\n",
    "        operations = operations + (i*k)\n",
    "    operations = operations + (r*(g*k))\n",
    "\n",
    "    for i in range(0, g):\n",
    "        operations = operations + (i*k)\n",
    "    operations = operations + (r*(k*(g-1)))\n",
    "    return operations\n",
    "    \n",
    "def diff_on_the_fly(g,k,r,gg,kk,rr):\n",
    "    return on_the_fly(gg,kk,rr)/float((on_the_fly(g,k,r)))\n",
    "\n",
    "def full_vector(g,k,r):\n",
    "    return(2*g-1)*k*(g+r)\n",
    "\n",
    "def diff_full_vector(g,k,r,gg,kk,rr):\n",
    "    return full_vector(gg,kk,rr)/float(full_vector(g,k,r))\n",
    "\n",
    "def diff_two_codes(g,k,r):\n",
    "    return full_vector(g,k,r)/float(on_the_fly(g,k,r))\n",
    "\n",
    "def calculate_avg_on_the_fly():\n",
    "    print('On-The-Fly')\n",
    "    # Gen 8,16:\n",
    "    diff = list()\n",
    "    g = 8\n",
    "    r = 0\n",
    "    s = symbol_size(g, sizes)\n",
    "\n",
    "    gg = 16\n",
    "    rr = 0\n",
    "    ss = symbol_size(gg, sizes)\n",
    "    \n",
    "    for k, kk in zip(s,ss):\n",
    "        diff.append(diff_on_the_fly(g,k,r,gg,kk,rr))\n",
    "        \n",
    "    dd = np.average(diff)\n",
    "    print('Diff: {} for g: {} and gg: {}'.format(dd,g,gg))\n",
    "    #for d in diff:\n",
    "    #    print(d)\n",
    "    \n",
    "    difff = list()\n",
    "    ggg = 32 \n",
    "    rrr = 0\n",
    "    sss = symbol_size(ggg, sizes)\n",
    "    for kk, kkk in zip(ss,sss):\n",
    "        difff.append(diff_on_the_fly(gg,kk,rr,ggg,kkk,rrr))\n",
    "    \n",
    "    dd = np.average(difff)\n",
    "    print('Diff: {} for g: {} and gg: {}'.format(dd,gg,ggg))    \n",
    "    #for d in difff:\n",
    "    #    print(d)  \n",
    "        \n",
    "def calculate_avg_full_vector():\n",
    "    print('Full Vector')\n",
    "    # Gen 8,16:\n",
    "    diff = list()\n",
    "    g = 8\n",
    "    r = 0\n",
    "    s = symbol_size(g, sizes)\n",
    "\n",
    "    gg = 16\n",
    "    rr = 0\n",
    "    ss = symbol_size(gg, sizes)\n",
    "    \n",
    "    for k, kk in zip(s,ss):\n",
    "        diff.append(diff_full_vector(g,k,r,gg,kk,rr))\n",
    "        \n",
    "    dd = np.average(diff)\n",
    "    print('Diff: {} for g: {} and gg: {}'.format(dd,g,gg))\n",
    "    #for d in diff:\n",
    "    ##    print(d)\n",
    "    \n",
    "    difff = list()\n",
    "    ggg = 32 \n",
    "    rrr = 0\n",
    "    sss = symbol_size(ggg, sizes)\n",
    "    for kk, kkk in zip(ss,sss):\n",
    "        difff.append(diff_full_vector(gg,kk,rr,ggg,kkk,rrr))\n",
    "    \n",
    "    dd = np.average(difff)\n",
    "    print('Diff: {} for g: {} and gg: {}'.format(dd,gg,ggg))    \n",
    "    # for d in difff:\n",
    "    #    print(d)     \n",
    "    \n",
    "def calc():\n",
    "    print('llll')\n",
    "    g = 8\n",
    "    r = 0\n",
    "    s = symbol_size(g, sizes)\n",
    "    diff = list()\n",
    "    for k in s:\n",
    "        diff.append(diff_two_codes(g,k,r))\n",
    "    d = np.average(diff)\n",
    "    print('g: {} diff: {}'.format(g,d))\n",
    "    \n",
    "    g = 16\n",
    "    r = 0\n",
    "    s = symbol_size(g, sizes)\n",
    "    diff = list()\n",
    "    for k in s:\n",
    "        diff.append(diff_two_codes(g,k,r))\n",
    "    d = np.average(diff)\n",
    "    print('g: {} diff: {}'.format(g,d))    \n",
    "\n",
    "    g = 32\n",
    "    r = 0\n",
    "    s = symbol_size(g, sizes)\n",
    "    diff = list()\n",
    "    for k in s:\n",
    "        diff.append(diff_two_codes(g,k,r))\n",
    "    d = np.average(diff)\n",
    "    print('g: {} diff: {}'.format(g,d))\n",
    "    \n",
    "calculate_avg_full_vector()\n",
    "calculate_avg_on_the_fly()\n",
    "calc()\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "ename": "IndentationError",
     "evalue": "unexpected indent (<ipython-input-14-bd53d30ef41c>, line 2)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;36m  File \u001b[0;32m\"<ipython-input-14-bd53d30ef41c>\"\u001b[0;36m, line \u001b[0;32m2\u001b[0m\n\u001b[0;31m    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\u001b[0m\n\u001b[0m    ^\u001b[0m\n\u001b[0;31mIndentationError\u001b[0m\u001b[0;31m:\u001b[0m unexpected indent\n"
     ]
    }
   ],
   "source": [
    ", value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "f = open('results/1510670606_full_vector_encoder_1_0_8_134217728')\n",
    "\n",
    "data = []\n",
    "\n",
    "for line in f:\n",
    "    temp = line.split(',')\n",
    "    index = 0\n",
    "    for item in temp: \n",
    "        temp[index] = parseData(index, item)\n",
    "        index = index + 1\n",
    "    data.append(temp)\n",
    "    \n",
    "        \n",
    "    \n",
    "    \n",
    "throughput = []\n",
    "latency = []\n",
    "\n",
    "for entry in data: \n",
    "    throughput.append(entry[5])\n",
    "    latency.append(entry[6])\n",
    "latency = [lat/(1000) for lat in latency]\n",
    "throughput = [th/(1000) for th in throughput] \n",
    "print(throughput[0])\n",
    "\n",
    "avg_latency = 0.0\n",
    "for val in latency:\n",
    "    avg_latency = avg_latency + val\n",
    "    \n",
    "avg_latency = avg_latency / len(latency)\n",
    "print avg_latency\n",
    "\n",
    "#df = pd.DataFrame(data, columns=['a'])\n",
    "#df.plot(kind='scatter', x='a')\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "ts = pd.Series(throughput)\n",
    "pd.to_numeric(ts)\n",
    "throughput_line = ts.plot(label='test', marker='o', linestyle='--')\n",
    "\n",
    "ts_2 = pd.Series(latency)\n",
    "pd.to_numeric(ts_2)\n",
    "latency_line  = ts_2.plot(marker='^', linestyle='--')\n",
    "plt.legend(['bytes/ms', 'Latency (Millisecons)'])\n",
    "plt.ylabel('bytes/ms | microseconds')\n",
    "plt.xlabel('# Data point')\n",
    "fig.savefig('results/graphs/full_vector_0_64_16.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "185.091932458\n",
      "13.7908244977\n",
      "2.11918413548\n",
      "t\n",
      "2.41145573908\n",
      "2.30046693432\n",
      "1.97750089697\n",
      "1.99841566616\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n",
      "534.278127\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAENCAYAAAAYIIIKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQ0REEvL/AVgPI0D5g4UOAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x111ca36d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8 \n",
    "## Full Vector - latency \n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "one_kb = 'results/encoder/thousand/1512452623_full_vector_encoder_binary8_0_8_128'\n",
    "one_mb = 'results/encoder/thousand/1512452934_full_vector_encoder_binary8_0_8_131072'\n",
    "ten_mb = 'results/encoder/thousand/1512453317_full_vector_encoder_binary8_0_8_1310720'\n",
    "twenty_mb = 'results/encoder/thousand/1512453778_full_vector_encoder_binary8_0_8_2621440'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512454344_full_vector_encoder_binary8_0_8_4194304'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512455182_full_vector_encoder_binary8_0_8_8388608'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512456573_full_vector_encoder_binary8_0_8_16777216'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512459046_full_vector_encoder_binary8_0_8_33554432'\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512463692_full_vector_encoder_binary8_0_8_67108864'\n",
    "\n",
    "\n",
    "def calc_delta(data):\n",
    "    for i in range(0, len(data)-1):\n",
    "        print('d: {}'.format(((data[i+1]-data[i])/float(data[i]))))\n",
    "\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print latencies[2]/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "print(latency[8])\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/full_vector_gen_8.eps')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13.1061896437\n",
      "101.322140355\n",
      "13.1048011196\n",
      "2.10868445901\n",
      "t\n",
      "2.11739018959\n",
      "2.51501772173\n",
      "2.35151874031\n",
      "2.0107896958\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x111a4b9d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 16\n",
    "## Full Vector - latency \n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def get_latency_t(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return latency\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512475943_full_vector_encoder_binary8_0_16_33554432'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512470804_full_vector_encoder_binary8_0_16_16777216'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512468085_full_vector_encoder_binary8_0_16_8388608'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512466613_full_vector_encoder_binary8_0_16_4194304'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512465749_full_vector_encoder_binary8_0_16_2097152'\n",
    "twenty_mb = 'results/encoder/thousand/1512465168_full_vector_encoder_binary8_0_16_1310720'\n",
    "ten_mb = 'results/encoder/thousand/1512464696_full_vector_encoder_binary8_0_16_655360'\n",
    "one_mb = 'results/encoder/thousand/1512464308_full_vector_encoder_binary8_0_16_65536'\n",
    "one_kb = 'results/encoder/thousand/1512463997_full_vector_encoder_binary8_0_16_64'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "mm = get_latency_t(one_mb)\n",
    "nn = get_latency_t(ten_mb)\n",
    "\n",
    "def calulate_avg_diffff(a1, a2):\n",
    "    ll = list()\n",
    "    for x,y in zip(a1, a2):\n",
    "        ll.append((y-x)/float(x))\n",
    "    return np.average(ll)\n",
    "\n",
    "#print(calulate_avg_diffff(mm, nn))\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print (latencies[2]-float(latencies[1]))/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "    \n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1KiB','1MB', '10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/full_vector_gen_16.eps')\n",
    "plt.show()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "82.5467896718\n",
      "14.1233696123\n",
      "2.10868713404\n",
      "t\n",
      "2.05370177557\n",
      "2.0735710227\n",
      "2.5096788308\n",
      "2.43921241017\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xl8VOXZ//HPRcJO2JewhV0QggJGEFwqKgqura1WxOWp\nC3bRVmufp/Znf7Xt7+mr9rH2UVu1WvcN69aqJbgLLhXZJUFkESEJJOwJgbBkuX5/nBNN0wRmIJOZ\nTL7v1yuvZM6cOefKZDLXXPd9n/s2d0dERCRSLeIdgIiINC1KHCIiEhUlDhERiYoSh4iIREWJQ0RE\noqLEISIiUVHiEBGRqChxiIhIVJQ4REQkKqnxDiAWunfv7gMHDox3GCIiTcbixYu3uXuPSPZNysQx\ncOBAFi1aFO8wRESaDDPbEOm+aqoSEZGoJFXiMLPzzOzBkpKSeIciIpK0kipxuPur7j6zU6dO8Q5F\nRCRpJVXiEBGR2FPiEBGRqChxiIhIVJQ4REQkKkocIiJNnLvzz7XbuOP1zxrlfEl1AaCZnQecN3To\n0HiHIiISc9t37+eFxQU8uzCfL7btoVPbllw5aSA909rE9Lzm7jE9QTxkZWW5rhwXkWTk7ny0bjuz\nFuTzem4RByqrOH5gFy6dkMG0zN60aZlyWMc1s8XunhXJvklVcYiIJKsdew7w4uICZi3IY922PXRs\nk8qMEzK4dHwGw3qlNWosShwiIgnK3Zm/bgezFuTxWlhdZA3owvWnDeXs0YdfXRwpJQ4RkQSzY88B\nXlpSwDML8li3NaguLp2QwaUTMjiqkauLuihxiIgkAHdnwRc7eGZBHnNyguriuAFduPOioLpo2yo+\n1UVdlDhEROJo554DvLgk6Lv4fOse0sLqYvr4DIanx7+6qIsSh4hII3N3Fq7fyTMfbyA7t4gDFVWM\ny+jM7y86lnMSrLqoS8InDjM7GvgR0B14293vj3NIIiKHpbjsAC8u2cisBXms3bKbtDapTD++P9Mn\nZDAivWO8w4tYXBKHmT0CnAtscffMGtunAncDKcBD7n67u68EvmtmLYAnACUOEWky3J1FG3byzMd5\nzM4p5EBFFWMzOnPHt47h3GP6JHx1UZd4VRyPAX8iSAQAmFkKcC8wBSgAFprZK+7+qZmdD3wPeDIO\nsYqIRK2krPzLvos1W3aT1jqVS47vz/TxGRzdu+lUF3WJS+Jw9/fMbGCtzeOBte6+DsDMngUuAD51\n91eAV8xsNvBMY8YqIhIpd2dxjepif0UVY/p35n++eQznHtubdq0SvncgIon0W/QF8mvcLgAmmNmp\nwIVAayC7vgeb2UxgJkBGRkbsohQRqaWkrJyXlgbVxerNQXVxcVZQXYzs07Sri7okUuKok7vPBeZG\nsN+DwIMQzFUV26hEpLlzd5bk7eTpj/OYvTyoLo7t35nffXM05x3bJ2mqi7ok0m+2Eehf43a/cFvE\nNDuuiMRayd5y/rakgFkL8lm1uZQOrVO5KKsf08dnMKpPp3iH1ygSKXEsBIaZ2SCChHEJcGk0B3D3\nV4FXs7Kyro1BfCLSjOVuLOGp+Rt4edkm9pZXcmy/Ttx+YVBdtG+dSG+lsRev4bizgFOB7mZWANzm\n7g+b2fXA6wTDcR9x9xVRHlcVh4g0mH3llWTnFPLk/A0szSumTcsWfH1MXy47YQCZfZtHdVEXrcch\nIlJL3vYynl6wgecW5rOzrJzBPdpz+QkDuHBcPzq1bRnv8GKi2a7HoYpDRA5XZZUzd9UWnpy/gXmr\nt9LCjDNH9uLyEwYwcUg3zCzeISYMVRwi0qxt272f5xbl8/T8PDYW76VnWmumjw8mGUzvFNslWBNJ\ns604REQiUT2U9smPNpAdTmE+cXA3bj3naKaM7EXLlBbxDjGhJVXiUFOViBzMnv0VvLxsE0/O38DK\nwl2ktQ6mML/shAyG9kzMKcwTkZqqRCTprdlcylPzN/DSko2U7q/g6N4duWLiAM5vhkNp66OmKhFp\n9sorq3hjxWaenL+e+et20CqlBecc05vLThjAuIzO6uw+AkmVONRUJSKFJXuZtSCfZxfksaV0P/26\ntOWnU0dwcVY/unVoHe/wkoKaqkSkyXN3Ply7nafmb+DNlZupcufUo3pw+cQBfO2onqS0UHVxKGqq\nEpFmoaSsnBeWFPD0/A2s27aHLu1acu3Jg5kxIYP+XdvFO7ykpcQhIk1O7sYSnvxoAy9/spF95cF6\n3f/77WOZltmbNi2b3op6TU1SJQ71cYgkr33llcxeHswbtSy/mLYtU/jG2L7MmNC8542KB/VxiEhC\n27B9D09/nMdzi/IpLitnSI/2XJbk80bFg/o4RKRJ27O/gndXbeH5RQXMW72VlBbGWaN6cdkJA5g4\nWPNGxZsSh4gkhNJ95bzz2RaycwqZu2or+yuq6NWxNTeeMYxLjm9e80YluqRKHOrjEGlaSsrKeXPl\nZubkFPL+mm0cqAySxfTxGUzLTCdrYFcNpU1A6uMQkUa1Y88B3vy0iOycIj5cu42KKqdPpzZMG92b\ns0enM7Z/F1ooWTQ69XGISELZWrqfNz4tYk5OER+t205lldO/a1uuPmkQ00b35th+ndRv0YQocYhI\nTGzetY/XVxSRnVPIgi92UOUwqHt7rjtlMGeP7s2oPh2VLJooJQ4RaTCbivfyWm4Rc3ILWbRhJ+4w\nrGcHrj9tGGePTmd4rzQliySgxCEiRyR/RxlzcgvJziliWX4xACPS07jpjKOYlpnOsF5a5yLZKHGI\nSNTWb9tDdm4hc3KKyNlYAkBm347851nDmZaZzuAeHeIcocRSUiUODccViZ21W3YzJ6eQ7NwiVhbu\nAmBM/878bNoIpmX2JqObJhVsLjQcV0Tq5O6s3ryb7JxC5uQWsnrzbgCyBnRh2ujeTM1Mp2/ntnGO\nUhqKhuOKyGFxdz4t3MWcnCKycwtZt3UPZjB+YFd+df4ozhqVriu4RYlDpLlzd3I2lpCdE4yG2rC9\njBYGE4d046oTB3HmqF70TFOykK8ocYg0Q1VVzuK8nczJKeL1FUVsLN5Lagtj0tDufO9rQ5gyspeW\nWZV6KXGINBMVlVV8/MUO5uQW8vqKzWwt3U+rlBacPKw7N54xjCkje9G5Xat4hylNgBKHSBLbX1HJ\nh2u3MSeniDdXbqa4rJy2LVOYPKIHZ41K57QRPUlrozUtJDpKHCJJZu+BSuat3sKc3CLeWbmF0v0V\npLVO5fSjezI1szdfO6oHbVtpeVU5fAmfOMzs68A5QEfgYXd/I84hiSSc6rUs5uQUMXf1FvaVV9Gl\nXUvOHt2bqaPTmTSkG61TlSykYcQlcZjZI8C5wBZ3z6yxfSpwN5ACPOTut7v734G/m1kX4PeAEocI\nsHPPAd78dDOvrSjig3Ati55prbnouP5My0xn/KCupKa0iHeYkoTiVXE8BvwJeKJ6g5mlAPcCU4AC\nYKGZveLun4a7/Dy8X6TZ2lK6j9dXbOa13ELmr9tBZZXTt3Nbrpg4gGlay0IaSVwSh7u/Z2YDa20e\nD6x193UAZvYscIGZrQRuB+a4+5JGDVQkARTsLOO13GDYbPWMs4PD6cmnZfYms6+mJ5fGlUh9HH2B\n/Bq3C4AJwA3AGUAnMxvq7n+u68FmNhOYCZCRkRHjUEVi64tte5iTW8hruUUsLwgmERyRnsaNpx/F\ntNHpDOvZQclC4iaREked3P0e4J4I9nsQeBCCuapiHZdIQ3J3Vm0uZU5OEa/lFrFqcykAx/bvzE+n\njmBaZjoDu7ePc5QigURKHBuB/jVu9wu3RUyz40pTUj3Vx5zcIFl8sS2YF+r4AV35xbkjmZqZTh9N\nIigJKJESx0JgmJkNIkgYlwCXRnMAd38VeDUrK+vaGMQncsTqmuojpYUxaUg3rjl5EFNGal4oSXzx\nGo47CzgV6G5mBcBt7v6wmV0PvE4wHPcRd18R5XFVcUhCqaisYtXmUpbmFbMkbyfvr9mmqT6kydN6\nHCINaGvpfpbm7WRpfjFL83ayvKCEsgOVAHTv0IoJg7pxVmY6k4f30FQfklCa7XocqjikMR2oqGJl\n4S6W5u1kSV4xS/N3kr9jLwCpLYxRfTpycVZ/xmZ0ZlxGF/p1aauRUJIUVHGIRKioZB9L8nYGFUVe\nMTkbS9hfUQVAr46tGZfRhXEZXRib0ZnMvp1o01JTfEjT0WwrDpGGsq+8khWbSr7sm1iaV0xhyT4A\nWqW2YHTfTlx+wgDGZnRh3IDO9O6k0U/SfCRV4lBTlRwOd6dg594vE8TS/GI+3VRCeWVQjffr0pas\ngV0Zl9GZsRldOLp3miYMlGZNTVXS7JQdqGB5wb9WE9t27wegbcsUjunXibFhk9PYjM4aHivNgpqq\nRELuzvrtZSzZsJOl+UGS+KyolMqq4APToO7tOWVYd8YO6MLY/p0ZkZ6mGWVFDiGpEoeaqqSqylma\nX8w/124Lqon8YorLygHo0DqVMf078/1ThzA2ozNj+neha3tdPyESLTVVSZPnHiSL7OWFZOcUsins\nxB7Ws8OXQ2HHZnRhaM8OpGjKcZE6xaypKlxMqb+7Lz+syEQaiLvzSUEJs5dvIjsnmLqjVUoLTjmq\nB/85dTinDe9Fp3a6wE4kFg6ZOMxsLnB+uO9iYIuZfejuP45xbCL/wt1ZXlBCdk4h/1heyMbivbRM\nMU4Z1oObzzyKM0b2oqOuxhaJuUgqjk7uvsvMrgGecPfbzCwhKw71cSQfdyd34y7+kbOJ2csLKdgZ\nJIuTh/XgpilHMWVkLzq1VbIQaUyRJI5UM+sNXAzcGuN4johmx00O7s6KTbuYnVPI7OWF5O0oI7WF\ncdKw7vzo9GGcOTJdzVAicRRJ4vg1wYy1H7j7QjMbDKyJbVjS3Lg7nxbuYvbyQmbnFLJhe5AsJg3t\nzvWTh3LmKM0gK5IoDpk43P154Pkat9cB34xlUNI8uDsrC0vJzgmSxRfb9ny5NsX3Tx3CmSPT6aLh\nsiIJJ5LO8UEE634PrLm/u58fu7AkWVUvkTp7edAMtW7bHloYTBrSnZmnDOasUem6tkIkwUXSVPV3\n4GHgVaAqtuEcGXWOJ67Vm0v5x/JCZi/fxOdbg2QxcUg3rjl5MGeN6kW3Dq3jHaKIROiQFwCa2cfu\nPqGR4mkQugAwMaypThY5hazdspsWBhMGdeOcY3ozNTOd7koWIgmjoS8AvNvMbgPeAPZXb3T3JYcZ\nnySxtVtKmb28iNk5m1i9eTdmMGFQV66cOIqzMtM1YaBIEogkcYwGLgdO46umKg9vi/D51t1f9lms\n2lyKGRw/sCu/vmAUU0el07OjkoVIMokkcVwEDHb3A7EORpqWl5dt5P65n/NZUZgsBnTll+eNZNro\n3vRSshBJWpEkjlygM7AlxrFIE/LOZ5u58a/LODq9I7edN5Jpmb1J76RkIdIcRJI4OgOfmdlC/rWP\nQ8Nxm6nPinZxwzNLGdWnI89dN5F2rZJqdn4ROYRI/uNvi3kUDUTDcWNv2+79XP3YIjq0SeWhK45X\n0hBphiK5cnxeYwTSEDRXVWztK69k5hOL2L5nP89fN0lNUyLNlD4uSkTcnVteXM6SvGLumzGO0f06\nxTskEYkTLa4sEbn33bX8fdkmfnLmUZw9une8wxGROFLikEPKzink92+s5htj+/KDyeo/EmnuIpnk\n8ETgl8CAcH8D3N0HxzY0SQTLC4r58XPLOG5AF3574WjMtGa3SHMXSR/Hw8BNBMvGVsY2HEkkRSX7\nuPaJRXRr35oHLj+ONi1T4h2SiCSASBJHibvPiXkkklDKDlRwzRML2b2vghe/P0kTEorIlyJJHO+a\n2R3AS8RhksNwxcFbCdY+/1ZjnLO5q6pyfvzXT/h00y4eujKLEekd4x2SiCSQSBJH9ZTqNafbPaJJ\nDs3sEeBcYIu7Z9bYPhW4G0gBHnL328MVB682sxcO93wSnTvfXMVrK4r4+TlHc9qIXvEOR0QSTCQX\nAE6OwXkfA/4EPFG9wcxSgHuBKUABsNDMXnH3T2NwfqnHS0sKuPfdz5k+PoOrTxoU73BEJAHVmzjM\n7DJ3f8rMflzX/e7+h8M9qbu/Z2YDa20eD6wNKwzM7FngAkCJo5EsXL+DW17MYdKQbvz6glEaQSUi\ndTrYdRztw+9p9Xw1tL5Afo3bBUBfM+tmZn8GxprZz+p7sJnNNLNFZrZo69atMQgvueXvKOO6JxfT\nt0tb7psxjpYpusRHROpWb8Xh7g+E33/VeOHUGcd24LsR7PegmRUC57Vq1eq42EeWPEr3lXP14wup\nrHIevjKLzu1axTskEUlgifSxciPQv8btfuG2iLn7q+4+s1MnzaMUqYrKKm6YtZR1W/dw/4xxDO7R\nId4hiUiCS6TEsRAYZmaDzKwVcAnwSjQHMLPzzOzBkpKSmASYjH6TvZK5q7by6wsymTS0e7zDEZEm\n4JCJIxzt1KDMbBbwETDczArM7Gp3rwCuB14HVgLPufuKaI6riiM6T3+8gUc/XM9VJw7i0gkZ8Q5H\nRJqISK7jWGNmLwKPNtTQWHefXs/2bCC7Ic4hB/fh2m384uUVTB7eg1vPOTre4YhIExJJU9WxwGrg\nITObH45eSshLidVUFZnPt+7me08tZmiPDtwzfSwpLTTsVkQid8jE4e6l7v4Xd58E/JRgKdlCM3vc\nzBJqjm01VR1acdkBrnl8ES1TWvDQlVmktWkZ75BEpImJqI/DzM43s78BdwF3AoOBV0mwZiVVHAdX\nXlnF955awsade3ng8uPo37VdvEMSkSYokqaqNQRXcN/h7mPd/Q/uvtndXwBei2140VHFUT935xcv\n5/LRuu387lujyRrYNd4hiUgTFUnn+DHuvruuO9z9hw0cj8TIwx98wawF+fxg8hC+MbZfvMMRkSYs\nkorjXjPrXH3DzLqEs9smHDVV1e2dzzbzm+yVTMtM5+Ypw+Mdjog0cZEkjmPcvbj6hrvvBMbGLqTD\np6aqf/dZ0S5ueGYpo/p05M6Lj6WFRlCJyBGKJHG0MLMu1TfMrCuRNXFJnG3bvZ+rH1tEhzapPHTF\n8bRrpT+biBy5SN5J7gQ+MrPnAQO+BfwmplHJEdtXXsnMJxaxfc9+nr9uEumd2sQ7JBFJEpEs5PSE\nmS0Gqhd0ujBRF1cys/OA84YOTajLSxqdu3PLi8tZklfMfTPGMbqfmu5EpOFEOsnhZwRrjr8C7Daz\nhJzYSH0cgT+9s5a/L9vET848irNH9453OCKSZA5ZcZjZDQRXi28GKgmaqxw4JrahyeGYvbyQO99c\nzTfG9uUHk5t35SUisRFJH8ePgOHhgkqSwJYXFHPz88s4bkAXbv/maC39KiIxEUlTVT6gCyMSXGHJ\nXq55fBHdO7TmgcuPo3Vqg8+GLyICRFZxrAPmmtlsYH/1Rnf/Q8yiOkzNtXO87EAF1zy+iLIDlTx5\n9QS6d2gd75BEJIlFUnHkAW8CrYC0Gl8Jpzl2jldVOT/+6yesLNzFH6ePZXh6Qv5pRCSJRDIc91cA\nZtbO3ctiH5JE4/dvrOK1FUX833NHMnlEz3iHIyLNQCTTqk80s08JhuRiZsea2X0xj0wO6cXFBdw3\n93Omj8/gqhMHxjscEWkmImmqugs4C9gO4O6fAKfEMig5tIXrd/Czl3KYNKQbv75glEZQiUijiegC\nQHfPr7WpMgaxSITyd5Rx3ZOL6dulLffNGEfLlEiv4xQROXKRjKrKN7NJgJtZS4LrOlbGNiypT+m+\ncq5+fCGVVc7DV2bRuV2reIckIs1MJB9Vvwv8AOgLbATGAN+PZVCHK9nX46iorOKGWUtZt3UP988Y\nx+AeHeIdkog0Q5EkjuHuPsPde7l7T3e/DDg61oEdjmQfjvub7JXMXbWVX1+QyaSh3eMdjog0U5Ek\njj9GuE1i6OmPN/Doh+u56sRBXDohIeeYFJFmot4+DjObCEwCepjZj2vc1RHQfBaN6MO12/jFyyuY\nPLwHt56TkMWeiDQjB+scbwV0CPepeTnyLoLFnKQR7Cuv5ObnPmFw9/bcM30sKVr6VUTirN7E4e7z\ngHlm9pi7b2jEmKSGZxfkUbRrH3+4eAJpbVrGOxwRkYiG45aZ2R3AKODL9Ufd/bSYRSVAUG3cN/dz\nxg/qysQh3eIdjogIEFnn+NME040MAn4FrAcWxjAmCT3zcR5bSvdz0xlH6cpwEUkYkSSObu7+MFDu\n7vPc/Sqg0aoNM2tvZo+b2V/MbEZjnTfe9h4Iqo0TBqvaEJHEEkniKA+/F5rZOWY2Fuh6JCc1s0fM\nbIuZ5dbaPtXMVpnZWjO7Jdx8IfCCu18LnH8k521Knv54A9t2B9WGiEgiiSRx/LeZdQJuBn4CPATc\neITnfQyYWnODmaUA9wLTgJHAdDMbCfQjWIUQmskcWWUHKvjzvM+ZNKQbEwar2hCRxHLIxOHu/3D3\nEnfPdffJ7n4cMORITuru7wE7am0eD6x193XufgB4FrgAKCBIHhHFmwyemr+BbbsPcNMUVRsikngO\n9434x4feJWp9+aqygCBh9AVeAr5pZvcDr9b3YDObaWaLzGzR1q1bYxBe4yg7UMED89Zx0tDuHD/w\niFoERURiIpLhuHVptCE+7r4H+E4E+z0IPAiQlZXlsY4rVp74aAPb9xzgpinD4h2KiEidDrfiiMUb\n80agf43b/cJtEWvqs+Pu2V/Bg++t4+Rh3TlugKoNEUlM9SYOMys1s111fJUCfWIQy0JgmJkNMrNW\nwCXAK9EcoKnPjvv4R+vZsUd9GyKS2OpNHO6e5u4d6/hKc/fDbeICwMxmAR8Bw82swMyudvcK4Hrg\ndYKFop5z9xVRHrfJVhy7w2rj1OE9GJfRJd7hiIjU64gSwOFy9+n1bM8Gso/guK8Cr2ZlZV17uMeI\nl8f/uZ7isnJu1HUbIpLgkmp4a1OtOEr3lfPge+s4bURPxvTvHO9wREQOKqkSR1Pt43jsw/WU7C3n\nxjM0kkpEEl9SJY6maNe+cv7y/jrOOLonx/RTtSEiiS+pEkdTbKp69IP17NpXob4NEWkykipxNLWm\nqpK95Tz0wTqmjOxFZt+mEbOISFIljqbmkQ++oHRfhfo2RKRJSarE0ZSaqkrKynnkgy84a1QvRvVR\ntSEiTUdSJY6m1FT18AfrKN2vvg0RaXqSKnE0FcVlB3jkw/VMy0zn6N4d4x2OiEhUlDji4KH3v2D3\n/gp+pL4NEWmCkipxNIU+jp17DvDoh19wzujejEhXtSEiTU9SJY6m0Mfxl/fXUVZeqWpDRJqspEoc\niW7HngM8/s/1nDO6N0f1Sot3OCIih0WJoxE9+F5YbZyuakNEmi4ljkayffd+nvhoPecd04dhqjZE\npAlLqsSRyJ3jD763jn3llfxQ1YaINHFJlTgStXN82+79PPHRBs4/tg9De3aIdzgiIkckqRJHonpg\n3ufsr1C1ISLJQYkjxraU7uPJ+Rv4+pi+DO6hakNEmj4ljhh7YN46yiudG1RtiEiSUOKIoS279vFU\nWG0M6t4+3uGIiDQIJY4Yun/e51RUOT88fWi8QxERaTBJlTgSaTju5l37ePrjPC4c25cB3VRtiEjy\nSKrEkUjDce+f+zlVVc4Np6lvQ0SSS1IljkRRVLKPZxbk8c1x/cjo1i7e4YiINCgljhi4b+5aqqqc\n609T34aIJB8ljga2qXgvzy7I56KsfvTvqmpDRJKPEkcDu2/uWhznB5NVbYhIclLiaEAbi/fy14X5\nXJTVn35dVG2ISHJS4mhA9767FsNUbYhIUkv4xGFmg83sYTN7Id6xHEzBzjKeX5TPt4/vT9/ObeMd\njohIzMQ0cZjZI2a2xcxya22famarzGytmd1ysGO4+zp3vzqWcTaE6mrj+5OHxDsUEZGYSo3x8R8D\n/gQ8Ub3BzFKAe4EpQAGw0MxeAVKA39Z6/FXuviXGMR6x/B1lPL+ogBkTMujdSdWGiCS3mCYOd3/P\nzAbW2jweWOvu6wDM7FngAnf/LXDu4Z7LzGYCMwEyMjIO9zCH5Y/vrKFFC+P76tsQkWYgHn0cfYH8\nGrcLwm11MrNuZvZnYKyZ/ay+/dz9QXfPcvesHj16NFy0h7Bh+x5eXLKRS8dn0Ktjm0Y7r4hIvMS6\nqeqIuft24LuR7Gtm5wHnDR3aeJ/8//jOWlJbGN8/VX0bItI8xKPi2Aj0r3G7X7jtiDX2JIfrt+3h\nb0s3MmPCAHqq2hCRZiIeiWMhMMzMBplZK+AS4JWGOHBjT6t+zztraJlifPfUwY1yPhGRRBDr4biz\ngI+A4WZWYGZXu3sFcD3wOrASeM7dVzTE+Rqz4li3dTd/X7qRyyYMoGeaqg0RaT5iPapqej3bs4Hs\nhj5fY/Zx/PGdtbRKbcF1X1Pfhog0Lwl/5Xg0Gqvi+Hzrbl5etpErJg6kR1rrmJ5LRCTRJFXiaCz3\nvL2G1qkpzDxFfRsi0vwkVeJojM7xtVtKeeWTTVwxaQDdO6jaEJHmJ6kSR2M0Vd399lratkzhulPU\ntyEizVNSJY5YW725lH8s38SVkwbStX2reIcjIhIXSZU4Yt1Udffba2jXMoWZJ6tvQ0Sar6RKHLFs\nqlpVVEp2TiH/ceJAuqjaEJFmLKkSRyzd/fZq2rdK5VpVGyLSzClxRGBl4S6yc4r4zokD6dxO1YaI\nNG9JlThi1cdx91trSGudyjUnqdoQEUmqxBGLPo4Vm0p4bUUR3zlpEJ3atWyw44qINFVJlThi4e63\n1pDWJpWrTxoU71BERBKCEsdB5G4s4Y1PN3P1SYPo1FbVhogIKHEc1F1vraFjm1SuUrUhIvKlpEoc\nDdk5nlNQwlsrN3PNyYPp2EbVhohItaRKHA3ZOX7XW6vp1LYl3zlx4JEHJiKSRJIqcTSUT/KLefuz\nLVx78iDSVG2IiPwLJY463PXWajq3a8mVkwbGOxQRkYSjxFHL0rydvLtqK9eePFjVhohIHZQ4arnr\nrTV0UbUhIlIvJY4aFm/YybzVW7nua0Po0Do13uGIiCSkpEocRzoc9663VtOtfSuumDiggSMTEUke\nSZU4jmQ4bkVlFSPS0/jh6cNo10rVhohIffQOGUpNacGt54yMdxgiIgkvqSoOERGJPSUOERGJihKH\niIhERYmrEQrtAAAKJklEQVRDRESiosQhIiJRUeIQEZGoKHGIiEhUlDhERCQq5u7xjqHBmdlWYMNh\nPrw7sK0Bw2koiis6iis6iis6yRjXAHfvEcmOSZk4joSZLXL3rHjHUZviio7iio7iik5zj0tNVSIi\nEhUlDhERiYoSx797MN4B1ENxRUdxRUdxRadZx6U+DhERiYoqDhERiUqzSxxm9oiZbTGz3Brb5prZ\nv41EMLP1ZpZjZsvC7xc0UjyPmVmZmaXV2HaXmbmZdQ9vV4ZxfWJmS8xsUoxi6Wpmb5rZmvB7l3D7\nqWE819TYd0y47Sc1fo8vwjg/M7PbDjOu/mb2rpl9amYrzOxHiRCbmbUxswXh32CFmf0q3H5HeMzl\nZvY3M+sch+ess5m9EB5jpZlNrHHfzbVeSzGLq57XVH3PT0szezz8X1tpZj+r8Zj1ZvZ+rWMvqz5u\n+DuUhNuWm9lbZtbzIHHV95r6pZltDI+zzMzOrvGYY8zso3D/HDNrE6PYar7vLAq3XRSet8pqvFeZ\n2RQzWxzuv9jMTovVc1ZTs0scwGPA1Cj2n+zuY4BvAfc0YjxrgQsAzKwFcBqwscb9e919jLsfC/wM\n+G2MYrkFeNvdhwFvh7er5QIX17g9Hfik1uP/M3z+xgBXmtmgw4irArjZ3UcCJwA/MLORCRDbfuC0\n8G8wBphqZicAbwKZ7n4MsJrg79OYcQHcDbzm7iOAY4GVELxhAmcCebX2j1Vcj/Hvr6n6np+LgNbu\nPho4DrjOzAbWeFxaGD9mdnQd53o//J84BlgI/OAgcdX3mgL43/A4Y9w9OzxfKvAU8F13HwWcCpTH\nKDYI33dqDK3NBS4E3qu13zbgvPA5uxJ4stb9DR0X0AwTh7u/B+yo6z4zaxF+svrvOu7uCOxsxHie\nBb4d/nwq8CHBi70uDRJbPbFcADwe/vw48PUa920A2phZLzMzgjeIOfUcvk34fc9hxFXo7kvCn0sJ\n3gT7xjs2D+wOb7YMv9zd33D36r/VfKBfY8ZlZp2AU4CHwzgPuHtxePf/Av8F1O7cjElcdb2mDvL8\nONA+fJNuCxwAdtV46HN89T8xHZhV1znD+NM4yP/EQV5T9TkTWO7un4SP2e7ulbGIrZ54V7r7qjq2\nL3X3TeHNFUBbM2sd67iaXeI4iFTgaWCNu/+8xvZ3w9JuHvDzOh8ZG6uBHhY0v0wnSCQ1ta1uNgAe\nAv5fjOLo5e6F4c9FQK9a979A8ElxErCE4FN4TXeY2TKgAHjW3bccSTDhJ9CxwMeJEJuZpYTH2AK8\n6e4f19rlKv79DTjWcQ0CtgKPmtlSM3vIzNpb0NS6sfrNrw6N+rcM1Xx+XiBIRoUEFdHv3b1m0nmR\n4FM3wHnAq7WOdXIYXx5wBvBIJAHUek0B3BA23TwS/v8BHAW4mb1uQdPwf9U6TEPG5sBbYdPTzEh+\nh9A3gSXuXvPvFpPnTInjKw8Aue7+m1rbJ7t7JjAa+JOZdWjEmF4CLgEmAO/Xuq+6qWoEwafDJ8JP\nDTHjwRC82p9UnyN4s6nv00x180Y6cLodQV9M+Ny/CNzo7jU/icYtNnevDI/RDxhvZpk14r2VoEp8\nupHjSgXGAfe7+1iCN+NfAv8H+MVBHtdof0uo8/kZD1QCfQiS381mNrjGQ7YDO83sEoIKoazWIaub\nXfoDjwL/E0EMtV9T9wODCZrjCoE7w11TgZOAGeH3b5jZ6TGK7aTweZ5G0IR2SgS/xyjgd8B1te5q\n8OcMlDhq+icwubrDqzZ3/xzYDIys6/4Y+StBJfGmu1fVt5O7f0QwR01E88xEabOZ9QYIv//Lp0x3\nLyJo651C0M9QX4y7gbkE/3RRM7OWBP/gT7v7S4kUW3iMYuBdwvZ8M/sP4Fxghtca894IcRUABTWq\nnxcIEskg4BMzW0+Q6JaYWXojxvWlep6fSwn6ZcrDauZDoPaglb8C91JPk0sNrxA01x0shn97Tbn7\n5vDDQBXwF4JkBsFz+p67b3P3MiCb4Dlt8NjcfWP4fQvwtxox1Pd79Av3uyJ8n6qtwZ6zakocX3mY\n4MXwXNjG+i/C0QaDOPzJE6Pm7huAW4H7DrafmY0AUgg+XTS0Vwg63Qi/v1zHPr8AflqrzfdfhM/p\nBKCuF/ZBhZXUw8BKd/9DosRmZj3sqxFBbQnecD8zs6kE/Qjnh28ydYlZXGECyDez4eGm0wmaMHq6\n+0B3H0jwRjgu3LdR4qrx+PqenzyCQSCYWXuCTuvPaj38bwSfil8/xGlOOlh89b2mqj+IhL5B0ClN\neL7RZtYu/P2/Bnza0LGFTYpp1T8T9K3k1rVvuE9nYDZwi7t/WM9uDfKc1fRvb5DJzsxmEXQ2dzez\nAuDLYYXu/oewY/FJM5sRbn7XzCoJOj5vcffNjRVPGNMD9Ty0bdg2CWDAlQf7Zz+CWG4nSKZXEyTN\ni2s/zt3/eZDD3mFmPwdaEXyKfekg+9bnROByIKfG7/x/EiC23sDjZpZC8CHsOXf/h5mtBVoDb4at\nh/Pd/buNGBfADcDTZtYKWAd8J5IHNXRc9bymfkbdz8+9BP0yKwhe04+6+/Ja8ZUSNMlQR8tsdXu9\nASXANbV3qKG+19R0MxtD0Oy5nrDpx913mtkfCEYeOZDt7rNjEFsv4G/h41OBZ9z9NTP7BvBHglaF\n2Wa2zN3PAq4HhgK/MLPqZsgza/Y/NeBz9iVdOS4iIlFRU5WIiERFiUNERKKixCEiIlFR4hARkago\ncYiISFSUOETqYF/NPrzCgtlvb7ZgssmDPWagmV3aQOd/yL6adE8koWg4rkgdzGy3u3cIf+4JPAN8\n6O71TiduZqcCP3H3cxsnSpH4UMUhcgjhxVQzgestMNDM3g8nu6u5FsrthBdUmdlNB9nvS+GVwrPD\nqibXzL4dbp9rZllmdr59tTbEKjP7Irz/ODObZ8FEeK/XuuJZJKaa3ZXjIofD3deFV4j3JJgTa4q7\n7zOzYQRzAGURrAfyZcVhZu3q2a+mqcAmdz8nfEynWud9hWBqFczsOWBeOMfSH4EL3H1rmGx+QzDT\nrEjMKXGIRK8lwUzJYwhmcz3qCPbLAe40s98B/3D32rMgA2DBNN573f1eC2bgzeSrKTtSCGZyFWkU\nShwiEbBgeu9KgmrjNoKZko8laO7dV8/DbjrUfu6+2szGAWcD/21mb7v7r2ud+wyC6c6rZy41YIW7\nT0QkDtTHIXIIZtYD+DPwp3AK8E5AYTj19uUEn/gBSglWUatW3341j90HKHP3p4A7qDVVt5kNIJj8\n7yJ33xtuXkWwyNfEcJ+WFqzHINIoVHGI1K169uGWBIsNPQlUT799H/CimV0BvMZXS6guByrN7BOC\ntbbr26+m0QSzzlYRrIXxvVr3/wfQDfh72Cy1yd3PNrNvAfeEfSKpwF0ES4eKxJyG44qISFTUVCUi\nIlFR4hARkagocYiISFSUOEREJCpKHCIiEhUlDhERiYoSh4iIREWJQ0REovL/AaHH2IozNrERAAAA\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10e88d790>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 32 \n",
    "## Full Vector - latency \n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "one_kb = 'results/encoder/thousand/1512476247_full_vector_encoder_binary8_0_32_32'\n",
    "one_mb = 'results/encoder/thousand/1512476560_full_vector_encoder_binary8_0_32_32768'\n",
    "ten_mb = 'results/encoder/thousand/1512476958_full_vector_encoder_binary8_0_32_327680'\n",
    "twenty_mb = 'results/encoder/thousand/1512477451_full_vector_encoder_binary8_0_32_655360'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512478066_full_vector_encoder_binary8_0_32_1048576'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512478997_full_vector_encoder_binary8_0_32_2097152'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512480568_full_vector_encoder_binary8_0_32_4194304'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512483577_full_vector_encoder_binary8_0_32_8388608'\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512489653_full_vector_encoder_binary8_0_32_16777216'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print latencies[2]/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/full_vector_gen_32.eps')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.10656660413\n",
      "1.70058183145\n",
      "1.73929909152\n",
      "1.73068158753\n",
      "1.68904305888\n",
      "1.48307229725\n",
      "1.62138957729\n",
      "1.9280537279\n",
      "1.93999208206\n",
      "t\n",
      "2.27817369247\n",
      "1.85602005612\n",
      "1.85846344363\n",
      "1.85846580124\n",
      "1.83780143833\n",
      "1.78252269969\n",
      "1.46964666906\n",
      "1.5684931916\n",
      "1.90267936333\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIRunDQL11wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x111adf850>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8,16,32 \n",
    "## Full Vector - latency comparison\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def read_gen_eight():\n",
    "    one_kb = 'results/encoder/thousand/1512452623_full_vector_encoder_binary8_0_8_128'\n",
    "    one_mb = 'results/encoder/thousand/1512452934_full_vector_encoder_binary8_0_8_131072'\n",
    "    ten_mb = 'results/encoder/thousand/1512453317_full_vector_encoder_binary8_0_8_1310720'\n",
    "    twenty_mb = 'results/encoder/thousand/1512453778_full_vector_encoder_binary8_0_8_2621440'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512454344_full_vector_encoder_binary8_0_8_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512455182_full_vector_encoder_binary8_0_8_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512456573_full_vector_encoder_binary8_0_8_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512459046_full_vector_encoder_binary8_0_8_33554432'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512463692_full_vector_encoder_binary8_0_8_67108864'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512475943_full_vector_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512470804_full_vector_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512468085_full_vector_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512466613_full_vector_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512465749_full_vector_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512465168_full_vector_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512464696_full_vector_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512464308_full_vector_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512463997_full_vector_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_thirdy_two():\n",
    "    one_kb = 'results/encoder/thousand/1512476247_full_vector_encoder_binary8_0_32_32'\n",
    "    one_mb = 'results/encoder/thousand/1512476560_full_vector_encoder_binary8_0_32_32768'\n",
    "    ten_mb = 'results/encoder/thousand/1512476958_full_vector_encoder_binary8_0_32_327680'\n",
    "    twenty_mb = 'results/encoder/thousand/1512477451_full_vector_encoder_binary8_0_32_655360'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512478066_full_vector_encoder_binary8_0_32_1048576'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512478997_full_vector_encoder_binary8_0_32_2097152'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512480568_full_vector_encoder_binary8_0_32_4194304'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512483577_full_vector_encoder_binary8_0_32_8388608'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512489653_full_vector_encoder_binary8_0_32_16777216'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "\n",
    "def calc_diff(x,y):\n",
    "    print (y/float(x))\n",
    "    \n",
    "def diff_gen(g, gg):\n",
    "    for x, y in zip(g,gg):\n",
    "        calc_diff(g[x],gg[y])\n",
    "\n",
    "eight_latencies = read_gen_eight()\n",
    "eight_data_size = np.array(eight_latencies.keys())\n",
    "eight_latency = np.array(eight_latencies.values())\n",
    "\n",
    "sixteen_latencies = read_gen_sixteen()\n",
    "diff_gen(eight_latencies,sixteen_latencies)\n",
    "sixteen_data_size = np.array(sixteen_latencies.keys())\n",
    "sixteen_latency = np.array(sixteen_latencies.values())\n",
    "\n",
    "thridy_two_latencies = read_gen_thirdy_two()\n",
    "print('t')\n",
    "diff_gen(sixteen_latencies,thridy_two_latencies)\n",
    "thridy_two_data_size = np.array(thridy_two_latencies.keys())\n",
    "thridy_two_latency = np.array(thridy_two_latencies.values())\n",
    "\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(eight_data_size, eight_latency)\n",
    "\n",
    "plt.xticks(sixteen_data_size, my_xticks)\n",
    "plt.plot(sixteen_data_size, sixteen_latency)\n",
    "\n",
    "plt.xticks(thridy_two_data_size, my_xticks)\n",
    "plt.plot(thridy_two_data_size, thridy_two_latency)\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['G=8', 'G=16', 'G=32'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/full_vector_gen_comparison.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY8AAAEnCAYAAABR1c9kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10e9894d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8 \n",
    "## Full Vector - throughput \n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[5] for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "    \n",
    "one_kb = 'results/1510751197_full_vector_encoder_1_0_8_128'\n",
    "one_mb = 'results/1510751438_full_vector_encoder_1_0_8_131072'\n",
    "ten_mb = 'results/1510752389_full_vector_encoder_1_0_8_1310720'\n",
    "twenty_mb = 'results/1510822030_full_vector_encoder_1_0_8_2621440'\n",
    "thirdy_two_mb = 'results/1510824925_full_vector_encoder_1_0_8_4194304'\n",
    "sixty_four_mb = 'results/1510836357_full_vector_encoder_1_0_8_8388608'\n",
    "one_hundre_and_twenty_eight_mb = 'results/1510847597_full_vector_encoder_1_0_8_16777216'\n",
    "two_hundre_and_fifty_six_mb = 'results/1510927849_full_vector_encoder_1_0_8_33554432'\n",
    "five_hundre_and_twelve_mb = 'results/1510972685_full_vector_encoder_1_0_8_67108864'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MiB','10MiB','20MiB','32MiB','64MiB','128MiB','256MiB','512MiB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "plt.plot(data_size, latency)\n",
    "plt.legend(['G=8'])\n",
    "plt.ylabel('Bytes Per microsecond')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/full_vector_through_put_gen_8.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1127efd10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 16 \n",
    "## Full Vector - throughput \n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[5] for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "    \n",
    "    five_hundre_and_twelve_mb = 'results/1511427894_full_vector_encoder_1_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/1511454980_full_vector_encoder_1_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/1511470001_full_vector_encoder_1_0_16_8388608'\n",
    "    sixty_four_mb = 'results/1511479334_full_vector_encoder_1_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/1511516065_full_vector_encoder_1_0_16_2097152'\n",
    "    twenty_mb = 'results/1511524913_full_vector_encoder_1_0_16_1310720'\n",
    "    ten_mb = 'results/1511529363_full_vector_encoder_1_0_16_655360'\n",
    "    one_mb = 'results/1511534339_full_vector_encoder_1_0_16_65536'\n",
    "    one_kb = 'results/1511534688_full_vector_encoder_1_0_16_64'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MiB','10MiB','20MiB','32MiB','64MiB','128MiB','256MiB','512MiB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "plt.plot(data_size, latency)\n",
    "plt.legend(['G=16'])\n",
    "plt.ylabel('Bytes Per microsecond')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/full_vector_through_put_gen_16.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMcYEZAnCGGNMQP8fm4L0K3CMgOgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1060da190>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 32 \n",
    "## Full Vector - throughput \n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[5] for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "    \n",
    "one_kb = 'results/1511252697_full_vector_encoder_1_0_32_32'\n",
    "one_mb = 'results/1511252611_full_vector_encoder_1_0_32_32768'\n",
    "ten_mb = 'results/1511252293_full_vector_encoder_1_0_32_327680'\n",
    "twenty_mb = 'results/1511221602_full_vector_encoder_1_0_32_655360'\n",
    "thirdy_two_mb = 'results/1511217072_full_vector_encoder_1_0_32_1048576'\n",
    "sixty_four_mb = 'results/1511210497_full_vector_encoder_1_0_32_2097152'\n",
    "one_hundre_and_twenty_eight_mb = 'results/1511201901_full_vector_encoder_1_0_32_4194304'\n",
    "two_hundre_and_fifty_six_mb = 'results/1511187820_full_vector_encoder_1_0_32_8388608'\n",
    "five_hundre_and_twelve_mb = 'results/1511156569_full_vector_encoder_1_0_32_16777216'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "plt.plot(data_size, latency)\n",
    "plt.legend(['G=32'])\n",
    "plt.ylabel('Bytes Per microsecond')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/full_vector_through_gen_32.eps')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "336.3206\n",
      "2043.096\n",
      "1494.6857\n",
      "1437.9984\n",
      "1407.4016\n",
      "1142.202\n",
      "989.0439\n",
      "984.4317\n",
      "982.3868\n",
      "126.422\n",
      "1217.3281\n",
      "858.5034\n",
      "821.6987\n",
      "800.3391\n",
      "749.1728\n",
      "603.6961\n",
      "507.4595\n",
      "506.3925\n",
      "54.3936\n",
      "660.137\n",
      "466.4345\n",
      "440.2093\n",
      "431.6178\n",
      "422.3622\n",
      "404.3491\n",
      "326.4833\n",
      "264.8809\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xd8XNWZ+P/PM9KoWJLlIsmqtmXL3XLBcu8FsCGmGkhI\nsmwIEPiyLOEXQgJhEzZlaUs2IQEn9EAoNjUQGwgYFyxXufeCm4ptuci2itXP7497JY9llasympH0\nvF+vQTN3zr3zaCzmmXvOuc8RYwxKKaVUTS5fB6CUUso/aYJQSilVK00QSimlaqUJQimlVK00QSil\nlKqVJgillFK10gShlFKqVpoglFJK1UoThFJKqVoF+jqA5oiKijK9e/f2dRhKKdWmbNiw4aQxJrqh\ndm06QfTu3ZuMjAxfh6GUUm2KiBx20k67mJRSStVKE4RSSqlaaYJQSilVK00QSimlaqUJQimlVK00\nQSillKqVJgillFK10gThR46cKuLtdUc4V1zm61CUUqptXyjX3vz6nzv5ctdxfvvPndwyuic/mNib\npG6dfB2WUqqDapMJQkTmAnNTUlJ8HUqLyT1XzNI9uVw7Ih4BXl99iNdWHWTO0DjumJzMyJ5dfR2i\nUqqDaZMJwhjzCfBJWlranb6OpaW8vzGbikrD/TP70Sc6nJ/NGchrqw7x1tojLNp2lLReXbljcjKX\nD44lwCW+Dlcp1QGIMcbXMTRZWlqaaQ+1mIwxzHxmOVHhwSy8e/xFzxWUlLNwfSavpB8kK+88vbp3\n4vaJycwblUhYcJvM70opHxORDcaYtIba6SC1H1h/KI8DJwu5eXTSJc+FBwdy+6Rklj04jee/exnd\nwoL41cc7mPDEVzz52W6Onyv2QcRKqY5Av4L6gQXrMwkPDuSq1Ng62wQGuLgqNY6rUuPYcDiPl74+\nwF+Xf8NLXx9g7vB47pjUh8HxnVsxaqVUe6cJwsfyi8tYvO0o141MoFOQs3+OUb26MqrXKI6cKuKV\n9IMszMjkg43ZTEzpzh2T+jC1fzQuHadQSjWTdjH52CdbjnK+rIJbauleakjP7p147JohrP75TH42\neyD7cwv4wWvrueIPK3hn3RGKyyq8ELFSqqPQQWofu/a5dIpLK/jsx5MRad63/tLyShZty+HFFQfZ\nefQc3cOC+P74Xnx/XC+6hwe3UMRKqbZOB6nbgD3H8tmSeYabRyc1OzkABAW6uH5kIov+cxJv3TmW\n4Uld+MOX+xj/xFc8/MFW9ucWtEDUSqmOQscgfGjB+kzcAcL1IxNa9LgiwoS+UUzoG8X+3HxeXnmI\nDzZm8fa6TGYMjOGOScmM79u9RZKSUqr90jMIHykpr+DDTVlcMTiWbmFBXnudlJgIHr8hlVU/n8GP\nZ/VjS+YZbn1pLVc/u5IPN2VRWl7ptddWSrVtmiB85Iudx8krKqv12gdv6B4ezI9n9Sf95zN44oZU\nSisqeWDBFqY8tZT5y77hbJEWCFRKXUwHqX3k+y+v5ZvcAr7+2QyflM6orDQs33eCl74+QPr+U3QK\nCuDmtCRun5hMz+5aIFCp9szpILWOQfhAVl4RK/ef5L4Z/XxWV8nlEqYPiGH6gBh25Jzl5ZUHeXPt\nYV5ffYgrh8Ryx+RkRvXq5pPYlFL+QROED7y3IQuAm0Yl+jgSy5D4SH5/8wgeunIgf1t9iDfXHObT\n7ccY2bMLd07uwxWDexAYoL2RSnU02sXUyiorDZOfWkpyVBh/v2Osr8OpVWFJOe9tyOLllQc5crqI\npG6h3HhZItERwUSEuIkIDiQ8JJDwYOsWYd/XJKJU26BdTH4q/ZuTZJ85z8/mDPR1KHUKCw7ktgm9\n+d64Xnyx8zgvfX2AP3y5r8H9Qt0BhIcEXpRArOThrk4inturnvNMMuEhgbg10SjlFzRBtLIF6zOJ\nDHVzxeAevg6lQQEuYfbQWGYPjaWwpJyCknLyi62fBcXl5BeXkW/fL/B4Pr+4rLrN4VNFF22rdHDC\nGuJ2XZxUqhLIRcnHej4sOICQwACCAl0EBwYQ7HYRXHU/0GU/tu4HBboIdIle/6GUQ5ogWlFeYSn/\n2nGcW8f2JMQd4OtwGiUsOJCw4EB6NKNgrDGG82UVVnKpSjbF5RSUlFUnnpo/C4qt546cLrqwraSc\nCieZphYuod5EEhTgqjfJBFcno5rbXQS7L9yPDHWTHBWmyUi1aZogWtFHm7Mprajk5rTWufbB34gI\nnYIC6RQUSEwzjmOMobissvqspLSikpKySkrKKykpr6i+X1pRUev2kvIK62dZpbVvjecKC8vtxxWU\nlldttx6XVThPTAldQpkxMIaZg2IY16d7m/tSoJQmiFZijGHB+kxSEyJ13YZmEhFCgwIIDQpoVqJp\niopKYycNK8lU3S+ukXyOnrHWGH9vQxZvrDlMp6AAJqVEMWtQD6YPjCE6QosnKv+nCaKVbMs+y+5j\n+fzmuqG+DkU1Q4DrQnJqyK1je1JcVsHqA6dYsus4X+3K5V87jwMwPKkLswbGMGNQDIPjOmtXlPJL\nmiBayYL1mQQHurhmeLyvQ1GtKMQdUH1BornWsOtoPkt2HWfJ7lx+/+VenvliL3GRIcwYGMOsQT0Y\n31e7opT/0ATRCs6XVvDx5hyuSo0jMtTt63CUj4gIg+M7Mzi+M/fN7MeJ/BKW7sllya7jfLgpmzfX\nHiHUHcDElChmDYphxsAYYjqH+Dps1YFpgmgFn24/Sn5JeYcdnFa1i44I5ua0JG5OS6K4rIK1B09b\nZxe7cvlyl9UVNSwxkpkDezBzUAxD4rUrSrUuvZK6Fdzy19UcO1fMsgen6f/gqkHGGPYcz2fJLuvs\nYlPmGYyBHp2DmTGwB7MGxTAxJUq7olST6ZXUfuLQyULWHjzNT68c0HByKC2Cs1kQ3b91glN+SUQY\nGNuZgbGduXd6CicLSli25wRLdh3n483ZvL3uCCFuFxP7RjFzUA9mDIwhNlK7olTL0wThZQszMnEJ\n3HiZg8J8S34Na+dD3xkw9efQ0z9rNanWFRUezLxRicwblUhJeQXrDp6u7oZasjsXgKEJnau7oobG\nR+LyUZVg1b5oF5MXlVdUMuGJrxiaEMkr/z66/sbGwP8NAXcnOJ8HRSetRDHtYUga0zoBqzbFGMO+\n3ILqrqiNR/KoNBATEczMQTHMGNiDSSlRjqbkqo5Fu5j8wPK9J8jNL3E2OH10C5zLhmufhyHXwfqX\nIf2P8PLlmihUrUSE/j0i6N8jgnum9eV0YSnL9uSyZFcu/9xylLfXWVOrJ/TtzsxB1tlFXGSor8NW\nbYieQXjRXa9nsPFIHqsfntlwhdKl/wMrnoYH90NYd2tbaeGFRKFnFKoRSssrWX/I6opasvs4h08V\nATAwNoLLenVlWEIkwxK70L9HuJZp74CcnkFogvCSE/kljH98CbdPSuaRqwY1vMP8SRAcAbd/eulz\nmihUMxhj+OZEIUt2HefrfSfZmnWGc8XlgFU5d0h8JMMSq25dSO4epmMY7Zx2MfnYBxuzKK80zrqX\n8g7D8W1wxW9rfz4oDCb+J4z+Iax/yaPraSZM+7kmClUvESElJpyUmHB+NLUvxhgOnSpia9YZtmad\nZWvWGd5Zl8mr6YcAiAgJJNU+w6hKHAldQnWKdgekCcILjDEsyMgkrVdXUmLCG95hj33WMOCq+tsF\nhcHE+2H0HZooVJOJCMlRYSRHhXHtiATAmlCx/0QBWzPPsjXbShwvrzxQXb22e1hQ9RlG1U8tONj+\n+U2CEJFBwP1AFLDEGDPfxyE12YbDeRw4Ucjd8/o622HPIogaAN0dtq9KFGk/hIyXNVGoZgsMcFVf\ne3HzaOust6S8gt1H89madYYtWWfZlnWW5Xv3VS/6FB8ZYiWMpEiGJXQhNTFSS8m0M15NECLyCvAt\nINcYM9Rj+2zgj0AA8JIx5gljzC7gbhFxAa8DbTZBLFifSVhQAFenxjXc+HweHEq3upAaKzi8nkTx\nMCQ1MLVWqXoEBwYwPKkLw5O68H17W2FJOTtyznkkjTN8tuNY9T7JUWEXnWkMie9MpyC/+R6qGsnb\n/3KvAX/G+sAHQEQCgOeAy4EsYL2IfGyM2Ski1wD3AG94OS6vKSgpZ9G2o1wzPJ6wYAdv774vwFTA\ngKub/qK1JopZkDLLuuBOE4VqIWHBgYxJ7saY5G7V284WlVV3S23JPMO6g6f5x+YcwFrBr3+PiIuS\nxsDYzgQF6syptsCrCcIYs0JEetfYPAbYb4w5ACAi7wDXAjuNMR8DH4vIIuAtb8bmLf/ckkNRaUX1\naXqDdi+C8B6QMKr5L+6ZKNa/BKue1UShvC6yk5vJ/aKZ3C+6elvuueLqAfAtWWf5YudxFmZkARAU\n4GJQXMRF4xkpMeEE6Mwpv+OLc78EINPjcRYwVkSmATcAwcDiunYWkbuAuwB69uzpvSibaEFGJv1i\nwhmZ1KXhxuUlsH8JDL0BXC34jSo4HCb9+MJgtiYK1cpiOocwa3AIswb3AKyJG1l55z2Sxhk+3JTN\nG2sOA+AOENwBLlwiCCACLpd4PBZcYm8Xe3vNx9T2/MX7VT/mQjvEOtOpdT/7tatSlzWRy27jsU3s\nbZ6Pqxp4HkNq2Va1QRwc1/O1b5vQm/49IlruH60WftM5aIxZBixz0O4F4AWwroPwblSNs+94PpuO\nnOHRqwc5mxJ46GsozYeBzeheqk99iWLaw5DY4DRopVqEiJDUrRNJ3Tpx9TBrbK6y0nDgZCFbs86w\n93gBFZWVVBqoNAZjrKRS/Rj7ceWFx7W2M2C4uF3V8/W1qzTU+frWUbC3UX3MKqZmG4/njf2fmtsu\n2cdUHaueNlS1s7ZdnRrXLhNENuDZ/5Job2vzFqzPxB0gXD8ywdkOuxeDOwySp3o3sNoSxUszIeVy\na9aTJgrlAy7XheszlH/yxUjReqCfiCSLSBDwbeBjH8TRokrLK/lgUzazBvWge7iD+eHGWNc/pMwA\ndyuVaq5KFPdvhVn/DTkbrUTx93mQ5Z9XpCulfMerCUJE3gZWAwNEJEtEfmiMKQf+A/gc2AUsNMbs\n8GYcrWHJruOcLix1Pjidswnycxq+OM4bLkoUj0H2Bk0USqlL1NnFJCI31LejMeaDhg5ujPlOHdsX\nU89AdFu0ICOTuMgQpnjM5KjXnsUgLuh3pXcDq09wOEx6AEbfCetfhHTPrqeHIbEFZlYppdqs+sYg\n5to/Y4AJwFf24+nAKqDBBOEtIjIXmJuSkuKrEC5y9Ox5Vuw9wb3TU5xP1du9GHqOv1C51ZdqTRQz\noN8V1qwnTRRKdUh1djEZY35gjPkB4AYGG2NuNMbcCAyxt/mMMeYTY8xdkZGRvgyj2nsZWVQauGmU\nw+6lvEOQu8M33Uv1qUoUP95mdT1lZViJ4s2brOm4ubvh3FFradQ2XAVYKeWMk1lMScaYox6PjwP+\ndwGCj1RWGhZuyGRC3+707N7J2U677d61gX6WIKrUdkaxr0aPo8sNIZEXbqFdLn5cfatje2DIhUnh\nSim/5CRBLBGRz4G37ce3AF96L6S2Zc2BU2SePs+DVwxwvtOexRA9CLr18V5gLcEzUWSuheKzNW5n\nLn58NuvC/fLi+o8dEFRHQqktsdSSZFpr5pdSHViDCcIY8x/2gPVke9MLxpgPvRtW27EgI5POIYFc\nOSTW2Q5Fp+HwKmsWUVsRHA4pMxu3T1kxlJy7OJmcP1NLkvG4ncm80LaitP7jBwRbiSI4AoI6QWAo\nuEOtNb3dIfbPUOtMpeq+26PNRds923u0CfCb60iV8glH/wfYM5Z8Nijtr84WlfHp9mN8e3QSIW6H\nC8O3RHG+tsAdYt3CY5q2f1nxpWcotZ21FJ+DsvNQft76ef60ta/ntrKipsXgCmwg0YTWnZiCwiAs\nxqqzFW7/1LMe1cY0mCDss4cnsWYziX0zxpjOXo6tvpj8YhbTR5uzKS2vdLZqXJU9iyA8FuJHei+w\n9sAdAu5YiHB4ZlYfY6y6V2VFduIovnC/zCOJXLTd4375+Rptz1tl2s/l2G2KLxyjsqzuOEIi7YTh\neYuxfseqJBLeA0K7tWxtLqWayMkZxFPAXHu9Br9gjPkE+CQtLe1OX8axYH0mQ+I7MzTB4WyqquJ8\nqfP0A6A1iVw4o/G2inIroZTkQ0GufTte45ZrXcWefxzKCi89hivQPvuwk0ZEjYQS7pFQghxOjFCq\nCZwkiOP+lBz8xfbss+w8eo5fXzvE+U4HV0BpQfvvXurIAgIhIMIaG+kc33D7koILSaPg2MUJJf84\n5B+Fo1ugMBdM5aX7B0XUchZSlUQ8urfCosDlsBtUKZuTBJEhIguAj4CSqo1OrqRuzxaszyQo0MW1\nwx0W5gNr7Qd3GCRP8V5gqm0JDrduDS03W1kBRacuPgvJ90wouXBsGxQssSYH1CQuCIuGyEToO8O6\nBid+pE41VvVykiA6A0XAFR7bDB140Lq4rIKPNmczZ2gskZ0cXjNYWWkX55upg5Wq8VwB9tlADJBa\nf9vSIuuMozqJHL+QSE7uha+fgRVPQ+cEGDDHSha9J0NgUKv8KqrtcDLN9QetEUhb8tn2Y+QXl3NL\nYwanczZZXQjeWvtBqSpBnSCoN3TtXfvzhadg3+fWGe3mt6wS8MGdrXVCBl4N/S63BtRVh+dkFlMi\n8Cdgor3pa+B+Y0yWNwPzZwvWZ5LULZRxfRpRR2nPIpAAq76RUr4U1h1G3Grdys7DgWVWstj7Gez4\nwLpKvvckK1kMmGN1S6kOyUkX06tY60PfZD/+nr3tcm8F5c8Onypk9YFT/OTy/rgas4bunk+h1wTo\n1K3htkq1Fneo3c00xxrnyMqA3f+0rvZf/KB1ixthJ4uroMcQHbfoQJwkiGhjzKsej18TEZ9eBuzL\n6yDezcjCJTAvrRHfqk4fhNydcOXj3gtMqeZyBUDPsdbtit/Aib3Wme/uxbD0f2Dp76BLT2sW3sCr\nrWrEerV5u+bkX/eUiHyPC7WYvgOc8l5IDfPVdRAVlYb3NmQxpX80cZGhznfcYxfnGzDHO4Ep5Q3R\n/a3bpAesKbd7P7WSRcYrsHY+hHa11jMZeBX0nWnNxlLtipMEcTvWGMT/Yc1eWgV0yIHrFXtPcOxc\nMb+aO7hxO+5eDDGDoVtyvc0WHVjEGzvf4LuDvsuc5DkEuvTbmfITET1g1L9bt5IC+OYra9xi3+ew\n9R2rNlafaVay6D/Haq/aPCezmA4D17RCLH5vwfpMuocFMXNQI/74i07DkVUw6f9rsOm7e99l56md\nPLLyEf6y5S/cNewuru5ztSYK5V+Cw2HwNdatohyOrLbOkqsSBj+GxDR73OJq6yxEtUkN1nsQkb+J\nSBePx11F5BXvhuV/ThaU8OWu41w/MoGgwEaUydj7uXUFbANrPxSUFrAldws/GPoD/jD9D3Ryd+LR\n9Ee55qNr+Gj/R5RXljfzN1DKCwICIXkyzH4c7t8Cd6fD9Eegogy+fAyeGw1/GgX/+i84ssYaCFdt\nhpOvpsOMMWeqHhhj8kSkw1Wa+3BjNuWVhltGN+LaB7AG+SLiIK7+t2ztsbWUm3ImJUxidOxoZiTN\nYGnmUv6y5S/8V/p/8cLWF7gz9U6+1fdbuF0+XdBPqdqJQOxQ6zb1ITibfeHMYs3zsOpZ62ru/rOt\ns4s+06xZVMpvOUkQLhHpaozJAxCRbg73azeMMSzIyGRkzy706xHhfMeyYtj/FQy/pcHifKuyVxHm\nDmNE9AgARIQZPWcwPWk6yzKXMX/LfH656pe8sPUF7hp2lyYK5f8iE2DMndat+KxV6n7PYtj5D9j0\nhlUavarsR//Z/rE+u7qIkw/6Z4DVIvKu/fgm4HfeC8n/bDxyhv25BTxxQwMlDmo6uNyq1tlAcT5j\nDOk56YyJHYM74OIPfRFhes/pTEuaxoqsFTy/5Xl+ueqX/HXrX7lr2F3M7TtXE4XyfyGRVhXj1HlQ\nXgqHvrbPLhZb110gEJtqlfzoPdGaQqvXDPmcGAeLz4vIYGCG/fArY8xOr0bVcDxV10HcuW/fPq+/\n3s/e28onW3NY94tZhAc34uTpk/th2/vw0DcQGFxns4NnD3LNR9fwX+P+i5sH3FzvIY0xfJ39Nc9v\nfp4dp3aQEJ7Anal3ck3fay5JLkr5PWPg6GbY+y8raWStt5ertburek2yrurWi0xblIhsMMakNdjO\nYYKYBPQzxrwqItFAuDHmYAvE2SxpaWkmIyPDq69RWFLOmN99yVWpcTx903DnO1ZWwu8HWt+Ebv5b\nvU3f3PUmT6x7gk9v+JTECGcX4FUlir9s+QvbTm4jPiyeO4bdwXV9r9NEodqu8hLI3gCHVloJI3Pd\nhYTRY4iVLHpPgl4TNWE0g9ME4aQW06+ANGAAVokNN/B3LtRmatcWbT1KYWlF4wenczZa1TMdFOdb\nmb2S3p17O04OYHU9TUmcwuSEyazMXslftvyFX6/+NS9ufZE7Uu/g+pTrNVGoticw2Dpb6DXBGugu\nL4HsjVbCOLwSNvwN1v7FahtTlTAmWgkjLMq3sbdDTvpLrgdGAhsBjDE5ItKIkdq2bUFGJn2iwxjV\nq2vjdtxtF+dLmVVvs5KKEjKOZXBDvxuaFJ+IMDlxMpMSJpGek878LfP5zZrf8NK2l7gj9Q6uS7mO\noAAt46zaqMBg6DXeuvFTa/wiZ6N1dnEo3RrsXvdXq23M4AtnF70nacJoAU4SRKkxxoiIARCRMC/H\n5Df25+az4XAeD88ZiDS2QNmexY76TTcc30BxRTETE5p3QiYiTEqYxMT4iazOWc3zW57nN2t+w4vb\nXuSOoXdwfb/rNVGoti8wCHqOs25TqhLGJithHE6HTX+HdS9YbaMHWWcXvSdZYxnh0b6NvQ1ykiAW\nishfgS4icidW6Y0XvRuWf1iYkUWgS7jhskaWOz71DZzYbZUlaMCq7FUEuYJI69Fgd6AjIsKEhAmM\njx/P6qOrmb95Pr9d+1srUaTewQ39btBEodqPwKALBQZ50LpAL2eTPYaxEja/ba13ARA14MIYRu9J\n9uJLqj5OB6kvx1pRToDPjTFfeDswJ7w5SF1WUcn4x5dwWc+uvPBvjfzwXvUn+NejcP9W6Nqr3qbX\n/+N6okKjePEK7+RcYwxrjq5h/pb5bMrdREynmOpEERxQ98wqpdqFijJrTe9DX1sJ48gaa114gKj+\nHoPekzpU/aiWHKQOw5ra+oWIDAAGiIjbGFPWEoH6qyW7cjlZUNr4wWmw5nb3GNpgcjhWeIz9Z/Zz\nXcp1TYyyYSLC+PjxjIsbx9pja5m/eT7/s/Z/eGnrS9yeejvz+s/TRKHarwC3VRcqMc2qSltRfnHC\n2PquVZ0WoHu/i2dJdY7zbex+wEkX0wpgsoh0BT4DMoBbgO96MzBfW5iRSUxEMFP7N7LfsvAUZK6B\nyQ822HRVzioAJsRPaEqIjSIijIsbx9jYsaw7to75W+bzxLoneGXbK9yeejs39ruRkEBdK1u1cwGB\nkDjKuk36sZUwjm250CW17T3YYC9/0z0FksZaV3xXj0GKfd9+XHXfc4zykm1S/7ZL9qm5jdqPM/i6\nBitEN5eTBCHGmCIR+SEw3xjzlIhs9mpUPnbsbDHL9uRyz7S+BAY0ojAfWNUsHRTnA2t6a0ynGFK6\ntN7CRyLC2LixjI0by/pj63l+8/M8se4JXt72MrcPtc4oNFGoDiMgEBJGWbeJ99sJY6s9rTYd9n8J\nFaVWW2MAYy16gKmxraqr3lzYdsnz9ezTFD1S/SNBiMh4rDOGH9rbArwXUsO8vaLc+xuzqDRwc1pT\nupcWQecEa5nGepRXlrPm6Bpm9ZzV+BlSLWR07Ghenf0q64+tZ/6W+Ty5/kle3m4lipv636SJQnU8\nAYGQcJl1m/ifrfvappGJphUmmzj5evxj4GHgQ2PMDhHpAyz1blj1M8Z8Yoy5KzIyssWPXVlpWJiR\nybg+3ejVvZEzesvOWwupDJjT4Lq9209uJ780v9nTW1vC6NjRvHLlK7xy5Sv0iezDU+ufYvb7s3l9\nx+ucLz/v6/CU6hhErKKeLpe1/GtAoDWGEuC2ZmsFBoM7xKqA6w612nhZgwnCGLPcGHMN8JyIhBtj\nDhhjWjm1tp61B09z+FRR0wanDyyHsiJHS4um56TjEhfj4sY1IUrvGB07mpevfJlXr3yVlC4pPJ3x\nNHPen8PfdvxNE4VSHZCTWUypwOtAN+uhnAD+zRizw9vB+cLCjEwiQgKZM7QJMxj2LIKgCKsiZQPS\ns9NJjUolMrjlz4KaKy02jZdiX2LD8Q3M3zKf/834X57d+CzdQrvRNbgr3UK70S24G11DutI1pCvd\nQrrRLcR6XLU9zB3ms64zpVTLcDIG8Vfg/zPGLAUQkWlYF8p5f+pNKzt7vozF245yU1oiIe5Gnr5V\nVsKez6DfrHortwLkFeex/eR27hlxTzOi9b5RPUbx0hUvsSl3E0uPLOVU8SnyivPIK87j0NlDnC4+\nXeeZRZArqDp5XJJIgi9NLOHucE0oSvkZJwkirCo5ABhjlrXXchsfb8mhpLySW9J6Nn7n7AwozG1w\n7QeANUfXYDBMjPf9+IMTI2NGMjKm9hXxzpefr04ap4tPc7r4tHW/5PRF2w+fO0xecR5F5UW1Hsft\nclefndSWQKqTjd0mwh2hCUUpL3OSIA6IyH8Bb9iPvwcc8F5IvrNwfSaD4jozNKFz43fevQhcgdDv\n8gabrsxeSWRwJEO6D2lClP4lNDCU0PBQ4sPjHbUvLi+uNYFUJZaq5zJPZJJXkkdhWWGtxwl0BdI1\nuCuRwZEEBQQRHBBMkCsId4CbIJf12B3gJigg6OLHriBrW9XNVeNnze32/apjBQcE43a5cUkjpz8r\n1QY5SRC3A/8NfIA1v+pre1u7sjPnHNuyz/LY3MFN+2a6Z7F19WVol3qbGWNYlbOK8XHjCWiFWQj+\nJiQwhLjwOOLCnY3xlFSUVCcRz595Jdb9syVnKa0opbSylLKKMgrLCimpLKGsoqx6e0mF/biylEpT\n2SK/R6ArsNbE4/k4OCD4QvKyfzrZVnPf2vYJCgjSJKW8rt4EISIBwC/a86ylKgszMgkKdHHdyITG\n73xyP5yzlUnKAAAgAElEQVTcC6PvaLDp3ry9nDx/0i+mt7YFwQHBxIbFEhsW2yLHK68stxKHnTwu\n+llzu0fiKa2wEk31YzvxlFaUUlZZdsn+JRUl5JfmU1JRUt3uop+Vpc3+Xdwud62JxDNZ1UwqIYEh\nRARFEOGOICIogs5Bna3HHrdwd3iH/PKiLlVvgjDGVNirybVrxWUVfLgpmyuHxNKlUxMuPtmz2Prp\ncHor0GbGH9qbQFcgga5AOrk7+TSOSlNJWWXZJcmj6n5tScXxtkrrfmF5IXkleRc9V1xeTEFZQYPx\nhbvDL0kcFyUTTTAdgpMupk0i8jHwLlDdIWyM+cBrUbWyf+08ztnzZdzSlCunwUoQsanQpeHB7fTs\ndPp37U90J61N35G5xFX9Db+1VVRWUFheSH5pfvXtXOm5ix7X3Ha04Ch7S/daj8vyG3yNxiaYriFd\n6dulr3ab+RknCSIEOAXM8NhmsMYk2oUF64+Q2DWUCX27N37nwpOQudZavKQBRWVFbMzdyPcHf78J\nUSrVMgJcAXQO6kznoCZMxsB7CSY6NJqpSVOZnjSdMbFjtNSLH2gwQRhjftAagTRGS9ZiyjxdRPr+\nUzwwqz8uVxMGp/d+ZhXnG9Bwcb51x9ZRXlmu3UuqTfNGgskpyGFF1goWH1jMe3vfIzQwlAnxE5iW\nNI2piVPpGtLIJX9Vi3ByJfXfgPuNMWfsx12BZ4wxPpvJZIz5BPgkLS3tzuYe692MTERgXlojV42r\nsnsxdE6EuOENNl2ZvZLQwNA6rylQqiOoK8Fcm3ItpRWlrD+2nqWZS1mauZQlR5bgEhcjokcwPWk6\n05Km0Tuyt28C74CcdDENq0oOAMaYPBFpF59wFZWGdzdkMblfNAldQht/gNIiqzjfyO81WJwPrPUf\nxsSO0SU/lapDUEAQExMmMjFhIr8Y+wt2nt7JssxlLD2ylGc2PMMzG54hOTKZaUnTmJE0g9SoVB0Q\n9yInCcIlIl2NMXkAItLN4X5+7+t9Jzh6tphHrx7ctAMcWAbl5x2t/XDk3BEy8zN1/EEph0SEId2H\nMKT7EO4dcS85BTlWsshcyhs73uDV7a/SLaQbUxOnMi1pGuPjxxMa2IQveqpOTj7onwFWi8i7WGsb\nzQN+59WoWsnCjEy6dnIza3ATFy/fswiCO1vr2TZAp7cq1Tzx4fHcOuhWbh10K/ml+azMXsnSzKV8\nefhLPtz/IcEBwYyPG2+NWyRNJSo0ytcht3lOBqlfF5EMLsxiusEYs9O7YXnfqYISvth5nO+P601w\nYBNOUSsr7OJ8l1u12huQnp1OUkQSPTs3oc6TUuoiEUERzEmew5zkOZRVlLEhd0N1V9SyrGXIaiE1\nOpXpSdOZnjSdPpF9tHZXE9SZIESkszHmnN2ldAx4y+O5bsaY060RoLd8uCmbsgrTtHUfALIyoOik\no9lLpRWlrDu2jmv7Xtu011JK1ckd4GZc3DjGxY3jZ6N/xt68vSzNXMqyzGX8ceMf+ePGP9IzoifT\nkqYxPWk6I2JGEOhqF73kXlffu/QW8C1gA9Z1D1LjZx+vR+clxlirxg1P6sKA2IimHWTPInC5HRXn\n25S7ifPl57W8hlJeJiIM6DaAAd0GcPfwuzlWeIwVWSv4KvMr3t79Nq/vfJ0uwV2YkjiFaUnTmBg/\n0edX1fuzOhOEMeZb9k/vrortA5szz7D3eAH/c31q0w+yezH0ngQhDS/4k56dTqArkDGxY5r+ekqp\nRosNi+XmATdz84CbKSwrJD07nWWZy1ietZyPv/kYt8vN2Lix1VNoYzo1cTyynXJ0niUiw4Denu3b\ncqmNhRmZhLoDmDu8CavGAZzcB6f2wZi7HDVPz0nnspjL9JuKUj4U5g7jit5XcEXvKyivLGdT7qbq\nWVG/WfMbfrPmNwzpPqQ6WfTv2r/Dj1s4uVDuFWAYsAOoqpXcpktt3Dm5DxNToogIcTftALsXWT8d\nFOfLLcplb95eHhj1QNNeSynV4gJdgYyOHc3o2NE8mPYgB84eqL4477nNz/HnzX8mITyh+krutB5p\nuAOa+HnRhjk5gxhnjGnihQL+qU90OH2iw5t+gD2LIXYYdGl4gHtVzipAp7cq5a9EhL5d+tK3S1/u\nSL2Dk+dPsjxzOcsyl/He3vd4c9ebhLnDqkt/TE6Y3GFKfzhJEKtFZHB7mNraIgpyIXMdTPu5o+bp\n2elEhUbRv2t/LwemlGoJUaFR3Nj/Rm7sfyPny8+zJmcNy7OWsyJrBV8c/gKXuBgePdwa6E6cRt8u\nfdttV5STBPE6VpI4BpRgz2IyxgzzamT+au9ngHE0vbWisoJVOauYnjS93f4BKdWehQaGMr3ndKb3\nnE6lqWTXqV0sz1p+0RTahPAEpiZOZWrSVEb3GN2uuqKcJIiXge8D27gwBtFx7V4MkT2t9R8asOPU\nDs6VntPprUq1Ay5xMSRqCEOihvD/Rvw/jhceZ0X2CpZnLuf9fe/z1u63qruipiZOZXLiZLqFdPN1\n2M3iJEGcMMZ87PVI2oLSIjiwFC67zVFxvvTsdARhfNz4VghOKdWaeoT14Kb+N3FT/5s4X36etUfX\nWl1RmVZXlCAMix5WPdCd0iWlzfUkOF1R7i3gE6wuJsC301xbcj2IRjmwFMqLHRXnA2t669CooXQJ\n6eLlwJRSvhQaGMq0pGlMS5qGGWfYeXonyzOXszxr+UVdUVXjFmmxaW2iqrMYY+pvIPJqLZuNL9eD\nqJKWlmYyMjJa7wU/uhd2fQIPfQMN9DOeLTnLlAVTuGvYXdw74t5WClAp5W+quqJWZK5gzdE1FFcU\n0ymwExMTJjIlcQqTEybTPbQJq1k2g4hsMMakNdSuTa4o5xOVFdYAdb/LG0wOAGuOrqHSVOr0VqU6\nuJpdUeuOrmN51nKWZy6/qCuqaqC7X5d+ftMVpRWrnMpcZxXnc9q9lJ1ORFAEQ6OGejkwpVRbERoY\nytQkKxGYcYZdp3dVd0U9u+lZnt30LPFh8dW1okbHjvZpV5QmCKeqivOlNFyczxhDenY64+PGa9VI\npVStRITB3QczuPtg7hlxD7lFuazIsmZFfbT/I97Z80712txTE6cyJXFKq3dF1fvpJSIuYJ4xZmEr\nxeOfjLGmtyZPhpCGF2rff2Y/uedzdXqrUsqxmE4xzOs/j3n951FcXsy6Y+uqCwsuObIEwVrjYmri\nVKYmTm2VWlH1JghjTKWIPAR07ARxci+c/gbG3eOoeXq2tXrchPgJ3oxKKdVOhQSGMCVxClMSp2CM\nYffp3SzLWsbyzOX8adOf+NOmP/HM1Ge4ovcVXo3DSf/HlyLyILAAKKza2NYXDGqU6uJ8zqe3pnRJ\nITYs1otBKaU6AhFhUPdBDOo+iHuG38OJohOsyFrB2LixXn9tJwniFvun51zNNr1gUKPt+RTiRkBk\nQoNNi8qK2HB8A7cOvLUVAlNKdTTRnaK5sf+NrfJaTqa5trsFgxqlIBey1sP0Rxw1zzieQVllGRMS\ntHtJKdW2uRpqICKdRORREXnBftxPRL7l/dD8xJ5PcVqcD6zxh5CAEEb1GOXduJRSyssaTBDAq0Ap\nUPWVOBv4rdci8jd77OJ8PYY4ap6ek87o2NEEBwR7OTCllPIuJwmirzHmKaAMwBhThFXyu/0rLYQD\ny6yL4xxMJ8vMz+TwucM6vVUp1S44SRClIhKKNTCNiPTFo2hfu/bNV1ZxPofdS6uydfU4pVT74WQW\n02PAZ0CSiLwJTAT+3Ysx+Y/diyEkEno5G3BOz0knITyBXp17eTkwpZTyPiezmP4lIhuAcVhdS/cb\nY056PTJfqyi3i/Nd6ag4X1lFGWuPruVbfb7lN4W2lFKqOepMECISAzwCpGCtJve4MeZcawXmc5lr\n4fxpx8X5Np/YTFF5kU5vVUq1G/WNQbyOdeX0n4Bw4NlWichf7FkMAUGQMstR8/TsdAIlkLGx3r+6\nUSmlWkN9XUxxxphf2Pc/F5GNrRGQXzDGShDJUyA4wtEuq3JWMTxmOOFB4V4OTimlWke9s5hEpKuI\ndBORbkBAjcft14k9cPqA49lLJ8+fZNfpXUxKmOTlwJRSqvXUdwYRCWzg4mseqs4i2nctpj1Vxfnm\nOGq+Kkentyql2p86E4QxpncrxuFfdi+G+JHQOd5R8/TsdLqFdGNAtwFeDkwppVqPkwvl/I6IzBWR\nF86ePdvyB88/BtkZMOBqR80rTSWrc1YzMX4iLmmTb6dSStWqTX6iGWM+McbcFRkZ2fIH3/Op9dPh\n9NZdp3aRV5Kn01uVUu1Om0wQXrVnMXTpBTGDHTVfmb0SQXT1OKVUu+Ok3HdfEQm2708Tkf8UkS7e\nD80HSgrgwHIYeLWj4nxgDVAP6j6IbiHte2KXUqrjcXIG8T5QISIpwAtAEvCWV6PylW++gooSx9Nb\nz5WeY8uJLTp7SSnVLjlJEJXGmHLgeuBPxpifAnHeDctH9iyG0K7Qc7yj5muPrqXCVOj1D0qpdslJ\ngigTke8AtwH/tLc1XL2urbmoOJ+TIrfW9NZwdzip0aleDk4ppVqfkwTxA2A88DtjzEERSQbe8G5Y\nPpC5Bs7nOb44zhhDek464+LG4Xa1v3yplFINJghjzE7gZ9hXURtjDhpjnvR2YK1ud1VxvpmOmh88\ne5Bjhcd0eqtSqt1yMotpLrAZa9EgRGSEiHzs7cBalTFWeY3kqY6L863MXgloeQ2lVPvlpIvpMWAM\ncAbAGLOZ9laHKXcX5B1yfHEcWNNbkyOTiQ93Vo5DKaXaGkeD1MaYmjUtKr0RjM9UFefr72z8obi8\nmIzjGXr2oJRq15xM19khIrdilfvuB/wnsMq7YbWy3YshYRR0djZ7N+N4BiUVJTq9VSnVrjk5g7gP\nGAKUYF0gdxa435tBtapzRyFno+OL48Ca3hocEMyoHqO8GJhSSvmWkzOIq+2V5apWl0NEbgLe9VpU\nrWlvVXE+Z9VbAdJz0knrkUZIYIiXglJKKd9zcgbxsMNtbdPuxdA1GaIHOmqeU5DDwbMHtTifUqrd\nq/MMQkTmAFcBCSLyrMdTnYFybwfWKkry4eByGH2n4+J86TnpADr+oJRq9+rrYsoBMoBrsJYerZIP\nPODNoFrN/iVQUdq46a3Zq4gNiyU5MtmLgSmllO/Vt+ToFmCLiOQB/zTGtK+prXChOF/SOEfNyyrL\nWHN0DVf2vhJxeMahlFJtlZMxiJuBfSLylIg466hvCyrKYO/n0H+24+J8W09spaCsQLuXlFIdgpNa\nTN8DRgLfAK+JyGoRuUtEnNWk8FdHVkPxmUZPbw2QAMbGjfViYEop5R8cLTlqjDkHvAe8g7UWxPXA\nRhG5z4uxedfuxRAQDH1nON4lPSed4dHDiQhq27lRKaWccFKs7xoR+RBYhrUOxBhjzBxgOPAT74bn\nJVXF+fpMg+BwR7ucOn+Knad26vRWpVSH4aTz/Ubg/4wxKzw3GmOKROSH3gnLy3J3wpkjMNl5flt9\ndDWg01uVUh1HgwnCGHNb1X0RiQJOGWOM/dwSL8bmPbsXWz8dFucDa3pr1+CuDOo+yEtBKaWUf6mz\ni0lExonIMhH5QERGish2YDtwXERmt16IXhAWBcO+DRE9HDWvNJWk56QzPn48LnE0bKOUUm1efWcQ\nfwYeASKBr4A5xpg19lTXt7EXEGqT0n5g3RzafXo3p4tPMzFBy3srpTqO+r4OBxpj/mWMeRc4ZoxZ\nA2CM2d06ofmPVTlWdXMdoFZKdST1nUF4Xjl9vsZzxgux+K2V2SsZ1G0QUaFRvg5FKdUIZWVlZGVl\nUVxc7OtQfCIkJITExETcbneT9q8vQQwXkXOAAKH2fezHHabOdUFpAVtyt3DbkNsabqyU8itZWVlE\nRETQu3fvDlcexxjDqVOnyMrKIjm5abXj6qvFFNDkyNqRtcfWUm7KdfxBqTaouLi4QyYHABGhe/fu\nnDhxosnH0Ck5DViVvYpOgZ0YET3C16EopZqgIyaHKs393TVB1MMYQ3pOOmPjxuIOaFofnlKqYzt+\n/Di33norffr0YdSoUYwfP54PP/zQ0b5vv/02qampDBs2jNmzZ3Py5EkvR3sxTRD1OHTuENkF2UyM\n1+4lpVTjGWO47rrrmDJlCgcOHGDDhg288847ZGVlNbhveXk5999/P0uXLmXr1q0MGzaMP//5z60Q\n9QV+kyBE5DoReVFEFojIFb6OBzymtybo9FalVON99dVXBAUFcffdd1dv69WrF/fd13CdU2MMxhgK\nCwsxxnDu3Dni4+O9Ge4lnC2E0EQi8grwLSDXGDPUY/ts4I9AAPCSMeYJY8xHwEci0hX4X+Bf3ozN\niZXZK+nduTdJEUm+DkUp1Uz//ckOduaca7hhIwyO78yv5g6p8/kdO3Zw2WWX1frcnj17uOWWW2p9\nbtmyZXTp0oX58+eTmppKWFgY/fr147nnnmuRuJ3yaoIAXsO6Ivv1qg0iEgA8B1wOZAHrReRjY8xO\nu8mj9vM+VVJRQsaxDG7od4OvQ1FKtRP33nsvK1euJCgoiPXr17N58+Y625aVlTF//nw2bdpEnz59\nuO+++3j88cd59NFHWy1eryYIY8wKEeldY/MYYL8x5gCAiLwDXCsiu4AngE+NMRu9GZcTG45voLii\nWKe3KtVO1PdN31uGDBnC+++/X/34ueee4+TJk6SlpTV4BrFv3z4A+vbtC8DNN9/ME0884f2gPXj7\nDKI2CUCmx+MsYCxwHzALiBSRFGPMX2rbWUTuAu4C6Nmzp9eCXJW9CrfLTVqPNK+9hlKqfZsxYwaP\nPPII8+fP55577gGgqKgIgAEDBtR7BpGQkMDOnTs5ceIE0dHRfPHFFwwa1LrVpH2RIGpljHkWeNZB\nuxeAFwDS0tK8VvIjPSedUT1G0cndyVsvoZRq50SEjz76iAceeICnnnqK6OhowsLCePLJJxvcNz4+\nnl/96ldMmTIFt9tNr169eO2117wftAdfJIhswHPUN9He5jeOFR5j/5n9XNv3Wl+HopRq4+Li4njn\nnXeatO/dd9990Qyo1uaLaa7rgX4ikiwiQcC3gY99EEedqqa36viDUqoj82qCEJG3gdXAABHJEpEf\nGmPKgf8APgd2AQuNMTu8GUdjrcxeSUynGFK6pPg6FKWU8hlvz2L6Th3bFwOLm3pcEZkLzE1JafkP\n8PLKctYcXcOsnrM6dA0XpZTymyupG8MY84kx5q7IyMgWP/b2k9vJL83X7iWlVIfXJhOEN6XnpOMS\nF+Pixvk6FKWU8ilNEDWkZ6eTGpVKZHDLn50opVRbognCQ15xHttPbtfqrUqpFtOcct+/+MUvSEpK\nIjw8/JLnFi5cyODBgxkyZAi33nprS4cNaIK4yJqjazAYHX9QSrWI5pT7Bpg7dy7r1q27ZPu+fft4\n/PHHSU9PZ8eOHfzhD39o6dABP7qS2h+szF5JZHAkQ7q3fs0WpVT705xy3wDjxtU+Fvriiy9y7733\n0rVrVwBiYmKaH2wt2mSC8MY0V2MMq3JWMT5uPAEuXY5bqXbn05/DsW0te8zYVJhTdwG95pb7rsve\nvXsBmDhxIhUVFTz22GPMnj27EYE70yYThDHmE+CTtLS0O1vqmHvz9nLy/EntXlJKeU1jyn3Xp7y8\nnH379rFs2TKysrKYMmUK27ZtqzepNEWbTBDekJ6TDsCEeF09Tql2qZ5v+t7SnHLf9X3YJyYmMnbs\nWNxuN8nJyfTv3599+/YxevToFo1fB6lt6dnp9O/an5hO3unLU0p1PDNmzKC4uJj58+dXb6tZ7ru2\nW0NnAtdddx3Lli0D4OTJk+zdu5c+ffq0ePyaIICisiI25m7U6a1KqRZVVe57+fLlJCcnM2bMGG67\n7TZH5b4BHnroIRITEykqKiIxMZHHHnsMgCuvvJLu3bszePBgpk+fztNPP0337t1bPn5jvLakgtel\npaWZjIyMZh9nWeYy7vvqPl664iXGxo1tgciUUv5g165drb7Ijr+p7T0QkQ3GmAZXQ9MzCKzpraGB\noYyMGenrUJRSym+0yQQhInNF5IWzZ8+2yPFW5axiTOwYggKCWuR4SinVHrTJBNGS1VyPnDtCZn6m\nTm9VSqka2mSCaElV01t1gFoppS6mCSI7naSIJHp27unrUJRSyq906ARRWlHKumPr9OI4pZSqRYdO\nEJtyN3G+/DyTEib5OhSlVDvVnHLfs2fPZvjw4QwZMoS7776biooKAH7/+98zePBghg0bxsyZMzl8\n+LBXYu/QCSI9O51AVyBjYsf4OhSlVDvU3HLfCxcuZMuWLWzfvp0TJ07w7rvvAjBy5EgyMjLYunUr\n8+bN46GHHvJK/B07QeSkc1nMZXRyd/J1KEqpdqi55b47d+4MWMX5SktLEREApk+fTqdO1ufWuHHj\nHCecxuqwxfpyi3LZm7eXB0Y94OtQlFKt4Ml1T7L79O4WPebAbgP52Zif1fl8S5T7vvLKK1m3bh1z\n5sxh3rx5l7R9+eWXmTNnThOib1ibTBAtsR7EqpxVgE5vVUq1nqaU+/78888pLi7mu9/9Ll999RWX\nX3559XN///vfycjIYPny5V6Jt00miJZYDyI9O52o0Cj6d+3fgpEppfxVfd/0vaWlyn2HhIRw7bXX\n8o9//KM6QXz55Zf87ne/Y/ny5QQHB3sl/jaZIJqrorKC1UdXMzVxanWfnlJKtbQZM2bwyCOPMH/+\nfO655x7g0nLfdSkoKCA/P5+4uDjKy8tZtGgRkydPBmDTpk386Ec/4rPPPvPacqPQQRPEjlM7OFty\nVqe3KqW8qqrc9wMPPMBTTz1FdHQ0YWFhjsp9FxYWcs0111BSUkJlZSXTp0+vHuz+6U9/SkFBATfd\ndBMAPXv25OOPP27x+DtkgkjPTkcQxseN93UoSql2Li4ujnfeeafR+/Xo0YP169fX+tyXX37Z3LAc\n6ZDTXGPDYrku5Tq6hLTs+q1KKdWedMgziOv7Xc/1/a73dRhKKeXXOuQZhFJKqYZpglBKtWtteVnl\n5mru764JQinVboWEhHDq1KkOmSSMMZw6dYqQkJAmH6NNjkG0xJXUSqn2LzExkaysLE6cOOHrUHwi\nJCSExMTEJu8vbTmzpqWlmYyMDF+HoZRSbYqIbDDGpDXUTruYlFJK1UoThFJKqVppglBKKVWrNj0G\nISIngKautRcFnGzBcFqKxtU4GlfjaFyN469xQfNi62WMiW6oUZtOEM0hIhlOBmlam8bVOBpX42hc\njeOvcUHrxKZdTEoppWqlCUIppVStOnKCeMHXAdRB42ocjatxNK7G8de4oBVi67BjEEopperXkc8g\nlFJK1aPdJggReUVEckVku8e2ZSJyyai/iBwSkW0istn+eW0rxfOaiBSJSITHtj+IiBGRKPtxhR3X\nFhHZKCITvBRLNxH5QkT22T+72tun2fHc4dF2hL3tQY/f46Ad524R+VUT40oSkaUislNEdojI/f4Q\nm4iEiMg6+99gh4j8t739afuYW0XkQxHp4oP3rIuIvGcfY5eIjPd47ic1/pa8Flcdf1N1vT9uEfmb\n/f/aLhF52GOfQyLydY1jb646rv07nLW3bRWRL0WkzkWZ6/mbekxEsu3jbBaRqzz2GSYiq+3220Qk\nxEuxeX7uZNjbbrJft1I8PqtE5HIR2WC33yAiM7z1nnlqtwkCeA2Y3Yj2040xI4B5wLOtGM9+4FoA\nEXEBM4Bsj+fPG2NGGGOGAw8Dj3splp8DS4wx/YAl9uMq24GbPR5/B9hSY/+f2u/fCOA2EUluQlzl\nwE+MMYOBccC9IjLYD2IrAWbY/wYjgNkiMg74AhhqjBkG7MX692nNuAD+CHxmjBkIDAd2gfXBCFwB\nHKnR3ltxvcalf1N1vT83AcHGmFRgFPAjEentsV+EHT8iMqiW1/ra/n9iGLAeuLeeuOr6mwL4P/s4\nI4wxi+3XCwT+DtxtjBkCTAPKvBQb2J87HtNVtwM3ACtqtDsJzLXfs9uAN2o839JxAe04QRhjVgCn\na3tORFz2N6Xf1vJ0ZyCvFeN5B7jFvj8NSMf6o65Ni8RWRyzXAn+z7/8NuM7jucNAiIj0EBHB+iD4\ntI7DV9UWLmxCXEeNMRvt+/lYH3YJvo7NWArsh277Zowx/zLGVP1brQE8y2Z6PS4RiQSmAC/bcZYa\nY87YT/8f8BBQc5DRK3HV9jdVz/tjgDD7wzgUKAXOeey6kAv/T3wHeLu217Tjj6Ce/yfq+ZuqyxXA\nVmPMFnufU8aYCm/EVke8u4wxe2rZvskYk2M/3AGEikiwt+NqtwmiHoHAm8A+Y8yjHtuX2qdky4FH\na93TO/YC0WJ1m3wHK2F4Cq063QdeAn7jpTh6GGOO2vePAT1qPP8e1je/CcBGrG/Vnp4Wkc1AFvCO\nMSa3OcHY3yhHAmv9ITYRCbCPkQt8YYxZW6PJ7Vz6QevtuJKBE8CrIrJJRF4SkTCxukizqz7katGq\n/5Y2z/fnPaykcxTrDOd/jTGeyeV9rG/RAHOBT2oca7Id3xFgFvCKkwBq/E0B3Gd3ubxi//8H0B8w\nIvK5WF26D9U4TEvGZoAv7S6ju5z8DrYbgY3GGM9/N6+8Zx0xQfwV2G6M+V2N7dONMUOBVODPIhLe\nijF9AHwbGAt8XeO5qi6mgVjf9l63vwV4jbGmttX85rkQ60Olrm8nVd0SscBMacZYif3evw/82Bjj\n+c3SZ7EZYyrsYyQCY0RkqEe8v8A663uzleMKBC4D5htjRmJ96D4GPAL8sp79Wu3fEmp9f8YAFUA8\nVpL7iYj08djlFJAnIt/G+sZfVOOQVd0lScCrwFMOYqj5NzUf6IPVjXYUeMZuGghMAr5r/7xeRGZ6\nKbZJ9vs8B6vra4qD32MI8CTwoxpPtfh7Bh0zQawCplcNPNVkjPkGOA4Mru15L1mAdWbwhTGmsq5G\nxpjVWPVXGqyh0gTHRSQOwP550bdGY8wxrL7Yy7HGAeqKsQBYhvU/V6OJiBvrf+Q3jTEf+FNs9jHO\nAMgFB+MAAATCSURBVEux+9tF5N+BbwHfNTXmjLdCXFlAlsfZzHtYCSMZ2CIih7AS2kYRiW3FuKrV\n8f7cijVuUmafnaQDNSePLACeo46uEg8fY3Wz1RfDJX9TxpjjdtKvBF7ESlpgvacrjDEnjTFFwGKs\n97TFYzPGZNs/c4EPPWKo6/dItNv9m/05VVOLvWdVOmKCeBnrH32h3Qd6EXt0P5mmFwFsNGPMYeAX\nwPP1tRORgUAA1reFlvYx1uAX9s9/1NLml8DPavTJXsR+T8cCtf0B18s+M3oZ2GWM+b2/xCYi0XJh\nBk4o1gfrbhGZjdXPf439YVIbr8Vlf9BnisgAe9NMrK6HGGNMb2NMb6wPvMvstq0Sl8f+db0/R7Am\nYyAiYViDx7tr7P4h1rfczxt4mUn1xVfX31TVFw7b9ViDw9ivlyoinezffyqws6Vjs7sCI6ruY419\nbK+trd2mC7AI+LkxJr2OZi3ynnlqk0uOOiEib2MN+kaJSBZQPV3PGPN7e4DvDRH5rr15qYhUYA1A\n/twYc7y14rFj+msdu4bafYcAAtxW3//UzYjlCayk+UOs5Hhzzf2MMavqOezTIvIoEIT1rfSDetrW\nZSLwfWCbx+/8iB/EFgf8TUQCsL5ULTTG/FNE9gPBwBd2r98aY8zdrRgXwH3AmyISBBwAfuBkp5aO\nq46/qYep/f15DmvcZAfW3/SrxpitNeLLx+pKoZYe1ar+dAHOAnfUbOChrr+p74jICKzuykPYXTbG\nmDwR+T3WTB8DLDbGLPJCbD2AD+39A4G3jDGficj1wJ+wegkWichmY8yVwH8AKcAvRaSq+/AKz/Gh\nFnzPqumV1EoppWrVEbuYlFJKOaAJQimlVK00QSillKqVJgillFK10gShlFKqVpogVIcmF6rl7hCr\nWutPxCqaWN8+vUXk1hZ6/ZfkQvE4pfyKTnNVHZqIFBhjwu37McBbQLoxps4y1yIyDXjQGPOt1olS\nKd/QMwilbPZFR3cB/yGW3iLytV20zXMtjiewLzwSkQfqaVfNvnJ2kX2Wsl1EbrG3LxORNBG5Ri6s\nTbBHRA7az48SkeViFXT7vMYVwEp5Vbu9klqppjDGHLCvmI7Bqvl0uTGmWET6YdW4ScNaj6L6DEJE\nOtXRztNsIMcYc7W9T2SN1/0Yq6QIIrIQWG7XEPoTcK0x5oSdVH6HVRlVKa/TBKFU3dxYlX1HYFUf\n7d+MdtuAZ0TkSeCfxpiaVXsBEKu89HljzHNiVYwdyoVSFQFYlUeVahWaIJTyIFbZ6Qqss4dfYVX2\nHY7VHVtcx24PNNTOGLNXRC4DrgJ+KyJLjDG/rvHas7DKcFdV2hRghzFmPEr5gI5BKGUTkWjgL8Cf\n7dLUkcBRuyT097G+wQPkY63KVaWudp7HjgeKjDF/B56mRglpEemFVcTuJmPMeXvzHqzFpMbbbdxi\nrQegVKvQMwjV0VVVy3VjLWrzBlBVFvp54H0R+TfgMy4svbkVqBCRLVhrMdfVzlMqVpXUSqy1GO6p\n8fy/A92Bj+zupBxjzFUiMg/+//buGAVAKIaC4Cu8P/JPbKHltnYzl1hIAtn5dhbXtnvvy0n4nTNX\nAJIREwBJIABIAgFAEggAkkAAkAQCgCQQACSBACA9FxS6SaTxXa0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1121695d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8,16,32 \n",
    "## Full Vector - throughput comparison\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[5] for entry in data] # convert to milliseconds \n",
    "    print(np.average(latency))\n",
    "    return np.average(latency)\n",
    "\n",
    "def read_gen_eight():\n",
    "    one_kb = 'results/1510751197_full_vector_encoder_1_0_8_128'\n",
    "    one_mb = 'results/1510751438_full_vector_encoder_1_0_8_131072'\n",
    "    ten_mb = 'results/1510752389_full_vector_encoder_1_0_8_1310720'\n",
    "    twenty_mb = 'results/1510822030_full_vector_encoder_1_0_8_2621440'\n",
    "    thirdy_two_mb = 'results/1510824925_full_vector_encoder_1_0_8_4194304'\n",
    "    sixty_four_mb = 'results/1510836357_full_vector_encoder_1_0_8_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/1510847597_full_vector_encoder_1_0_8_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/1510927849_full_vector_encoder_1_0_8_33554432'\n",
    "    five_hundre_and_twelve_mb = 'results/1510972685_full_vector_encoder_1_0_8_67108864'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/1511427894_full_vector_encoder_1_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/1511454980_full_vector_encoder_1_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/1511470001_full_vector_encoder_1_0_16_8388608'\n",
    "    sixty_four_mb = 'results/1511479334_full_vector_encoder_1_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/1511516065_full_vector_encoder_1_0_16_2097152'\n",
    "    twenty_mb = 'results/1511524913_full_vector_encoder_1_0_16_1310720'\n",
    "    ten_mb = 'results/1511529363_full_vector_encoder_1_0_16_655360'\n",
    "    one_mb = 'results/1511534339_full_vector_encoder_1_0_16_65536'\n",
    "    one_kb = 'results/1511534688_full_vector_encoder_1_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_thirdy_two():\n",
    "    one_kb = 'results/1511252697_full_vector_encoder_1_0_32_32'\n",
    "    one_mb = 'results/1511252611_full_vector_encoder_1_0_32_32768'\n",
    "    ten_mb = 'results/1511252293_full_vector_encoder_1_0_32_327680'\n",
    "    twenty_mb = 'results/1511221602_full_vector_encoder_1_0_32_655360'\n",
    "    thirdy_two_mb = 'results/1511217072_full_vector_encoder_1_0_32_1048576'\n",
    "    sixty_four_mb = 'results/1511210497_full_vector_encoder_1_0_32_2097152'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/1511201901_full_vector_encoder_1_0_32_4194304'\n",
    "    two_hundre_and_fifty_six_mb = 'results/1511187820_full_vector_encoder_1_0_32_8388608'\n",
    "    five_hundre_and_twelve_mb = 'results/1511156569_full_vector_encoder_1_0_32_16777216'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "eight_latencies = read_gen_eight()\n",
    "eight_data_size = np.array(eight_latencies.keys())\n",
    "eight_latency = np.array(eight_latencies.values())\n",
    "\n",
    "sixteen_latencies = read_gen_sixteen()\n",
    "sixteen_data_size = np.array(sixteen_latencies.keys())\n",
    "sixteen_latency = np.array(sixteen_latencies.values())\n",
    "\n",
    "thridy_two_latencies = read_gen_thirdy_two()\n",
    "thridy_two_data_size = np.array(thridy_two_latencies.keys())\n",
    "thridy_two_latency = np.array(thridy_two_latencies.values())\n",
    "\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "plt.plot(eight_data_size, eight_latency)\n",
    "\n",
    "plt.xticks(sixteen_data_size, my_xticks)\n",
    "plt.plot(sixteen_data_size, sixteen_latency)\n",
    "\n",
    "plt.xticks(thridy_two_data_size, my_xticks)\n",
    "plt.plot(thridy_two_data_size, thridy_two_latency)\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['G=8', 'G=16', 'G=32'])\n",
    "plt.ylabel('Bytes Per microsecond')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/full_vector_throughput_gen_comparison.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "170.601161996\n",
      "12.5145348663\n",
      "2.11998107964\n",
      "t\n",
      "2.39688843232\n",
      "2.31317858658\n",
      "1.9797070488\n",
      "2.00283028044\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAENCAYAAAAYIIIKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEyUOERGJyf8HjRVBbmm2sNEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1062125d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8 \n",
    "## On The fly - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512501259_on_the_fly_encoder_binary8_0_8_67108864'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512496496_on_the_fly_encoder_binary8_0_8_33554432'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512493964_on_the_fly_encoder_binary8_0_8_16777216'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512492545_on_the_fly_encoder_binary8_0_8_8388608'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512491693_on_the_fly_encoder_binary8_0_8_4194304'\n",
    "twenty_mb = 'results/encoder/thousand/1512491120_on_the_fly_encoder_binary8_0_8_2621440'\n",
    "ten_mb = 'results/encoder/thousand/1512490653_on_the_fly_encoder_binary8_0_8_1310720'\n",
    "one_mb = 'results/encoder/thousand/1512490269_on_the_fly_encoder_binary8_0_8_131072'\n",
    "one_kb = 'results/encoder/thousand/1512489957_on_the_fly_encoder_binary8_0_8_128'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print (latencies[2]-float(latencies[1]))/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/on_the_fly_gen_8.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "102.18009828\n",
      "13.0392211084\n",
      "2.12470880956\n",
      "t\n",
      "2.10778455771\n",
      "2.52915881181\n",
      "2.35885300174\n",
      "2.0260116244\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUXCIiEhcFBwiIhKX/w+xKxRliDyjJgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1130035d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 16 \n",
    "## On The fly - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512513739_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512508485_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512505713_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512504214_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512503337_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "twenty_mb = 'results/encoder/thousand/1512502740_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "ten_mb = 'results/encoder/thousand/1512502264_on_the_fly_encoder_binary8_0_16_655360'\n",
    "one_mb = 'results/encoder/thousand/1512501875_on_the_fly_encoder_binary8_0_16_65536'\n",
    "one_kb = 'results/encoder/thousand/1512501563_on_the_fly_encoder_binary8_0_16_64'\n",
    "\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print (latencies[2]-float(latencies[1]))/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/on_the_fly_gen_16.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "85.693500136\n",
      "12.8405712563\n",
      "2.12056889935\n",
      "t\n",
      "2.03719422218\n",
      "2.07883043154\n",
      "2.48792871386\n",
      "2.45240696245\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xl8VfWd//HXh4R9CTthCwFBkE2QCO51FxektbZV1Gqr\n0k5rp3vHTvurXaedcepIW1tr1aoVdNSqVUFUGLdaRTaBsCNKCCSsSUAgZPv8/jgnehuTcC/k5m7v\n5+ORR3LPPcsnNzf3cz7f7znfr7k7IiIi0WqT6ABERCS1KHGIiEhMlDhERCQmShwiIhITJQ4REYmJ\nEoeIiMREiUNERGKixCEiIjFR4hARkZhkJzqAeOjdu7fn5+cnOgwRkZSxdOnS3e7eJ5p10zJx5Ofn\ns2TJkkSHISKSMsxsS7TrqqlKRERiklaJw8ymmdk9FRUViQ5FRCRtpVXicPdn3X1mTk5OokMREUlb\naZU4REQk/pQ4REQkJkocIiISEyUOERGJiRKHiEiKc3f+sWk3//3C+lY5XlrdAGhm04Bpw4cPT3Qo\nIiJxt/dAFU8s3cojb2/lvd0HyOnYlutPy6dP1/ZxPa65e1wPkAgFBQWuO8dFJB25O4ve28ucRUXM\nLyylqraOgiE9mDElj0vG9adD26yj2q+ZLXX3gmjWTauKQ0QkXZUfrOKvy7YxZ9EW3t11gK4dspkx\nJY+rJ+cxMrdrq8aixCEikqTcnaVbypizqIjnVpVQVVPHhMHd+a8rxzNt/AA6tju66uJYKXGIiCSZ\nikPVPLWsmEfe3sr6Hfvp0j6bzxYMYsbkIYwe0C3R4SlxiIgkA3dn+dbyoLpYuZ3K6jrGD8rhV1eM\nY9qJA+jcPnk+rpMnEhGRDLS/spqnl29j9qIi1pXup1O7LD41cRDXTMlj7MDkHHdPiUNEJAFWFpcz\n+60inlmxnUPVtYwZ0I1ffGos0ycMpEsSVReNSe7oADM7Afg60BtY6O5/SHBIIiJH5YPDNTzzznbm\nvL2Fwm376Ng2i2kn9mfGlCGcOCgHM0t0iFFJSOIws/uBy4Cd7j42YvlUYBaQBdzr7r9y97XAl82s\nDfAQoMQhIimlcFsFc94u4m/Lt3GgqpZRuV356fQxfHLiQLp1aJvo8GKWqIrjAeB3BIkAADPLAu4C\nLgCKgcVm9oy7rzGzy4F/Af6SgFhFRGJ2sKqGZ1dsZ86iIlYUV9A+uw2XjR/AjCl5nJTXPWWqi8Yk\nJHG4+2tmlt9g8WRgk7tvBjCzR4HpwBp3fwZ4xszmAnNaM1YRkVisLdnHnEVFPL18G/sP1zCibxdu\nmzaaKyYOIqdT6lUXjUmmPo6BwNaIx8XAFDM7G7gCaA/Ma2pjM5sJzATIy8uLX5QiIg0cqqrluZXb\nmfN2EcuLymmX3YZLx/VnxpQ8Cob0SOnqojHJlDga5e6vAK9Esd49wD0QjFUV36hERGDjjv3MXlTE\nk8uK2VdZw7A+nfnhpSfw6ZMG0aNzu0SHFzfJlDi2AYMjHg8Kl0VNo+OKSLxVVtfyfGEJcxYVsfj9\nMtpmGVPH9mfG5DxOGdYz7aqLxiRT4lgMjDCzoQQJ4ypgRiw7cPdngWcLCgpujkN8IpLB3t99gNmL\ntvD40mLKD1aT36sT3794FFdOGkSvLvEdxjzZJOpy3EeAs4HeZlYM3Obu95nZLcALBJfj3u/uq2Pc\nryoOEWkxNbV1LFi7k9mLtvD6xt1ktTEuHN2Pa6YM4bTjetGmTfpXF43RfBwiIg2UVlTyyNtFPLq4\niB37DtM/pwNXT87jcycPpl+3DokOLy4ydj4OVRwicrTq6py/b9rN7EVbWLB2J3XunDWiDz+bnse5\no/qSnaWZtuup4hCRjFZ2oIrHl25lzqIi3t9zkJ6d2/GZgkFcM3kIeb06JTq8VqOKQxWHiDTD3VlW\nVMbDbxUxN5wg6eT8HnzzguOZOjaX9tmJmSApVajiEJGM8cHhGp5evo2H39rCutJggqQrThrIjCl5\njMpN/ARJiZSxFYeISGPWluzj4be28HQ4yODo/t34j0+NY/qE5JogKVXoFRORtFRZXcu8VSU8/NYW\nlhWVfzjI4LWn5DFhcGoPMphoaZU41MchIu/vPsCct4t4fMlWyg5WM7R3MAzIlZMG0b1T+g4D0prS\nKnHoznGRzNTUjXrXnhLcqKfqomWlVeIQkcxSWlHJo4uLePTtrZTuq6R/Tge+dcHxaX2jXjJIq8Sh\npiqR9FdX57zx7m4efuufb9T76fQxulGvlehyXBFJCbpRL750Oa6IpIXgRr1yZr+1hed0o17SUOIQ\nkaSzvfwQzxeW8sTSYtaW7KNL+2yuOnmwbtRLEkocIpIUtu49yPOFJcxbVco7W8sBGDtQN+olo7T6\nS6hzXCS1vL/7APMKS3h+VSmrtlUAMG5gDt+bOpKLx/ZnaO/OCY5QGqPOcRFpVe/u+oDnV5Uwd1Up\na0v2ATBhcHcuGZfLxWP7M7inOroTQZ3jIpI03J2NOz9g3qqgsli/Yz8ABUN68P8uG83UsbkM7N4x\nwVFKLJQ4RKTFuTtrS/aHfRYlvLvrAGZwcn5PfjxtNFPH9ic3RzfopSolDhFpEe7O6u37gsqisJT3\ndh+gjcEpw3pxw+lDuWhMP/p2VbJIB0ocInLU3J0VxRU8v6qEeYUlbN17iKw2xmnH9WLmWcO4cHQ/\nenVpn+gwpYUpcYhITOrqnOVby5i3qpT5haVsKz9E2yzj9OG9+dq5I7jghH706KxRaNNZWiUOXY4r\nEh+1dc6S9/fyfGEpzxeWsGPfYdplteGs43vzrQuO5/wT+pHTqW2iw5RWklaJQ8Oqi7Scmto63n5/\nL8+vKmX+6lJ27T9M++w2nD2yD5eM68+5o/rStYOSRSZKq8QhIsemuraOtzbvYd6qUl5cXcqeA1V0\nbJvFuaP6cvG4XM4Z2Vd3cIsSh0imq6qp4413d/P8qhJeXLOD8oPVdG6Xxbkn9OOSsbl8YmQfOrXT\nR4V8RO8GkQx0sKqG1zbsYn5hKQvX7WR/ZQ1d22dz/uh+XDw2l7OO70OHthp5VhqnxCGSISoOVrNw\n3Q5eWF3Kqxt2UVldR49Obbl4bC4XjcnljBG9NUy5REWJQySN7dxfyUtrdjC/sJQ3391DTZ2T260D\nnysYzEVjc5mc31Mz5knMlDhE0szWvQd5YXVwj8XSojLcYWjvztx05jCmjs1l/MAc2rSxRIcpKSzp\nE4eZfRK4FOgG3OfuLyY4JJGkUj+I4PzCIFmsCUecHd2/G988P5gpb0TfLpgpWUjLSEjiMLP7gcuA\nne4+NmL5VGAWkAXc6+6/cvengafNrAfw34ASh2S8+qE+5hcGl81u3h0MIjgprwc/uOQELhqTq3m4\nJW4SVXE8APwOeKh+gZllAXcBFwDFwGIze8bd14Sr/DB8XiQj1d+Q9+LqoIO7pKKS7DbGqcf14otn\nDOXC0f3o202DCEr8JSRxuPtrZpbfYPFkYJO7bwYws0eB6Wa2FvgV8Ly7L2vVQEUSrLK6ln+8u5v5\nhaW8tGYHZQeraZ/dhk8c34fvXjSS80ZpqA9pfcnUxzEQ2BrxuBiYAnwNOB/IMbPh7n53Yxub2Uxg\nJkBeXl6cQxWJnw8O1/DK+p3MLyzl5XU7OVBVS9f22Zx3Ql+mhvdY6IY8SaSkf/e5+2+A30Sx3j3A\nPRBMHRvvuERaUtmBKl5au4MXV5fy2sbdVNXU0btLOy6fMJCLxvTjtON60y5bl81KckimxLENGBzx\neFC4LGoaHVdSSWlFJS+uCa6EWvTeXmrrnIHdO3LtlCFMHZvLpCE9yNJls5KEkilxLAZGmNlQgoRx\nFTAjlh1odFxJdu/tPvDhPRbvbC0HYHjfLvzLJ45j6thcxgzopstmJekl6nLcR4Czgd5mVgzc5u73\nmdktwAsEl+Pe7+6rY9yvKg5JGtW1dawr2c/yrWUs21LGsqJyivYeBGD8oBy+e9FILhqTy/C+XRIc\nqUhszD39ugMKCgp8yZIliQ5DMsyu/YdZVlTGsqIylheVs7K4nMrqOgD6dG3PSXndmTK0FxeO6ceg\nHrrHQpKLmS1194Jo1k2mpqpjpopDWktVTR1rS/axvCioJJYVlVFcdgiAtlnG6AE5XD05j5PyejAx\nrzsDu3dUE5SkDVUcIlHYsa/yoySxpYxV2yo4XBNUE7ndOnDSkO5MHNyDk4Z0Z8yAHA1JLiknYysO\nkZZQVVPH6u0VLCsqZ3nY7LStPKgm2mW1YczAblx7ypAPq4kB3TsmOGKR1pVWiUNNVXI0SioOsWxL\neVhRlFG4fR9VYTUxIKcDE4f04ItnDGViXnfGDOimOSsk46mpSjJKZXUtq7fv+zBJLC8qp6SiEoB2\n2W0YPzCHiXndw2qiB7k5GvtJMoOaqkQIRpDdXlEZXgobJIk12/dRVRtUEwO7d6Qgvycn5XVnYl4P\nRvfvpruzRaKQVolDTVWya/9h5q8u5Y2Nu1lWVMbO/YcB6NC2DeMHducLZ+QHndh53TWSrMhRUlOV\npLy9B6qYX1jKcyu389bmPdQ5DOrRkUlDenBSXvA1qn9X2mqKVJEmxa2pKpxMabC7rzyqyERaSPnB\nKl5YXcpzK0v4x7t7qK1zhvXuzC3nDOfS8QM4vp9mvBOJlyMmDjN7Bbg8XHcpsNPM3nD3b8U5NpF/\nUnGwmhfXlDJ3VQl/37ibmjpnSK9OfOmsYVw2fgAn9O+qZCHSCqKpOHLcfZ+Z3QQ85O63mVlSVhzq\n40g/+yqrWbBmB3NXlvDaxl1U1zqDenTkxjOHMm38AA0KKJIA0SSObDPrD3wW+EGc4zkmGh03PXxw\nuIaFa3fw3MoSXl2/i6raOgbkdOCG0/K5bPwAxg/KUbIQSaBoEsdPCUas/bu7LzazYcDG+IYlmeZg\nVQ0L1+5k7soSXl6/k8M1deR268C1pwzh0vH9mTi4O200N4VIUjhi4nD3x4HHIx5vBj4dz6AkMxyq\nquXl9UGyWLhuB5XVdfTp2p6rJ+dx6fj+TMrroWQhkoSi6RwfSjDvd37k+u5+efzCknRVWV3LK+t3\nMXdVCQvX7uBgVS29u7TjM5MGc+n4/pyc31Oz3okkuWiaqp4G7gOeBeriG86xUed4cjpcU8vrG3bz\n3MrtLFi7kw8O19Czczs+OXEgl43rz+ShPcnWPRYiKeOINwCa2SJ3n9JK8bQI3QCYeFU1dbyxaTfP\nrtzOS6t3sP9wDd07tWXqmFwuHd+fU4f1UrIQSSItfQPgLDO7DXgROFy/0N2XHWV8kqaqa+v4x7t7\neG7Fdl5YXcq+yhq6dsjmorG5XDa+P6cP7627t0XSQDSJYxxwHXAuHzVVefhYMlxNbR1vbd7L3FXb\nmV9YStnBarq0z+bC0f24dHx/zhjRW8OQi6SZaBLHZ4Bh7l4V72Aktdz7+mb+8Mq77DlQRed2WZw/\nuh+XjuvPWcf30Qx4ImksmsRRCHQHdsY5Fkkhjy3Zys/nruXMEb25ZkoeZ4/sq2QhkiGiSRzdgXVm\ntph/7uPQ5bgZ6s139/DvT67izBG9uf+Gk9VvIZJhokkct8U9ihaiy3Hjb/OuD/jyw0vJ792Z3804\nSUlDJANpPg6JWvnBKj71+39Qcaiap79yOnm9OiU6JBFpIbFcjqvTRYlKVU0dX/rLUraVHeKe6yYp\naYhksLSaOlbiw935wVOrWPTeXu783AQK8nsmOiQRSSBVHHJEd7+6mceXFvOv543gkxMHJjocEUmw\naAY5PB34MTAkXN8Ad/dh8Q1NksH8whL+c/46pp04gG+ePyLR4YhIEoimqeo+4JsE08bWxjccSSYr\ni8v5xv++w8S87tx+5XhNniQiQHSJo8Ldn497JJJUtpcf4sYHl9Crc3vuua5AN/eJyIeiSRwvm9nt\nwJMkYJDDcMbBHxDMfX5laxwz0x04XMONDy6hsqqW2V+ZQp+u7RMdkogkkWgSR/2Q6pHX9x7TIIdm\ndj9wGbDT3cdGLJ8KzAKygHvd/VfhjIM3mtkTR3s8iV5tnfOvjyxnw4793H/DyRzfr2uiQxKRJBPN\n1LHnxOG4DwC/Ax6qX2BmWcBdwAVAMbDYzJ5x9zVxOL404T/mrWXhup38bPoYPnF8n0SHIyJJqMnE\nYWbXuvvDZvatxp539zuO9qDu/pqZ5TdYPBnYFFYYmNmjwHQgqsRhZjOBmQB5eXlHG1pGm71oC/f9\n/T1uOC2f607NT3Q4IpKkmruPo3P4vWsTXy1tILA14nExMNDMepnZ3cBEM/t+Uxu7+z3uXuDuBX36\n6Ew5Vq9v3MWP/raac0b24f9dNjrR4YhIEmuy4nD3P4bff9J64TQaxx7gy9Gsq0EOj86mnfv5yuxl\njOjbhd/OOImsNrrsVkSalkx3jm8DBkc8HhQui5q7P+vuM3Nyclo0sHS254PDfOGBxbTPzuK+G06m\nS3uNQiMizUumxLEYGGFmQ82sHXAV8EwsOzCzaWZ2T0VFRVwCTDeHa2r50l+WsnPfYe69voCB3Tsm\nOiQRSQFHTBzh1U4tysweAd4ERppZsZnd6O41wC3AC8Ba4DF3Xx3LflVxRM/dufWvq1iypYw7PjuB\nCYO7JzokEUkR0bRLbDSzvwJ/bqlLY9396iaWzwPmHe1+1ccRvd/+3yaeWr6N71x4PJeO75/ocEQk\nhUTTVHUisAG418zeMrOZZtYtznEdFVUc0Xl2xXbueGkDV5w0kK+eoyQrIrE5YuJw9/3u/id3Pw34\nN4KpZEvM7EEz06dOillWVMa3H1/B5Pye/PKKcRq4UERiFlUfh5ldbmZPAXcCvwaGAc9yDM1K8aDO\n8eZt3XuQmQ8toX9OB+6+bhLtszVwoYjELpqmqo0Ed3Df7u4T3f0Od9/h7k8A8+MbXmzUVNW0fZXV\n3PjgYqpq6rjv+pPp2bldokMSkRQVTef4eHf/oLEn3P1fWzgeiYOa2jpumbOczbsO8OAXJzO8b5dE\nhyQiKSyaiuMuM/vwWk0z6xGObpt01FTVuJ8+t4bXNuziZ58cy+nDeyc6HBFJcdEkjvHuXl7/wN3L\ngInxC+noqanq4x544z0eenMLN585lKsna/BHETl20SSONmbWo/6BmfUkuiYuSbCX1+3kp8+t4fwT\n+nHrxSckOhwRSRPRJIBfA2+a2eOAAVcCv4hrVHLM1pXu42uPLGdUbjdmXTVBAxeKSIuJZiKnh8xs\nKVA/odMVyTq5ku4cD+zcX8mNDyyhc/ss7ruhgM4auFBEWlC0gxyuI5hz/BngAzNLysZy9XFAZXUt\nMx9ayt4DVdz7+ZPpn6OBC0WkZR3xVNTMvkZwt/gOoJagucqB8fENTWJVV+d8+/EVrCgu5+5rJzFu\nUOYmUBGJn2jaML4OjAwnVJIk9j8LNjB3ZQnfv3gUF43JTXQ4IpKmommq2groxogk9+SyYn77f5v4\nXMFgZp41LNHhiEgai6bi2Ay8YmZzgcP1C939jrhFdZQytXN88ft7ufWvqzh1WC9+9smxGrhQROIq\nmoqjCHgJaAd0jfhKOpnYOb5lzwFmPrSEQT06cve1k2iXnUyTOopIOormctyfAJhZJ3c/GP+QJFoV\nh6r54gOLceD+G04mp1PbRIckIhkgmmHVTzWzNQSX5GJmJ5rZ7+MemTSruraOr8xeStHeg/zx2knk\n9+6c6JBEJENE065xJ3ARsAfA3VcAZ8UzKGmeu/OjvxXyxqY9/PKK8UwZ1ivRIYlIBomqQdzdtzZY\nVBuHWCRK977+Ho+8vZWvnH0cV04alOhwRCTDRHNV1VYzOw1wM2tLcF/H2viGJU15cXUp//H8Wi4Z\nl8t3LhyZ6HBEJANFU3F8GfgqMBDYBkwAvhLPoI5Wus/HUbitgq8/+g7jB+bw689MoI0GLhSRBIgm\ncYx092vcvZ+793X3a4GkHKM7nS/HLa2o5KYHl9CjU1v+9PkCOrbTfOEikhjRJI7fRrlM4uRgVQ03\nPbSY/ZXV3Hv9yfTt1iHRIYlIBmuyj8PMTgVOA/qY2bcinuoG6HS3Ff37k6tYs30ff/p8AaMHdEt0\nOCKS4ZrrHG8HdAnXibxTfB/BZE7SCgq3VfD0O9u55ZzhnHdCv0SHIyLSdOJw91eBV83sAXff0oox\nSYTfLNxI1w7Z3KyBC0UkSURzOe5BM7sdGAN82Lju7ufGLSoBgmrjxTU7+Mb5I8jpqOFERCQ5RNM5\nPptguJGhwE+A94HFcYxJQvXVxhdOH5roUEREPhRN4ujl7vcB1e7+qrt/EWi1asPMOpvZg2b2JzO7\nprWOm2j11cZNZwxTtSEiSSWaxFEdfi8xs0vNbCLQ81gOamb3m9lOMytssHyqma03s01mdmu4+Arg\nCXe/Gbj8WI6bSn6zcCPdOmRzw+n5iQ5FROSfRJM4fm5mOcC3ge8A9wLfOMbjPgBMjVxgZlnAXcDF\nwGjgajMbDQwimIUQMmSMrPpq40ZVGyKShI6YONz9OXevcPdCdz/H3ScBxx3LQd39NWBvg8WTgU3u\nvtndq4BHgelAMUHyiCredDBL1YaIJLGj/SD+1pFXidlAPqosIEgYA4EngU+b2R+AZ5va2MxmmtkS\nM1uya9euOITXOgq3VfCSqg0RSWLRXI7bmFYbXc/dDwBfiGK9e8ysBJjWrl27SfGPLD5UbYhIsjva\nisNbNIrANmBwxONB4bKopfogh6o2RCQVNDdW1X4aTxAGdIxDLIuBEWY2lCBhXAXMiGUHZjYNmDZ8\n+PA4hBd/9dXGF87IT3QoIiJNarLicPeu7t6tka+u7n60TVwAmNkjwJvASDMrNrMb3b0GuAV4gWCi\nqMfcfXUs+03liqO+2rjpzGF066BqQ0SS1zElgKPl7lc3sXweMK+Vw0kKdy5Q34aIpIa0urw1VWcA\nLNxWwYK1qjZEJDWkVeJI1aYqVRsikkrSKnGkYsWhakNEUk1aJY5UrDhUbYhIqkmrxJFqVG2ISCpS\n4kigOxdsJKdjW1UbIpJS0ipxpFIfx6risNo4Y6iqDRFJKWmVOFKpj2PWwg3kdGzL9ao2RCTFpFXi\nSBVBtbFT1YaIpKS0Shyp0lSlakNEUllaJY5UaKpStSEiqS6tEkcqqK82dCWViKQqJY5WVF9t3Hzm\nULqq2hCRFKXE0YruXBD2bZyWn+hQRESOWloljmTuHF9ZXM7Cdao2RCT1pVXiSObO8VnhXeKqNkQk\n1aVV4khWqjZEJJ0ocbQCVRsikk6UOOJM1YaIpBsljjibtWAj3Tup2hCR9KHEEUcrttZXG8NUbYhI\n2kirxJFsl+POWhhUG58/dUiiQxERaTFplTiS6XLcFVvL+T9VGyKShtIqcSQTVRsikq6UOOJA1YaI\npDMljjhQtSEi6UyJo4Wp2hCRdKfE0cLuXLBB922ISFpT4mhB72wt5+X1u7j5zGF0aZ+d6HBEROIi\n6ROHmQ0zs/vM7IlEx3Iks1RtiEgGiGviMLP7zWynmRU2WD7VzNab2SYzu7W5fbj7Zne/MZ5xtgRV\nGyKSKeL9CfcA8DvgofoFZpYF3AVcABQDi83sGSAL+GWD7b/o7jvjHGOLULUhIpkironD3V8zs/wG\niycDm9x9M4CZPQpMd/dfApfFM554qa82vnvRSFUbIpL2EtHHMRDYGvG4OFzWKDPrZWZ3AxPN7PvN\nrDfTzJaY2ZJdu3a1XLRRmLVgAz1UbYhIhkj602N33wN8OYr17gHuASgoKPB4x1VveVEZL6/fxfem\nqtoQkcyQiIpjGzA44vGgcNkxS8TouLMWbqRHp7Z8/tT8VjumiEgiJSJxLAZGmNlQM2sHXAU80xI7\nbu3RcZcXlfHK+l3cfJaupBKRzBHvy3EfAd4ERppZsZnd6O41wC3AC8Ba4DF3X91Cx2vVikPVhohk\nonhfVXV1E8vnAfPicLxngWcLCgpubul9N1RfbahvQ0QyTdLfOR6L1qw4VG2ISKZKq8TRWn0c6tsQ\nkUyWVomjtdy5QNWGiGSutEocrdFUtayojFc37GLmWcep2hCRjJRWiaM1mqpmfVhtaHY/EclMaZU4\n4i2y2uisakNEMlRaJY54N1Wp2hARSbPEEc+mKlUbIiKBtEoc8aRqQ0QkoMQRhaVbVG2IiNRLq8QR\nrz6OWQs30rNzO1UbIiKkWeKIRx/H0i1lvLZhFzPPGqZqQ0SENEsc8VBfbVx3iqoNERFQ4miWqg0R\nkY9T4miGqg0RkY9Lq8TRkp3jqjZERBqXVomjJTvH71ywQdWGiEgj0ipxtJSlW/by+sbdfEnVhojI\nxyhxNOLOBWHfhu7bEBH5GCWOBiKrjU7tVG2IiDSkxNGAqg0RkealVeI41quqVG2IiBxZWiWOY72q\nStWGiMiR6bQ6VFNbx6jcrpw3qq+qDRGRZugTMpSd1YYfXDo60WGIiCS9tGqqEhGR+FPiEBGRmChx\niIhITJQ4REQkJkocIiISEyUOERGJiRKHiIjERIlDRERiYu6e6BhanJntArYc5ea9gd0tGE5LUVyx\nUVyxUVyxSce4hrh7n2hWTMvEcSzMbIm7FyQ6joYUV2wUV2wUV2wyPS41VYmISEyUOEREJCZKHB93\nT6IDaILiio3iio3iik1Gx6U+DhERiYkqDhERiUnGJQ4zu9/MdppZYcSyV8zsY1cimNn7ZrbKzN4J\nv09vpXgeMLODZtY1YtmdZuZm1jt8XBvGtcLMlpnZaXGKpaeZvWRmG8PvPcLlZ4fx3BSx7oRw2Xci\nfo/3wjjXmdltRxnXYDN72czWmNlqM/t6MsRmZh3M7O3wb7DazH4SLr893OdKM3vKzLon4DXrbmZP\nhPtYa2anRjz37QbvpbjF1cR7qqnXp62ZPRj+r601s+9HbPO+mb3eYN/v1O83/B0qwmUrzWyBmfVt\nJq6m3lM/NrNt4X7eMbNLIrYZb2ZvhuuvMrMOcYot8nNnSbjsM+Fx6yzis8rMLjCzpeH6S83s3Hi9\nZpEyLnEADwBTY1j/HHefAFwJ/KYV49kETAcwszbAucC2iOcPufsEdz8R+D7wyzjFciuw0N1HAAvD\nx/UKgc9GPL4aWNFg+++Gr98E4HozG3oUcdUA33b30cApwFfNbHQSxHYYODf8G0wApprZKcBLwFh3\nHw9sIPilD3K/AAAIHklEQVT7tGZcALOA+e4+CjgRWAvBByZwIVDUYP14xfUAH39PNfX6fAZo7+7j\ngEnAl8wsP2K7rmH8mNkJjRzr9fB/YjywGPhqM3E19Z4C+J9wPxPcfV54vGzgYeDL7j4GOBuojlNs\nEH7uRFxaWwhcAbzWYL3dwLTwNbse+EuD51s6LiADE4e7vwbsbew5M2sTnln9vJGnuwFlrRjPo8Dn\nwp/PBt4geLM3pkViayKW6cCD4c8PAp+MeG4L0MHM+pmZEXxAPN/E7juE3w8cRVwl7r4s/Hk/wYfg\nwETH5oEPwodtwy939xfdvf5v9RYwqDXjMrMc4CzgvjDOKncvD5/+H+B7QMPOzbjE1dh7qpnXx4HO\n4Yd0R6AK2Bex6WN89D9xNfBIY8cM4+9KM/8TzbynmnIhsNLdV4Tb7HH32njE1kS8a919fSPLl7v7\n9vDhaqCjmbWPd1wZlziakQ3MBja6+w8jlr8clnavAj9sdMv42AD0saD55WqCRBKpY32zAXAv8LM4\nxdHP3UvCn0uBfg2ef4LgTPE0YBnBWXik283sHaAYeNTddx5LMOEZ6ERgUTLEZmZZ4T52Ai+5+6IG\nq3yRj38AxzuuocAu4M9mttzM7jWzzhY0tW6r//BrRKv+LUORr88TBMmohKAi+m93j0w6fyU46waY\nBjzbYF9nhvEVAecD90cTQIP3FMDXwqab+8P/P4DjATezFyxoGv5eg920ZGwOLAibnmZG8zuEPg0s\nc/fIv1tcXjMljo/8ESh09180WH6Ou48FxgG/M7MurRjTk8BVwBTg9QbP1TdVjSI4O3woPGuIGw8u\nwWt4pvoYwYdNU2cz9c0bucB5dgx9MeFr/1fgG+4eeSaasNjcvTbcxyBgspmNjYj3BwRV4uxWjisb\nOAn4g7tPJPgw/jHw78CPmtmu1f6W0OjrMxmoBQYQJL9vm9mwiE32AGVmdhVBhXCwwS7rm10GA38G\n/iuKGBq+p/4ADCNojisBfh2umg2cAVwTfv+UmZ0Xp9jOCF/niwma0M6K4vcYA/wn8KUGT7X4awZK\nHJH+AZxT3+HVkLu/C+wARjf2fJz8L0El8ZK71zW1kru/STBGTVTjzMRoh5n1Bwi//9NZpruXErT1\nXkDQz9BUjB8ArxD808XMzNoS/IPPdvcnkym2cB/lwMuE7flmdgNwGXCNN7jmvRXiKgaKI6qfJwgS\nyVBghZm9T5DolplZbivG9aEmXp8ZBP0y1WE18wbQ8KKV/wXuookmlwjPEDTXNRfDx95T7r4jPBmo\nA/5EkMwgeE1fc/fd7n4QmEfwmrZ4bO6+Lfy+E3gqIoamfo9B4XqfDz+nGmqx16yeEsdH7iN4MzwW\ntrH+k/Bqg6Ec/eCJMXP3LcAPgN83t56ZjQKyCM4uWtozBJ1uhN//1sg6PwL+rUGb7z8JX9MpQGNv\n7GaFldR9wFp3vyNZYjOzPvbRFUEdCT5w15nZVIJ+hMvDD5nGxC2uMAFsNbOR4aLzCJow+rp7vrvn\nE3wQnhSu2ypxRWzf1OtTRHARCGbWmaDTel2DzZ8iOCt+4QiHOaO5+Jp6T9WfiIQ+RdApTXi8cWbW\nKfz9PwGsaenYwibFrvU/E/StFDa2brhOd2AucKu7v9HEai3ymkX62AdkujOzRwg6m3ubWTHw4WWF\n7n5H2LH4FzO7Jlz8spnVEnR83uruO1ornjCmPzaxacewbRLAgOub+2c/hlh+RZBMbyRImp9tuJ27\n/6OZ3d5uZj8E2hGcxT7ZzLpNOR24DlgV8Tv/exLE1h940MyyCE7CHnP358xsE9AeeClsPXzL3b/c\ninEBfA2YbWbtgM3AF6LZqKXjauI99X0af33uIuiXWU3wnv6zu69sEN9+giYZGmmZrW+vN6ACuKnh\nChGaek9dbWYTCJo93yds+nH3MjO7g+DKIwfmufvcOMTWD3gq3D4bmOPu883sU8BvCVoV5prZO+5+\nEXALMBz4kZnVN0NeGNn/1IKv2Yd057iIiMRETVUiIhITJQ4REYmJEoeIiMREiUNERGKixCEiIjFR\n4hBphH00+vBqC0a//bYFg002t02+mc1ooePfax8NuieSVHQ5rkgjzOwDd+8S/twXmAO84e5NDidu\nZmcD33H3y1onSpHEUMUhcgThzVQzgVsskG9mr4eD3UXOhfIrwhuqzOybzaz3ofBO4blhVVNoZp8L\nl79iZgVmdrl9NDfEejN7L3x+kpm9asFAeC80uONZJK4y7s5xkaPh7pvDO8T7EoyJdYG7V5rZCIIx\ngAoI5gP5sOIws05NrBdpKrDd3S8Nt8lpcNxnCIZWwcweA14Nx1j6LTDd3XeFyeYXBCPNisSdEodI\n7NoSjJQ8gWA01+OPYb1VwK/N7D+B59y94SjIAFgwjPchd7/LghF4x/LRkB1ZBCO5irQKJQ6RKFgw\nvHctQbVxG8FIyScSNPdWNrHZN4+0nrtvMLOTgEuAn5vZQnf/aYNjn08w3Hn9yKUGrHb3UxFJAPVx\niByBmfUB7gZ+Fw4BngOUhENvX0dwxg+wn2AWtXpNrRe57wHAQXd/GLidBkN1m9kQgsH/PuPuh8LF\n6wkm+To1XKetBfMxiLQKVRwijasffbgtwWRDfwHqh9/+PfBXM/s8MJ+PplBdCdSa2QqCubabWi/S\nOIJRZ+sI5sL4lwbP3wD0Ap4Om6W2u/slZnYl8JuwTyQbuJNg6lCRuNPluCIiEhM1VYmISEyUOERE\nJCZKHCIiEhMlDhERiYkSh4iIxESJQ0REYqLEISIiMVHiEBGRmPx/SAaJ8fOnXRgAAAAASUVORK5C\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x111c50090>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 32 \n",
    "## On The fly - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512527656_on_the_fly_encoder_binary8_0_32_16777216'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512521487_on_the_fly_encoder_binary8_0_32_8388608'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512518431_on_the_fly_encoder_binary8_0_32_4194304'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512516828_on_the_fly_encoder_binary8_0_32_2097152'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512515884_on_the_fly_encoder_binary8_0_32_1048576'\n",
    "twenty_mb = 'results/encoder/thousand/1512515252_on_the_fly_encoder_binary8_0_32_655360'\n",
    "ten_mb = 'results/encoder/thousand/1512514755_on_the_fly_encoder_binary8_0_32_327680'\n",
    "one_mb = 'results/encoder/thousand/1512514356_on_the_fly_encoder_binary8_0_32_32768'\n",
    "one_kb = 'results/encoder/thousand/1512514043_on_the_fly_encoder_binary8_0_32_32'\n",
    "\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print (latencies[2]-float(latencies[1]))/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/on_the_fly_gen_32.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.78195488722\n",
      "1.66622794629\n",
      "1.73091732615\n",
      "1.73477741231\n",
      "1.67037904274\n",
      "1.46890406092\n",
      "1.60605483334\n",
      "1.91364033729\n",
      "1.93578937073\n",
      "t\n",
      "2.2585995086\n",
      "1.89417803031\n",
      "1.86737610287\n",
      "1.86373759515\n",
      "1.86287821786\n",
      "1.80048986893\n",
      "1.47990435149\n",
      "1.5608842633\n",
      "1.88938868307\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11310f110>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8,16,32 \n",
    "## On The Fly - latency comparison\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512501259_on_the_fly_encoder_binary8_0_8_67108864'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512496496_on_the_fly_encoder_binary8_0_8_33554432'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512493964_on_the_fly_encoder_binary8_0_8_16777216'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512492545_on_the_fly_encoder_binary8_0_8_8388608'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512491693_on_the_fly_encoder_binary8_0_8_4194304'\n",
    "    twenty_mb = 'results/encoder/thousand/1512491120_on_the_fly_encoder_binary8_0_8_2621440'\n",
    "    ten_mb = 'results/encoder/thousand/1512490653_on_the_fly_encoder_binary8_0_8_1310720'\n",
    "    one_mb = 'results/encoder/thousand/1512490269_on_the_fly_encoder_binary8_0_8_131072'\n",
    "    one_kb = 'results/encoder/thousand/1512489957_on_the_fly_encoder_binary8_0_8_128'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512513739_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512508485_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512505713_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512504214_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512503337_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512502740_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512502264_on_the_fly_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512501875_on_the_fly_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512501563_on_the_fly_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_thirdy_two():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512527656_on_the_fly_encoder_binary8_0_32_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512521487_on_the_fly_encoder_binary8_0_32_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512518431_on_the_fly_encoder_binary8_0_32_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512516828_on_the_fly_encoder_binary8_0_32_2097152'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512515884_on_the_fly_encoder_binary8_0_32_1048576'\n",
    "    twenty_mb = 'results/encoder/thousand/1512515252_on_the_fly_encoder_binary8_0_32_655360'\n",
    "    ten_mb = 'results/encoder/thousand/1512514755_on_the_fly_encoder_binary8_0_32_327680'\n",
    "    one_mb = 'results/encoder/thousand/1512514356_on_the_fly_encoder_binary8_0_32_32768'\n",
    "    one_kb = 'results/encoder/thousand/1512514043_on_the_fly_encoder_binary8_0_32_32'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def calc_diff(x,y):\n",
    "    print (y/float(x))\n",
    "    \n",
    "def diff_gen(g, gg):\n",
    "    for x, y in zip(g,gg):\n",
    "        calc_diff(g[x],gg[y])\n",
    "\n",
    "eight_latencies = read_gen_eight()\n",
    "eight_data_size = np.array(eight_latencies.keys())\n",
    "eight_latency = np.array(eight_latencies.values())\n",
    "\n",
    "sixteen_latencies = read_gen_sixteen()\n",
    "diff_gen(eight_latencies,sixteen_latencies)\n",
    "sixteen_data_size = np.array(sixteen_latencies.keys())\n",
    "sixteen_latency = np.array(sixteen_latencies.values())\n",
    "\n",
    "print('t')\n",
    "thridy_two_latencies = read_gen_thirdy_two()\n",
    "diff_gen(sixteen_latencies,thridy_two_latencies)\n",
    "thridy_two_data_size = np.array(thridy_two_latencies.keys())\n",
    "thridy_two_latency = np.array(thridy_two_latencies.values())\n",
    "\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(eight_data_size, my_xticks)\n",
    "plt.plot(eight_data_size, eight_latency)\n",
    "\n",
    "plt.xticks(sixteen_data_size, my_xticks)\n",
    "plt.plot(sixteen_data_size, sixteen_latency)\n",
    "\n",
    "plt.xticks(thridy_two_data_size, my_xticks)\n",
    "plt.plot(thridy_two_data_size, thridy_two_latency)\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['G=8', 'G=16', 'G=32'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/on_the_fly_gen_comparison.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSxzGGGNC8v8Bjy9ZMqPll2QAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11261e750>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def fv_read_gen_eight():\n",
    "    one_kb = 'results/encoder/thousand/1512452623_full_vector_encoder_binary8_0_8_128'\n",
    "    one_mb = 'results/encoder/thousand/1512452934_full_vector_encoder_binary8_0_8_131072'\n",
    "    ten_mb = 'results/encoder/thousand/1512453317_full_vector_encoder_binary8_0_8_1310720'\n",
    "    twenty_mb = 'results/encoder/thousand/1512453778_full_vector_encoder_binary8_0_8_2621440'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512454344_full_vector_encoder_binary8_0_8_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512455182_full_vector_encoder_binary8_0_8_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512456573_full_vector_encoder_binary8_0_8_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512459046_full_vector_encoder_binary8_0_8_33554432'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512463692_full_vector_encoder_binary8_0_8_67108864'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def fv_read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512475943_full_vector_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512470804_full_vector_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512468085_full_vector_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512466613_full_vector_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512465749_full_vector_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512465168_full_vector_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512464696_full_vector_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512464308_full_vector_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512463997_full_vector_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def fv_read_gen_thirdy_two():\n",
    "    one_kb = 'results/encoder/thousand/1512476247_full_vector_encoder_binary8_0_32_32'\n",
    "    one_mb = 'results/encoder/thousand/1512476560_full_vector_encoder_binary8_0_32_32768'\n",
    "    ten_mb = 'results/encoder/thousand/1512476958_full_vector_encoder_binary8_0_32_327680'\n",
    "    twenty_mb = 'results/encoder/thousand/1512477451_full_vector_encoder_binary8_0_32_655360'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512478066_full_vector_encoder_binary8_0_32_1048576'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512478997_full_vector_encoder_binary8_0_32_2097152'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512480568_full_vector_encoder_binary8_0_32_4194304'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512483577_full_vector_encoder_binary8_0_32_8388608'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512489653_full_vector_encoder_binary8_0_32_16777216'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512501259_on_the_fly_encoder_binary8_0_8_67108864'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512496496_on_the_fly_encoder_binary8_0_8_33554432'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512493964_on_the_fly_encoder_binary8_0_8_16777216'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512492545_on_the_fly_encoder_binary8_0_8_8388608'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512491693_on_the_fly_encoder_binary8_0_8_4194304'\n",
    "    twenty_mb = 'results/encoder/thousand/1512491120_on_the_fly_encoder_binary8_0_8_2621440'\n",
    "    ten_mb = 'results/encoder/thousand/1512490653_on_the_fly_encoder_binary8_0_8_1310720'\n",
    "    one_mb = 'results/encoder/thousand/1512490269_on_the_fly_encoder_binary8_0_8_131072'\n",
    "    one_kb = 'results/encoder/thousand/1512489957_on_the_fly_encoder_binary8_0_8_128'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512513739_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512508485_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512505713_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512504214_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512503337_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512502740_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512502264_on_the_fly_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512501875_on_the_fly_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512501563_on_the_fly_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_thirdy_two():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512527656_on_the_fly_encoder_binary8_0_32_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512521487_on_the_fly_encoder_binary8_0_32_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512518431_on_the_fly_encoder_binary8_0_32_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512516828_on_the_fly_encoder_binary8_0_32_2097152'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512515884_on_the_fly_encoder_binary8_0_32_1048576'\n",
    "    twenty_mb = 'results/encoder/thousand/1512515252_on_the_fly_encoder_binary8_0_32_655360'\n",
    "    ten_mb = 'results/encoder/thousand/1512514755_on_the_fly_encoder_binary8_0_32_327680'\n",
    "    one_mb = 'results/encoder/thousand/1512514356_on_the_fly_encoder_binary8_0_32_32768'\n",
    "    one_kb = 'results/encoder/thousand/1512514043_on_the_fly_encoder_binary8_0_32_32'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "\n",
    "fv_eight_latencies = fv_read_gen_eight()\n",
    "fv_eight_data_size = np.array(fv_eight_latencies.keys())\n",
    "fv_eight_latency = np.array(fv_eight_latencies.values())\n",
    "\n",
    "fv_sixteen_latencies = fv_read_gen_sixteen()\n",
    "fv_sixteen_data_size = np.array(fv_sixteen_latencies.keys())\n",
    "fv_sixteen_latency = np.array(fv_sixteen_latencies.values())\n",
    "\n",
    "fv_thridy_two_latencies = fv_read_gen_thirdy_two()\n",
    "fv_thridy_two_data_size = np.array(fv_thridy_two_latencies.keys())\n",
    "fv_thridy_two_latency = np.array(fv_thridy_two_latencies.values())\n",
    "\n",
    "of_eight_latencies = of_read_gen_eight()\n",
    "of_eight_data_size = np.array(of_eight_latencies.keys())\n",
    "of_eight_latency = np.array(of_eight_latencies.values())\n",
    "\n",
    "of_sixteen_latencies = of_read_gen_sixteen()\n",
    "of_sixteen_data_size = np.array(of_sixteen_latencies.keys())\n",
    "of_sixteen_latency = np.array(of_sixteen_latencies.values())\n",
    "\n",
    "of_thridy_two_latencies = of_read_gen_thirdy_two()\n",
    "of_thridy_two_data_size = np.array(of_thridy_two_latencies.keys())\n",
    "of_thridy_two_latency = np.array(of_thridy_two_latencies.values())\n",
    "\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "\n",
    "plt.xticks(fv_eight_data_size, my_xticks)\n",
    "plt.plot(fv_eight_data_size, fv_eight_latency, marker='v', linestyle='')\n",
    "\n",
    "plt.xticks(fv_sixteen_data_size, my_xticks)\n",
    "plt.plot(fv_sixteen_data_size, fv_sixteen_latency, marker='v', linestyle='')\n",
    "\n",
    "plt.xticks(fv_thridy_two_data_size, my_xticks)\n",
    "plt.plot(fv_thridy_two_data_size, fv_thridy_two_latency, marker='v', linestyle='')\n",
    "\n",
    "plt.xticks(of_eight_data_size, my_xticks)\n",
    "plt.plot(of_eight_data_size, of_eight_latency, marker='p', linestyle='')\n",
    "\n",
    "plt.xticks(of_sixteen_data_size, my_xticks)\n",
    "plt.plot(of_sixteen_data_size, of_sixteen_latency, marker='p', linestyle='')\n",
    "\n",
    "plt.xticks(of_thridy_two_data_size, my_xticks)\n",
    "plt.plot(of_thridy_two_data_size, of_thridy_two_latency, marker='p', linestyle='')\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['FV G=8', 'FV G=16', 'FV G=32', 'OF G=8', 'OF G=16', 'OF G=32'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/fv_ov_comparison.eps')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAD8CAYAAAB+UHOxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkSIAACBSBAAARIoAAIBIEQAAEKn/B51HDlzd5udSAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1182af610>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    data = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return data\n",
    "\n",
    "def fv_read_gen_eight():\n",
    "    one_kb = 'results/encoder/thousand/1512452623_full_vector_encoder_binary8_0_8_128'\n",
    "    one_mb = 'results/encoder/thousand/1512452934_full_vector_encoder_binary8_0_8_131072'\n",
    "    ten_mb = 'results/encoder/thousand/1512453317_full_vector_encoder_binary8_0_8_1310720'\n",
    "    twenty_mb = 'results/encoder/thousand/1512453778_full_vector_encoder_binary8_0_8_2621440'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512454344_full_vector_encoder_binary8_0_8_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512455182_full_vector_encoder_binary8_0_8_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512456573_full_vector_encoder_binary8_0_8_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512459046_full_vector_encoder_binary8_0_8_33554432'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512463692_full_vector_encoder_binary8_0_8_67108864'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def fv_read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512475943_full_vector_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512470804_full_vector_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512468085_full_vector_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512466613_full_vector_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512465749_full_vector_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512465168_full_vector_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512464696_full_vector_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512464308_full_vector_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512463997_full_vector_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def fv_read_gen_thirdy_two():\n",
    "    one_kb = 'results/encoder/thousand/1512476247_full_vector_encoder_binary8_0_32_32'\n",
    "    one_mb = 'results/encoder/thousand/1512476560_full_vector_encoder_binary8_0_32_32768'\n",
    "    ten_mb = 'results/encoder/thousand/1512476958_full_vector_encoder_binary8_0_32_327680'\n",
    "    twenty_mb = 'results/encoder/thousand/1512477451_full_vector_encoder_binary8_0_32_655360'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512478066_full_vector_encoder_binary8_0_32_1048576'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512478997_full_vector_encoder_binary8_0_32_2097152'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512480568_full_vector_encoder_binary8_0_32_4194304'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512483577_full_vector_encoder_binary8_0_32_8388608'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512489653_full_vector_encoder_binary8_0_32_16777216'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512501259_on_the_fly_encoder_binary8_0_8_67108864'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512496496_on_the_fly_encoder_binary8_0_8_33554432'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512493964_on_the_fly_encoder_binary8_0_8_16777216'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512492545_on_the_fly_encoder_binary8_0_8_8388608'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512491693_on_the_fly_encoder_binary8_0_8_4194304'\n",
    "    twenty_mb = 'results/encoder/thousand/1512491120_on_the_fly_encoder_binary8_0_8_2621440'\n",
    "    ten_mb = 'results/encoder/thousand/1512490653_on_the_fly_encoder_binary8_0_8_1310720'\n",
    "    one_mb = 'results/encoder/thousand/1512490269_on_the_fly_encoder_binary8_0_8_131072'\n",
    "    one_kb = 'results/encoder/thousand/1512489957_on_the_fly_encoder_binary8_0_8_128'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512513739_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512508485_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512505713_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512504214_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512503337_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512502740_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512502264_on_the_fly_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512501875_on_the_fly_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512501563_on_the_fly_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_thirdy_two():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512527656_on_the_fly_encoder_binary8_0_32_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512521487_on_the_fly_encoder_binary8_0_32_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512518431_on_the_fly_encoder_binary8_0_32_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512516828_on_the_fly_encoder_binary8_0_32_2097152'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512515884_on_the_fly_encoder_binary8_0_32_1048576'\n",
    "    twenty_mb = 'results/encoder/thousand/1512515252_on_the_fly_encoder_binary8_0_32_655360'\n",
    "    ten_mb = 'results/encoder/thousand/1512514755_on_the_fly_encoder_binary8_0_32_327680'\n",
    "    one_mb = 'results/encoder/thousand/1512514356_on_the_fly_encoder_binary8_0_32_32768'\n",
    "    one_kb = 'results/encoder/thousand/1512514043_on_the_fly_encoder_binary8_0_32_32'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "\n",
    "fv_eight_latencies = fv_read_gen_eight()\n",
    "fv_eight_data_size = np.array(fv_eight_latencies.keys())\n",
    "fv_eight_latency = np.array(fv_eight_latencies.values())\n",
    "\n",
    "fv_sixteen_latencies = fv_read_gen_sixteen()\n",
    "fv_sixteen_data_size = np.array(fv_sixteen_latencies.keys())\n",
    "fv_sixteen_latency = np.array(fv_sixteen_latencies.values())\n",
    "\n",
    "fv_thridy_two_latencies = fv_read_gen_thirdy_two()\n",
    "fv_thridy_two_data_size = np.array(fv_thridy_two_latencies.keys())\n",
    "fv_thridy_two_latency = np.array(fv_thridy_two_latencies.values())\n",
    "\n",
    "of_eight_latencies = of_read_gen_eight()\n",
    "of_eight_data_size = np.array(of_eight_latencies.keys())\n",
    "of_eight_latency = np.array(of_eight_latencies.values())\n",
    "\n",
    "of_sixteen_latencies = of_read_gen_sixteen()\n",
    "of_sixteen_data_size = np.array(of_sixteen_latencies.keys())\n",
    "of_sixteen_latency = np.array(of_sixteen_latencies.values())\n",
    "\n",
    "of_thridy_two_latencies = of_read_gen_thirdy_two()\n",
    "of_thridy_two_data_size = np.array(of_thridy_two_latencies.keys())\n",
    "of_thridy_two_latency = np.array(of_thridy_two_latencies.values())\n",
    "\n",
    "data_to_plot= [fv_eight_latency[0], of_eight_latency[0], fv_sixteen_latency[0], \n",
    "               of_sixteen_latency[0], fv_thridy_two_latency[0], of_thridy_two_latency[0]]\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "\n",
    "\n",
    "# Create a figure instance\n",
    "fig = plt.figure(1, figsize=(9, 6))\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "\n",
    "# Create an axes instance\n",
    "ax = fig.add_subplot(111)\n",
    "\n",
    "# Create the boxplot\n",
    "bp = ax.boxplot(data_to_plot)\n",
    "\n",
    "# Save the figure\n",
    "fig.savefig('results/graphs/fv_ov_comparison_full_data.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144.74408218\n",
      "14.1107096352\n",
      "2.1822824446\n",
      "t\n",
      "2.18010197913\n",
      "2.20869153412\n",
      "1.97427959603\n",
      "2.00687909101\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n",
      "356.124283\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEWkSFQ4REWkSFQ4REWmS/w+m0MYnzo6h7wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11311c7d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8 \n",
    "## Actual On The fly - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512428817_actual_on_the_fly_encoder_binary8_0_8_67108864'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512424410_actual_on_the_fly_encoder_binary8_0_8_33554432'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512422042_actual_on_the_fly_encoder_binary8_0_8_16777216'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512420706_actual_on_the_fly_encoder_binary8_0_8_8388608'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512419890_actual_on_the_fly_encoder_binary8_0_8_4194304'\n",
    "twenty_mb = 'results/encoder/thousand/1512419331_actual_on_the_fly_encoder_binary8_0_8_2621440'\n",
    "ten_mb = 'results/encoder/thousand/1512418874_actual_on_the_fly_encoder_binary8_0_8_1310720'\n",
    "one_mb = 'results/encoder/thousand/1512418495_actual_on_the_fly_encoder_binary8_0_8_131072'\n",
    "one_kb = 'results/encoder/thousand/1512418184_actual_on_the_fly_encoder_binary8_0_8_128'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print (latencies[2]-float(latencies[1]))/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "print(latency[8])\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/actual_on_the_fly_gen_8.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "79.8724699687\n",
      "13.5074755091\n",
      "2.1858874924\n",
      "t\n",
      "2.12458362774\n",
      "2.346810412\n",
      "2.28408595934\n",
      "2.03123576381\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAELCAYAAADOeWEXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x113017a50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 16 \n",
    "## Actual On The fly - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512440213_actual_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512435529_actual_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512433044_actual_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512431668_actual_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512430840_actual_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "twenty_mb = 'results/encoder/thousand/1512430275_actual_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "ten_mb = 'results/encoder/thousand/1512429814_actual_on_the_fly_encoder_binary8_0_16_655360'\n",
    "one_mb = 'results/encoder/thousand/1512429432_actual_on_the_fly_encoder_binary8_0_16_65536'\n",
    "one_kb = 'results/encoder/thousand/1512429121_actual_on_the_fly_encoder_binary8_0_16_64'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print (latencies[2]-float(latencies[1]))/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/actual_on_the_fly_gen_16.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "66.9768057285\n",
      "12.5640800076\n",
      "2.19826784046\n",
      "t\n",
      "2.07516600669\n",
      "2.08103916055\n",
      "2.40725723207\n",
      "2.3746728919\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11311c750>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 32 \n",
    "## Actual On The fly - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512452319_actual_on_the_fly_encoder_binary8_0_32_16777216'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512447186_actual_on_the_fly_encoder_binary8_0_32_8388608'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512444550_actual_on_the_fly_encoder_binary8_0_32_4194304'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512443124_actual_on_the_fly_encoder_binary8_0_32_2097152'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512442266_actual_on_the_fly_encoder_binary8_0_32_1048576'\n",
    "twenty_mb = 'results/encoder/thousand/1512441685_actual_on_the_fly_encoder_binary8_0_32_655360'\n",
    "ten_mb = 'results/encoder/thousand/1512441215_actual_on_the_fly_encoder_binary8_0_32_327680'\n",
    "one_mb = 'results/encoder/thousand/1512440829_actual_on_the_fly_encoder_binary8_0_32_32768'\n",
    "one_kb = 'results/encoder/thousand/1512440517_actual_on_the_fly_encoder_binary8_0_32_32'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print latencies[1]/float(latencies[0])\n",
    "print (latencies[2]-float(latencies[1]))/float(latencies[1])\n",
    "print latencies[3]/float(latencies[2])\n",
    "print 't'\n",
    "i = 4\n",
    "for x in range(i, (len(latencies)-1)):\n",
    "    print (latencies[x+1]/float(latencies[x]))\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/actual_on_the_fly_gen_32.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.7141581063\n",
      "1.49772279855\n",
      "1.43793225758\n",
      "1.44030766712\n",
      "1.40744367584\n",
      "1.37160179629\n",
      "1.4573738917\n",
      "1.68606678114\n",
      "1.7065298858\n",
      "t\n",
      "2.11420108606\n",
      "1.77285659837\n",
      "1.65757086594\n",
      "1.666958954\n",
      "1.64803306869\n",
      "1.6096999701\n",
      "1.42740489704\n",
      "1.50437891685\n",
      "1.7587361825\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWV36f2IYtN6kS2Y1AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10e8d4850>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8,16,32 \n",
    "## Actual On The Fly - latency comparison\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512428817_actual_on_the_fly_encoder_binary8_0_8_67108864'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512424410_actual_on_the_fly_encoder_binary8_0_8_33554432'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512422042_actual_on_the_fly_encoder_binary8_0_8_16777216'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512420706_actual_on_the_fly_encoder_binary8_0_8_8388608'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512419890_actual_on_the_fly_encoder_binary8_0_8_4194304'\n",
    "    twenty_mb = 'results/encoder/thousand/1512419331_actual_on_the_fly_encoder_binary8_0_8_2621440'\n",
    "    ten_mb = 'results/encoder/thousand/1512418874_actual_on_the_fly_encoder_binary8_0_8_1310720'\n",
    "    one_mb = 'results/encoder/thousand/1512418495_actual_on_the_fly_encoder_binary8_0_8_131072'\n",
    "    one_kb = 'results/encoder/thousand/1512418184_actual_on_the_fly_encoder_binary8_0_8_128'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512440213_actual_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512435529_actual_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512433044_actual_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512431668_actual_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512430840_actual_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512430275_actual_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512429814_actual_on_the_fly_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512429432_actual_on_the_fly_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512429121_actual_on_the_fly_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_thirdy_two():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512452319_actual_on_the_fly_encoder_binary8_0_32_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512447186_actual_on_the_fly_encoder_binary8_0_32_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512444550_actual_on_the_fly_encoder_binary8_0_32_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512443124_actual_on_the_fly_encoder_binary8_0_32_2097152'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512442266_actual_on_the_fly_encoder_binary8_0_32_1048576'\n",
    "    twenty_mb = 'results/encoder/thousand/1512441685_actual_on_the_fly_encoder_binary8_0_32_655360'\n",
    "    ten_mb = 'results/encoder/thousand/1512441215_actual_on_the_fly_encoder_binary8_0_32_327680'\n",
    "    one_mb = 'results/encoder/thousand/1512440829_actual_on_the_fly_encoder_binary8_0_32_32768'\n",
    "    one_kb = 'results/encoder/thousand/1512440517_actual_on_the_fly_encoder_binary8_0_32_32'\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def calc_diff(x,y):\n",
    "    print (y/float(x))\n",
    "    \n",
    "def diff_gen(g, gg):\n",
    "    for x, y in zip(g,gg):\n",
    "        calc_diff(g[x],gg[y])\n",
    "\n",
    "eight_latencies = read_gen_eight()\n",
    "eight_data_size = np.array(eight_latencies.keys())\n",
    "eight_latency = np.array(eight_latencies.values())\n",
    "\n",
    "sixteen_latencies = read_gen_sixteen()\n",
    "diff_gen(eight_latencies,sixteen_latencies)\n",
    "sixteen_data_size = np.array(sixteen_latencies.keys())\n",
    "sixteen_latency = np.array(sixteen_latencies.values())\n",
    "\n",
    "print('t')\n",
    "thridy_two_latencies = read_gen_thirdy_two()\n",
    "diff_gen(sixteen_latencies,thridy_two_latencies)\n",
    "thridy_two_data_size = np.array(thridy_two_latencies.keys())\n",
    "thridy_two_latency = np.array(thridy_two_latencies.values())\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(eight_data_size, my_xticks)\n",
    "plt.plot(eight_data_size, eight_latency)\n",
    "\n",
    "plt.xticks(sixteen_data_size, my_xticks)\n",
    "plt.plot(sixteen_data_size, sixteen_latency)\n",
    "\n",
    "plt.xticks(thridy_two_data_size, my_xticks)\n",
    "plt.plot(thridy_two_data_size, thridy_two_latency)\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['G=8', 'G=16', 'G=32'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "plt.tight_layout()\n",
    "fig.savefig('results/graphs/actual_on_the_fly_gen_comparison.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x: 0 and y: 0\n",
      "8 1KiB diff: 1.1902635105\n",
      "x: 1 and y: 1\n",
      "8 1MiB diff: 1.52205306065\n",
      "x: 2 and y: 2\n",
      "8 10MiB diff: 1.3891052864\n",
      "x: 3 and y: 3\n",
      "8 20MiB diff: 1.34894082695\n",
      "x: 4 and y: 4\n",
      "8 32MiB diff: 1.30559880178\n",
      "x: 5 and y: 5\n",
      "8 64MiB diff: 1.44414974787\n",
      "x: 6 and y: 6\n",
      "8 64MiB diff: 1.50415695984\n",
      "x: 7 and y: 7\n",
      "8 128MiB diff: 1.50661119289\n",
      "x: 8 and y: 8\n",
      "8 256MiB diff: 1.5002575014\n",
      "Average: 1.41234854314\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEMCAYAAADTfFGvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxlSLJQ5jjDHV8v8BfEWpJtHu55gAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1125e3950>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## DIFF TOF and FV for G=8\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def fv_read_gen_eight():\n",
    "    one_kb = 'results/encoder/thousand/1512452623_full_vector_encoder_binary8_0_8_128'\n",
    "    one_mb = 'results/encoder/thousand/1512452934_full_vector_encoder_binary8_0_8_131072'\n",
    "    ten_mb = 'results/encoder/thousand/1512453317_full_vector_encoder_binary8_0_8_1310720'\n",
    "    twenty_mb = 'results/encoder/thousand/1512453778_full_vector_encoder_binary8_0_8_2621440'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512454344_full_vector_encoder_binary8_0_8_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512455182_full_vector_encoder_binary8_0_8_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512456573_full_vector_encoder_binary8_0_8_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512459046_full_vector_encoder_binary8_0_8_33554432'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512463692_full_vector_encoder_binary8_0_8_67108864'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512428817_actual_on_the_fly_encoder_binary8_0_8_67108864'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512424410_actual_on_the_fly_encoder_binary8_0_8_33554432'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512422042_actual_on_the_fly_encoder_binary8_0_8_16777216'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512420706_actual_on_the_fly_encoder_binary8_0_8_8388608'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512419890_actual_on_the_fly_encoder_binary8_0_8_4194304'\n",
    "    twenty_mb = 'results/encoder/thousand/1512419331_actual_on_the_fly_encoder_binary8_0_8_2621440'\n",
    "    ten_mb = 'results/encoder/thousand/1512418874_actual_on_the_fly_encoder_binary8_0_8_1310720'\n",
    "    one_mb = 'results/encoder/thousand/1512418495_actual_on_the_fly_encoder_binary8_0_8_131072'\n",
    "    one_kb = 'results/encoder/thousand/1512418184_actual_on_the_fly_encoder_binary8_0_8_128'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def print_difference(gen, fv, of):\n",
    "    i = 0\n",
    "    ds = ''\n",
    "    avg = 0.0\n",
    "    for x,y in zip(fv,of):\n",
    "        if i == 0:\n",
    "            ds = '1KiB'\n",
    "        elif i == 1:\n",
    "            ds = '1MiB'\n",
    "        elif i == 2:\n",
    "            ds = '10MiB'            \n",
    "        elif i == 3:\n",
    "            ds = '20MiB'\n",
    "        elif i == 4:\n",
    "            ds = '32MiB'            \n",
    "        elif i == 5:\n",
    "            ds = '64MiB'   \n",
    "        elif i == 6:\n",
    "            ds = '64MiB'            \n",
    "        elif i == 7:\n",
    "            ds = '128MiB'            \n",
    "        elif i == 8:\n",
    "            ds = '256MiB'            \n",
    "        elif i == 9:\n",
    "            ds = '512MiB'\n",
    "\n",
    "        res = fv[x]/float(of[y])\n",
    "        avg = avg + res\n",
    "        print('x: {} and y: {}'.format(x,y))\n",
    "        print('{} {} diff: {}'.format(gen, ds, res))\n",
    "        i = i + 1\n",
    "    avg = avg / float(len(fv))\n",
    "    print('Average: {}'.format(avg))\n",
    "        \n",
    "fv_eight_latencies = fv_read_gen_eight()\n",
    "fv_eight_data_size = np.array(fv_eight_latencies.keys())\n",
    "fv_eight_latency = np.array(fv_eight_latencies.values())\n",
    "\n",
    "\n",
    "of_eight_latencies = of_read_gen_eight()\n",
    "of_eight_data_size = np.array(of_eight_latencies.keys())\n",
    "of_eight_latency = np.array(of_eight_latencies.values())\n",
    "\n",
    "print_difference('8', fv_eight_latencies, of_eight_latencies)\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "\n",
    "plt.xticks(fv_eight_data_size, my_xticks)\n",
    "plt.plot(fv_eight_data_size, fv_eight_latency, linestyle='-')\n",
    "\n",
    "plt.xticks(of_eight_data_size, my_xticks)\n",
    "plt.plot(of_eight_data_size, of_eight_latency, linestyle='-')\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['FV G=8','TOF G=8'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/fv_aov_gen_8_comparison.eps')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x: 0 and y: 0\n",
      "8 1KiB diff: 1.36234984367\n",
      "x: 1 and y: 1\n",
      "8 1MiB diff: 1.72820750538\n",
      "x: 2 and y: 2\n",
      "8 10MiB diff: 1.68023879424\n",
      "x: 3 and y: 3\n",
      "8 20MiB diff: 1.62089469158\n",
      "x: 4 and y: 4\n",
      "8 32MiB diff: 1.56682120335\n",
      "x: 5 and y: 5\n",
      "8 64MiB diff: 1.56151624324\n",
      "x: 6 and y: 6\n",
      "8 64MiB diff: 1.6734377027\n",
      "x: 7 and y: 7\n",
      "8 128MiB diff: 1.7228423924\n",
      "x: 8 and y: 8\n",
      "8 256MiB diff: 1.70550055875\n",
      "Average: 1.62464543726\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncRhjjCkXSxzGGGPK5f8DOEOyfJfo91cAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10e8bf050>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## DIFF TOF and FV for G=16\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def fv_read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512475943_full_vector_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512470804_full_vector_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512468085_full_vector_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512466613_full_vector_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512465749_full_vector_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512465168_full_vector_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512464696_full_vector_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512464308_full_vector_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512463997_full_vector_encoder_binary8_0_16_64'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512440213_actual_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512435529_actual_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512433044_actual_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512431668_actual_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512430840_actual_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512430275_actual_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512429814_actual_on_the_fly_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512429432_actual_on_the_fly_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512429121_actual_on_the_fly_encoder_binary8_0_16_64'   \n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512440213_actual_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512435529_actual_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512433044_actual_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512431668_actual_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512430840_actual_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512430275_actual_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512429814_actual_on_the_fly_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512429432_actual_on_the_fly_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512429121_actual_on_the_fly_encoder_binary8_0_16_64'    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "\n",
    "def print_difference(gen, fv, of):\n",
    "    i = 0\n",
    "    ds = ''\n",
    "    avg = 0.0\n",
    "    for x,y in zip(fv,of):\n",
    "        if i == 0:\n",
    "            ds = '1KiB'\n",
    "        elif i == 1:\n",
    "            ds = '1MiB'\n",
    "        elif i == 2:\n",
    "            ds = '10MiB'            \n",
    "        elif i == 3:\n",
    "            ds = '20MiB'\n",
    "        elif i == 4:\n",
    "            ds = '32MiB'            \n",
    "        elif i == 5:\n",
    "            ds = '64MiB'   \n",
    "        elif i == 6:\n",
    "            ds = '64MiB'            \n",
    "        elif i == 7:\n",
    "            ds = '128MiB'            \n",
    "        elif i == 8:\n",
    "            ds = '256MiB'            \n",
    "        elif i == 9:\n",
    "            ds = '512MiB'\n",
    "\n",
    "        res = fv[x]/float(of[y])\n",
    "        avg = avg + res\n",
    "        print('x: {} and y: {}'.format(x,y))\n",
    "        print('{} {} diff: {}'.format(gen, ds, res))\n",
    "        i = i + 1\n",
    "    avg = avg / float(len(fv))\n",
    "    print('Average: {}'.format(avg))\n",
    "        \n",
    "fv_eight_latencies = fv_read_gen_eight()\n",
    "fv_eight_data_size = np.array(fv_eight_latencies.keys())\n",
    "fv_eight_latency = np.array(fv_eight_latencies.values())\n",
    "\n",
    "\n",
    "of_eight_latencies = of_read_gen_eight()\n",
    "of_eight_data_size = np.array(of_eight_latencies.keys())\n",
    "of_eight_latency = np.array(of_eight_latencies.values())\n",
    "\n",
    "print_difference('8', fv_eight_latencies, of_eight_latencies)\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "\n",
    "plt.xticks(fv_eight_data_size, my_xticks)\n",
    "plt.plot(fv_eight_data_size, fv_eight_latency, linestyle='-')\n",
    "\n",
    "plt.xticks(of_eight_data_size, my_xticks)\n",
    "plt.plot(of_eight_data_size, of_eight_latency, linestyle='-')\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['FV G=16','TOF G=16'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/fv_aov_gen_16_comparison.eps')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/larsnielsen/miniconda2/lib/python2.7/site-packages/matplotlib/__init__.py:1405: UserWarning: \n",
      "This call to matplotlib.use() has no effect because the backend has already\n",
      "been chosen; matplotlib.use() must be called *before* pylab, matplotlib.pyplot,\n",
      "or matplotlib.backends is imported for the first time.\n",
      "\n",
      "  warnings.warn(_use_error_msg)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x: 0 and y: 0\n",
      "8 1KiB diff: 1.46801058531\n",
      "x: 1 and y: 1\n",
      "8 1MiB diff: 1.80927650555\n",
      "x: 2 and y: 2\n",
      "8 10MiB diff: 1.8838786563\n",
      "x: 3 and y: 3\n",
      "8 20MiB diff: 1.8071094939\n",
      "x: 4 and y: 4\n",
      "8 32MiB diff: 1.74723815671\n",
      "x: 5 and y: 5\n",
      "8 64MiB diff: 1.72916580804\n",
      "x: 6 and y: 6\n",
      "8 64MiB diff: 1.72296042331\n",
      "x: 7 and y: 7\n",
      "8 128MiB diff: 1.79626723854\n",
      "x: 8 and y: 8\n",
      "8 256MiB diff: 1.84508668758\n",
      "Average: 1.75655483947\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ntpuW7QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1061aab90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## DIFF TOF and FV for G=32\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def fv_read_gen_eight():\n",
    "    one_kb = 'results/encoder/thousand/1512476247_full_vector_encoder_binary8_0_32_32'\n",
    "    one_mb = 'results/encoder/thousand/1512476560_full_vector_encoder_binary8_0_32_32768'\n",
    "    ten_mb = 'results/encoder/thousand/1512476958_full_vector_encoder_binary8_0_32_327680'\n",
    "    twenty_mb = 'results/encoder/thousand/1512477451_full_vector_encoder_binary8_0_32_655360'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512478066_full_vector_encoder_binary8_0_32_1048576'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512478997_full_vector_encoder_binary8_0_32_2097152'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512480568_full_vector_encoder_binary8_0_32_4194304'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512483577_full_vector_encoder_binary8_0_32_8388608'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512489653_full_vector_encoder_binary8_0_32_16777216'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512452319_actual_on_the_fly_encoder_binary8_0_32_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512447186_actual_on_the_fly_encoder_binary8_0_32_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512444550_actual_on_the_fly_encoder_binary8_0_32_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512443124_actual_on_the_fly_encoder_binary8_0_32_2097152'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512442266_actual_on_the_fly_encoder_binary8_0_32_1048576'\n",
    "    twenty_mb = 'results/encoder/thousand/1512441685_actual_on_the_fly_encoder_binary8_0_32_655360'\n",
    "    ten_mb = 'results/encoder/thousand/1512441215_actual_on_the_fly_encoder_binary8_0_32_327680'\n",
    "    one_mb = 'results/encoder/thousand/1512440829_actual_on_the_fly_encoder_binary8_0_32_32768'\n",
    "    one_kb = 'results/encoder/thousand/1512440517_actual_on_the_fly_encoder_binary8_0_32_32'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "\n",
    "\n",
    "def print_difference(gen, fv, of):\n",
    "    i = 0\n",
    "    ds = ''\n",
    "    avg = 0.0\n",
    "    for x,y in zip(fv,of):\n",
    "        if i == 0:\n",
    "            ds = '1KiB'\n",
    "        elif i == 1:\n",
    "            ds = '1MiB'\n",
    "        elif i == 2:\n",
    "            ds = '10MiB'            \n",
    "        elif i == 3:\n",
    "            ds = '20MiB'\n",
    "        elif i == 4:\n",
    "            ds = '32MiB'            \n",
    "        elif i == 5:\n",
    "            ds = '64MiB'   \n",
    "        elif i == 6:\n",
    "            ds = '64MiB'            \n",
    "        elif i == 7:\n",
    "            ds = '128MiB'            \n",
    "        elif i == 8:\n",
    "            ds = '256MiB'            \n",
    "        elif i == 9:\n",
    "            ds = '512MiB'\n",
    "\n",
    "        res = fv[x]/float(of[y])\n",
    "        avg = avg + res\n",
    "        print('x: {} and y: {}'.format(x,y))\n",
    "        print('{} {} diff: {}'.format(gen, ds, res))\n",
    "        i = i + 1\n",
    "    avg = avg / float(len(fv))\n",
    "    print('Average: {}'.format(avg))\n",
    "        \n",
    "fv_eight_latencies = fv_read_gen_eight()\n",
    "fv_eight_data_size = np.array(fv_eight_latencies.keys())\n",
    "fv_eight_latency = np.array(fv_eight_latencies.values())\n",
    "\n",
    "\n",
    "of_eight_latencies = of_read_gen_eight()\n",
    "of_eight_data_size = np.array(of_eight_latencies.keys())\n",
    "of_eight_latency = np.array(of_eight_latencies.values())\n",
    "\n",
    "print_difference('8', fv_eight_latencies, of_eight_latencies)\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "\n",
    "plt.xticks(fv_eight_data_size, my_xticks)\n",
    "plt.plot(fv_eight_data_size, fv_eight_latency, linestyle='-')\n",
    "\n",
    "plt.xticks(of_eight_data_size, my_xticks)\n",
    "plt.plot(of_eight_data_size, of_eight_latency, linestyle='-')\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['FV G=32','TOF G=32'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/fv_aov_gen_32_comparison.eps')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.1902635105\n",
      "1.52205306065\n",
      "1.3891052864\n",
      "1.34894082695\n",
      "1.30559880178\n",
      "1.44414974787\n",
      "1.50415695984\n",
      "1.50661119289\n",
      "1.5002575014\n",
      "AVG: 1.41234854314\n",
      "t\n",
      "1.36234984367\n",
      "1.72820750538\n",
      "1.68023879424\n",
      "1.62089469158\n",
      "1.56682120335\n",
      "1.56151624324\n",
      "1.6734377027\n",
      "1.7228423924\n",
      "1.70550055875\n",
      "AVG: 1.62464543726\n",
      "t\n",
      "1.46801058531\n",
      "1.80927650555\n",
      "1.8838786563\n",
      "1.8071094939\n",
      "1.74723815671\n",
      "1.72916580804\n",
      "1.72296042331\n",
      "1.79626723854\n",
      "1.84508668758\n",
      "AVG: 1.75655483947\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMSGxxGGMMSYk/x/uzGWdpQmMcwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x113143b10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def fv_read_gen_eight():\n",
    "    one_kb = 'results/encoder/thousand/1512452623_full_vector_encoder_binary8_0_8_128'\n",
    "    one_mb = 'results/encoder/thousand/1512452934_full_vector_encoder_binary8_0_8_131072'\n",
    "    ten_mb = 'results/encoder/thousand/1512453317_full_vector_encoder_binary8_0_8_1310720'\n",
    "    twenty_mb = 'results/encoder/thousand/1512453778_full_vector_encoder_binary8_0_8_2621440'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512454344_full_vector_encoder_binary8_0_8_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512455182_full_vector_encoder_binary8_0_8_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512456573_full_vector_encoder_binary8_0_8_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512459046_full_vector_encoder_binary8_0_8_33554432'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512463692_full_vector_encoder_binary8_0_8_67108864'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def fv_read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512475943_full_vector_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512470804_full_vector_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512468085_full_vector_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512466613_full_vector_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512465749_full_vector_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512465168_full_vector_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512464696_full_vector_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512464308_full_vector_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512463997_full_vector_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def fv_read_gen_thirdy_two():\n",
    "    one_kb = 'results/encoder/thousand/1512476247_full_vector_encoder_binary8_0_32_32'\n",
    "    one_mb = 'results/encoder/thousand/1512476560_full_vector_encoder_binary8_0_32_32768'\n",
    "    ten_mb = 'results/encoder/thousand/1512476958_full_vector_encoder_binary8_0_32_327680'\n",
    "    twenty_mb = 'results/encoder/thousand/1512477451_full_vector_encoder_binary8_0_32_655360'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512478066_full_vector_encoder_binary8_0_32_1048576'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512478997_full_vector_encoder_binary8_0_32_2097152'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512480568_full_vector_encoder_binary8_0_32_4194304'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512483577_full_vector_encoder_binary8_0_32_8388608'\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512489653_full_vector_encoder_binary8_0_32_16777216'\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512428817_actual_on_the_fly_encoder_binary8_0_8_67108864'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512424410_actual_on_the_fly_encoder_binary8_0_8_33554432'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512422042_actual_on_the_fly_encoder_binary8_0_8_16777216'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512420706_actual_on_the_fly_encoder_binary8_0_8_8388608'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512419890_actual_on_the_fly_encoder_binary8_0_8_4194304'\n",
    "    twenty_mb = 'results/encoder/thousand/1512419331_actual_on_the_fly_encoder_binary8_0_8_2621440'\n",
    "    ten_mb = 'results/encoder/thousand/1512418874_actual_on_the_fly_encoder_binary8_0_8_1310720'\n",
    "    one_mb = 'results/encoder/thousand/1512418495_actual_on_the_fly_encoder_binary8_0_8_131072'\n",
    "    one_kb = 'results/encoder/thousand/1512418184_actual_on_the_fly_encoder_binary8_0_8_128'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512440213_actual_on_the_fly_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512435529_actual_on_the_fly_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512433044_actual_on_the_fly_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512431668_actual_on_the_fly_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512430840_actual_on_the_fly_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512430275_actual_on_the_fly_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512429814_actual_on_the_fly_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512429432_actual_on_the_fly_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512429121_actual_on_the_fly_encoder_binary8_0_16_64'    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def of_read_gen_thirdy_two():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512452319_actual_on_the_fly_encoder_binary8_0_32_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512447186_actual_on_the_fly_encoder_binary8_0_32_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512444550_actual_on_the_fly_encoder_binary8_0_32_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512443124_actual_on_the_fly_encoder_binary8_0_32_2097152'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512442266_actual_on_the_fly_encoder_binary8_0_32_1048576'\n",
    "    twenty_mb = 'results/encoder/thousand/1512441685_actual_on_the_fly_encoder_binary8_0_32_655360'\n",
    "    ten_mb = 'results/encoder/thousand/1512441215_actual_on_the_fly_encoder_binary8_0_32_327680'\n",
    "    one_mb = 'results/encoder/thousand/1512440829_actual_on_the_fly_encoder_binary8_0_32_32768'\n",
    "    one_kb = 'results/encoder/thousand/1512440517_actual_on_the_fly_encoder_binary8_0_32_32'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "\n",
    "def print_difference(gen, fv, of):\n",
    "    i = 0\n",
    "    ds = ''\n",
    "    for x,y in zip(fv,of):\n",
    "        if i == 0:\n",
    "            ds = '1KiB'\n",
    "        elif i == 1:\n",
    "            ds = '1MiB'\n",
    "        elif i == 2:\n",
    "            ds = '10MiB'            \n",
    "        elif i == 3:\n",
    "            ds = '20MiB'\n",
    "        elif i == 4:\n",
    "            ds = '32MiB'            \n",
    "        elif i == 5:\n",
    "            ds = '64MiB'   \n",
    "        elif i == 6:\n",
    "            ds = '64MiB'            \n",
    "        elif i == 7:\n",
    "            ds = '128MiB'            \n",
    "        elif i == 8:\n",
    "            ds = '256MiB'            \n",
    "        elif i == 9:\n",
    "            ds = '512MiB'\n",
    "        if y == 0:\n",
    "            res = 0\n",
    "        else:\n",
    "            res = float(x)/float(y)\n",
    "        print('{} {} diff: {}'.format(gen, ds, res))\n",
    "        i = i + 1\n",
    "        \n",
    "def calc_diff(x,y):\n",
    "    res = y/float(x)\n",
    "    print (res)\n",
    "    return res\n",
    "def diff_gen(g, gg):\n",
    "    res = list()\n",
    "    for x, y in zip(g,gg):\n",
    "        res.append(calc_diff(g[x],gg[y]))\n",
    "    print('AVG: {}'.format(np.average(res)))\n",
    "        \n",
    "fv_eight_latencies = fv_read_gen_eight()\n",
    "fv_eight_data_size = np.array(fv_eight_latencies.keys())\n",
    "fv_eight_latency = np.array(fv_eight_latencies.values())\n",
    "\n",
    "fv_sixteen_latencies = fv_read_gen_sixteen()\n",
    "fv_sixteen_data_size = np.array(fv_sixteen_latencies.keys())\n",
    "fv_sixteen_latency = np.array(fv_sixteen_latencies.values())\n",
    "\n",
    "fv_thridy_two_latencies = fv_read_gen_thirdy_two()\n",
    "fv_thridy_two_data_size = np.array(fv_thridy_two_latencies.keys())\n",
    "fv_thridy_two_latency = np.array(fv_thridy_two_latencies.values())\n",
    "\n",
    "of_eight_latencies = of_read_gen_eight()\n",
    "\n",
    "diff_gen(of_eight_latencies,fv_eight_latencies)\n",
    "\n",
    "of_eight_data_size = np.array(of_eight_latencies.keys())\n",
    "of_eight_latency = np.array(of_eight_latencies.values())\n",
    "\n",
    "#print_difference('8', fv_eight_latencies, of_eight_latencies)\n",
    "\n",
    "print('t')\n",
    "of_sixteen_latencies = of_read_gen_sixteen()\n",
    "diff_gen(of_sixteen_latencies, fv_sixteen_latencies)\n",
    "of_sixteen_data_size = np.array(of_sixteen_latencies.keys())\n",
    "of_sixteen_latency = np.array(of_sixteen_latencies.values())\n",
    "\n",
    "print('t')\n",
    "of_thridy_two_latencies = of_read_gen_thirdy_two()\n",
    "diff_gen(of_thridy_two_latencies, fv_thridy_two_latencies)\n",
    "of_thridy_two_data_size = np.array(of_thridy_two_latencies.keys())\n",
    "of_thridy_two_latency = np.array(of_thridy_two_latencies.values())\n",
    "\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "\n",
    "plt.xticks(fv_eight_data_size, my_xticks)\n",
    "plt.plot(fv_eight_data_size, fv_eight_latency, marker='v', linestyle='')\n",
    "\n",
    "plt.xticks(fv_sixteen_data_size, my_xticks)\n",
    "plt.plot(fv_sixteen_data_size, fv_sixteen_latency, marker='v', linestyle='')\n",
    "\n",
    "plt.xticks(fv_thridy_two_data_size, my_xticks)\n",
    "plt.plot(fv_thridy_two_data_size, fv_thridy_two_latency, marker='v', linestyle='')\n",
    "\n",
    "plt.xticks(of_eight_data_size, my_xticks)\n",
    "plt.plot(of_eight_data_size, of_eight_latency, marker='p', linestyle='')\n",
    "\n",
    "plt.xticks(of_sixteen_data_size, my_xticks)\n",
    "plt.plot(of_sixteen_data_size, of_sixteen_latency, marker='p', linestyle='')\n",
    "\n",
    "plt.xticks(of_thridy_two_data_size, my_xticks)\n",
    "plt.plot(of_thridy_two_data_size, of_thridy_two_latency, marker='p', linestyle='')\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['FV G=8', 'FV G=16', 'FV G=32', 'OTF G=8', 'OTF G=16', 'OTF G=32'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/fv_aov_comparison.eps')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xl8XHW9//HXJ3ubpnua7htLKV1pIwIisggWRCgFCgL+\n8Mq9CCIXuSqi8hC9ilRQlKvIlYsKXhC6AIqscpHVBZoudC/UbkmbNmnakqZpkib5/P44J2UaJu20\nzcyZJO/n49HHzJw5Z85nppPznu/3nPM95u6IiIi0lhF1ASIikp4UECIiEpcCQkRE4lJAiIhIXAoI\nERGJSwEhIiJxKSBERCQuBYSIiMSlgBARkbiyoi7gSPTv399HjhwZdRkiIh3KggULtrl74cHm69AB\nMXLkSEpKSqIuQ0SkQzGzDYnMpy4mERGJSwEhIiJxKSBERCSuDr0PIp69e/dSVlZGXV1d1KUkVV5e\nHkOHDiU7OzvqUkSkk0paQJjZb4DzgQp3Hx9O6wvMBkYC64GZ7r4jfO6bwDVAE/Dv7v7i4ay3rKyM\ngoICRo4ciZkd8ftIR+5OVVUVZWVljBo1KupyRKSTSmYX00PAtFbTbgVedvdjgJfDx5jZ8cDlwLhw\nmV+aWebhrLSuro5+/fp12nAAMDP69evX6VtJIhKtpAWEu78ObG81+ULg4fD+w8D0mOmPu3u9u68D\n1gAnHu66O3M4tOgK71FEopXqndRF7l4e3t8CFIX3hwClMfOVhdNERKSV37y5jj8v35L09UR2FJMH\nF8M+5Atim9m1ZlZiZiWVlZVJqKx9bN26lSuuuILRo0czdepUTj75ZJ566qmEln3ssceYMGECEydO\nZNq0aWzbti3J1YpIR1FT38iP/7yal1dWJH1dqQ6IrWY2CCC8bXmHm4BhMfMNDad9iLs/4O7F7l5c\nWHjQM8Uj4e5Mnz6d0047jbVr17JgwQIef/xxysrKDrpsY2MjN910E6+88gpLlixh4sSJ/OIXv0hB\n1SLSETy7ZDO1DU3M/Miwg898hFIdEE8DV4f3rwb+GDP9cjPLNbNRwDHA2ymurd385S9/IScnh+uu\nu27ftBEjRnDjjTcedFl3x93ZvXs37k51dTWDBw9OZrki0oHMnl/K0QN6MGV476SvK5mHuT4GnA70\nN7My4HZgFjDHzK4BNgAzAdx9uZnNAVYAjcAN7t50pDV870/LWbG5+khfZj/HD+7J7Z8Zd8B5li9f\nzpQpU+I+t3r1ai677LK4z7366qv07t2b+++/nwkTJpCfn88xxxzDfffdd8R1i0jHt6ZiFws37uTb\n541NyYEqSQsId/9sG0+d1cb8dwB3JKueKN1www28+eab5OTkMH/+fBYvXtzmvHv37uX+++9n0aJF\njB49mhtvvJE777yT2267LYUVi0g6mj2/lKwM46IpqTmGp9OdSR3rYL/0k2XcuHE88cQT+x7fd999\nbNu2jeLi4oO2IN577z0AjjrqKABmzpzJrFmzkl+0iKS1hsZmnly4ibPGDqB/j9yUrLNTB0RUzjzz\nTL71rW9x//33c/311wNQW1sLwJgxYw7YghgyZAgrVqygsrKSwsJCXnrpJcaOHZuSukUkff1l1Vaq\ndjdwWQp2TrdQQCSBmfGHP/yBm2++mbvuuovCwkLy8/P50Y9+dNBlBw8ezO23385pp51GdnY2I0aM\n4KGHHkp+0SKS1uaUlFHUM5fTjknd0ZsKiCQZNGgQjz/++GEte9111+13BJSIdG1b3q/j1dUVXH/6\nUWRlpu7gUw33LSKS5p5YWEazw8zi1HUvgQJCRCStNTc7c0pKOWl0X0b0y0/pujtlQASjeHRuXeE9\nigi8tW47G6pqU7pzukWnC4i8vDyqqqo69Qa05XoQeXl5UZciIkk2p6SUgrwszh0/KOXr7nQ7qYcO\nHUpZWRnpPJBfe2i5opyIdF7v79nLc0vLuWTqUPKyD+sSOUek0wVEdna2rrImIp3C0+9spr6xOZLu\nJeiEXUwiIp3FnPmlHDewgAlDekWyfgWEiEgaWrG5mqWb3ueyjwyL7AqSCggRkTQ0p6SUnMwMpk+O\n7uKaCggRkTRT39jEHxZv4pxxRfTJz4msDgWEiEia+fPyreys3RvZzukWCggRkTQzp6SUIb278bGj\n+kdahwJCRCSNlO2o5c0127hk6lAyMqLZOd1CASEikkbmlpQBcGlx9CfCKiBERNJEU7Mzb0EZpx7d\nn6F9ukddjgJCRCRd/HXNNjbt3JPyYb3booAQEUkTc0pK6d09m3PGFUVdCqCAEBFJCzt2N/Dn5VuZ\nPnkIuVmpH5gvHgWEiEga+MPiTTQ0RTcwXzwKCBGRiLk7s+eXMnFoL8YO6hl1OfsoIEREIrZ00/us\n2rKLS9Nk53QLBYSISMRmzy8lNyuDCyYNjrqU/SggREQitKehiacXb+a8CYPo1S076nL2o4AQEYnQ\n88vK2VXfmDbnPsRSQIiIRGhOSSkj+nXnpNF9oy7lQxQQIiIR2VC1m3+s3c7M4uiuGncgCggRkYjM\nKSklw+DiKdEPzBdPJAFhZjeb2XIzW2Zmj5lZnpn1NbOXzOy98LZPFLWJiKRCY1Mz8xaUcfqYAQzs\nlRd1OXGlPCDMbAjw70Cxu48HMoHLgVuBl939GODl8LGISKf0+nuVbK2uZ2YaDOvdlqi6mLKAbmaW\nBXQHNgMXAg+Hzz8MTI+oNhGRpJs9v5R++TmceVx6DMwXT8oDwt03AT8GNgLlwPvu/megyN3Lw9m2\nAHE/NTO71sxKzKyksrIyJTWLiLSnyl31vLyyghlThpCTlb67gqPoYupD0FoYBQwG8s3sqth53N0B\nj7e8uz/g7sXuXlxYWJj0ekVE2ttTi8pobPa0Gpgvniii65PAOnevdPe9wJPAKcBWMxsEEN5WRFCb\niEhStQzMN2V4b44eUBB1OQcURUBsBE4ys+4WHPh7FrASeBq4OpznauCPEdQmIpJUCzfu5J+Vu9O+\n9QDBzuKUcve3zGwesBBoBBYBDwA9gDlmdg2wAZiZ6tpERJJtzvxSuudk8umJ6TUwXzwpDwgAd78d\nuL3V5HqC1oSISKe0u76RZ5Zs5vyJg+iRG8nm95Ck7+5zEZFO5tkl5exuaErLgfniUUCIiKTI7JJS\nRhfmM3VExxgoQgEhIpICayp2sWDDDi5L04H54lFAiIikwJySMrIyjBlpOjBfPAoIEZEk29vUzJML\nyzjzuAEUFuRGXU7CFBAiIkn28soKttU0dIhzH2IpIEREkmxuSSkDCnL5xLEda3ggBYSISBJtra7j\nldUVXDJ1KFmZHWuT27GqFRHpYOYtKKPZ4dIOcu5DLAWEiEiSuDtzS0o5cVRfRvXPj7qcQ6aAEBFJ\nkrfWbWd9VS2XdcDWAyggRESSZs78UgpyszhvwqCoSzksCggRkSSortvLc8vK+czkwXTLyYy6nMOi\ngBARSYKnF2+mbm9zh+1eAgWEiEhSzCkp5biBBUwc2ivqUg6bAkJEpJ2tLK9mSdn7zOxAA/PFo4AQ\nEWlnc0pKyc40pp8wJOpSjogCQkSkHdU3NvHUok2cc/xA+ubnRF3OEVFAiIi0o5dWbGVn7V5mdrCB\n+eJRQIiItKPZ80sZ3CuPU4/uH3UpR+yQAsLM+pjZxGQVIyLSkZXtqOXNNdu4pHgYmRkdd+d0i4MG\nhJm9amY9zawvsBD4HzO7J/mliYh0LPMWlAFw6dSOc9W4A0mkBdHL3auBGcDv3P2jwCeTW5aISMfS\n3OzMLSnjY0f1Z1jf7lGX0y4SCYgsMxsEzASeSXI9IiId0t/+WcWmnXs6xc7pFokExH8CLwJr3H2+\nmY0G3ktuWSIiHcvsklJ6dcvmnOOLoi6l3WQdbAZ3nwvMjXm8Frg4mUWJiHQkO2sbeHH5Fj77kWHk\nZXfMgfniOWhAmNko4EZgZOz87n5B8soSEek4/rBoEw2NzZ2qewkSCAjgD8CvgT8BzcktR0SkY3F3\nZpeUMX5IT8YN7rgD88WTSEDUuft/Jb0SEZEOaNmmalaWV/P9C8dFXUq7SyQg7jWz24E/A/UtE919\nYdKqEhHpIGaXbCQ3K4MLJnfsgfniSSQgJgCfA87kgy4mDx8fFjPrDTwIjA9f6wvAamA2wb6O9cBM\nd99xuOsQEUm2ur1N/HHxZs4dP5Be3bKjLqfdJRIQlwKj3b2hHdd7L/CCu19iZjlAd+BbwMvuPsvM\nbgVuBb7RjusUEWlXzy8rZ1ddY6fbOd0ikfMglgG922uFZtYLOI1gxzfu3uDuO4ELgYfD2R4GprfX\nOkVEkmHO/DKG9e3GSaP6RV1KUiTSgugNrDKz+ey/D+JwD3MdBVQCvzWzScAC4CagyN3Lw3m2AJ3n\nbBMR6XQ2VO3m72ur+OrZx5LRCQbmiyeRgLg9CeucAtzo7m+Z2b0E3Un7uLubmcdb2MyuBa4FGD58\neDuXJiKSmLklZWQYXFLcOQbmiyeRM6lfa+d1lgFl7v5W+HgeQUBsNbNB7l4ejv1U0UY9DwAPABQX\nF8cNERGRZGpqduYtKOO0YwsZ1Ktb1OUkTcovGOTuW4BSMxsTTjoLWAE8DVwdTrsa+GOqaxMRScTr\n71aypbqOy4o7587pFol0MSXDjcCj4RFMa4F/IQirOWZ2DbCBYPRYEZG0M3t+Kf3yczhrbOfeVRpJ\nQLj7YqA4zlNnpboWEZFDsa2mnv9buZXPnzKSnKzOfdXmRAbr+xjwXWBEOL8R7EcendzSRETSz1ML\nN9HY7J323IdYibQgfg3cTHA4alNyyxERSV/BwHylTB7Wm2OLCqIuJ+kSCYj33f35pFciIpLmFpXu\nZE1FDXfOmBB1KSmRSEC8YmZ3A0+iwfpEpAubM7+UbtmZnD9xUNSlpEQiAfHR8DZ2p/IRDdYnItLR\n7K5v5E/vbObTEwdRkNf5BuaLJ5ET5c5IRSEiIuns2aXl7G5o4rIusHO6RZsBYWZXufsjZvYf8Z53\n93uSV5aISHqZM7+U0YX5FI/oE3UpKXOgg3jzw9uCNv6JiHQJaypqKNmwg5nFwzDrnAPzxdNmC8Ld\nfxXefi915YiIpJ+5JaVkZhgzpnS+q8YdSOc+DVBE5AjtbWrmiYVlnDFmAAMK8qIuJ6UUECIiB/CX\nVRVsq2noUjunWxw0IMwsMxWFiIiko7klpRQW5HLGmMKoS0m5RFoQ75nZ3WZ2fNKrERFJIxXVdbyy\nupKLpwwlK7Prdbgk8o4nAe8CD5rZP8zsWjPrmeS6REQiN29hGU3NzsxOfNW4AzloQLj7Lnf/H3c/\nBfgGwSVIy83sYTM7OukViohEoLpuL4/+YyMnjuzL6MIeUZcTiYT2QZjZBWb2FPAz4CfAaOBPwHNJ\nrk9EJOXcnVvmLmFLdR1fnzbm4At0UomMxfQe8Apwt7v/LWb6PDM7LTlliYhE58E31vHC8i3c9umx\nfGRk36jLiUwiATHR3WviPeHu/97O9YiIROqttVXMemEV544fyDWnjoq6nEglspP6PjPr3fLAzPqY\n2W+SWJOISCQqquv48mOLGNG3O3ddMrFLDasRTyIBMdHdd7Y8cPcdwAnJK0lEJPX2NjXz5d8voqau\nkfuvmtplhvQ+kEQCIsPM9g1faGZ9SaxrSkSkw7j7xdW8vX47d86YwJiBGo8UEtvQ/wT4u5nNBQy4\nBLgjqVWJiKTQC8vKeeD1tXzupBFMP6FrDch3IIlcMOh3ZrYAaLlw0Ax3X5HcskREUmNtZQ1fm7uE\nScN6c9v5Y6MuJ60k2lW0CtjRMr+ZDXf3jUmrSkQkBWobGrn+kYVkZxq/vHIKuVkaei7WQQPCzG4k\nOHt6K9BE0M3kwMTkliYikjzuzrefWsa7Fbt46F9OZEjvblGXlHYSaUHcBIxx96pkFyMikiqPvrWR\npxZt4uZPHssnju16I7UmIpGjmEqB95NdiIhIqrxTupP//NMKTh9TyI1naki5tiTSglgLvGpmzwL1\nLRPd/Z6kVSUikiQ7djfwpUcXUliQy09nTiYjo2ufDHcgiQTExvBfTvhPRKRDamp2bpq9mMpd9cy7\n/mT65GuTdiCJHOb6PQAz6+7utckvSUQkOX7+l/d4/d1K7rhoPBOH9j74Al1cIsN9n2xmKwgOdcXM\nJpnZL5NemYhIO3p1dQX3vvweM6YM4YoTh0ddToeQyE7qnwGfAqoA3P0d4IiH+Q6vM7HIzJ4JH/c1\ns5fM7L3wts/BXkNEJBFlO2r5yuzFjCkq4I7pE7r8IHyJSugiq+5e2mpSUzus+yZgZczjW4GX3f0Y\n4OXwsYjIEalvbOKGRxfS1OTcf9VUuuXoZLhEJXSYq5mdAriZZZvZ19h/w37IzGwo8GngwZjJFwIP\nh/cfBqYfyTpERAC+/8wK3il7n7svncSo/vlRl9OhJBIQ1wE3AEOATcBk4EtHuN6fAbcAzTHTity9\nPLy/BSg6wnWISBf35MIyHvnHRr542mimjR8YdTkdTiIBMcbdr3T3Incf4O5XAYc9opWZnQ9UuPuC\ntuZxdycYziPe8teaWYmZlVRWVh5uGSLSya3aUs23nlrKiaP68vVPdd3rSh+JRALi5wlOS9THgAvM\nbD3wOHCmmT0CbDWzQQDhbUW8hd39AXcvdvfiwkKdHi8iH7arbi/XP7KQgrxsfnHFCWRlJrS7VVpp\n8zwIMzsZOAUoNLP/iHmqJ3DYe3nc/ZvAN8N1nA58zd2vMrO7gauBWeHtHw93HSLSdbk7X5+7hI3b\na3ns305iQEFe1CV1WAc6US4H6BHOE3t5pWqCiwa1t1nAHDO7BtgAzEzCOkSkk3vwjXW8sHwLt316\nLCeO6ht1OR1amwHh7q8Br5nZQ+6+IRkrd/dXgVfD+1XAWclYj4h0DW+trWLWC6s4d/xArjl1VNTl\ndHiJjMVUG3b/jAP2tdXc/cykVSUicogqquv48mOLGN63O3ddMlEnw7WDRPbcPEowzMYo4HvAemB+\nEmsSETkkjU3NfPmxRdTUNfLfV02lIC876pI6hUQCop+7/xrY6+6vufsXALUeRCRt3P3iat5et507\nZ0xgzMCCgy8gCUmki2lveFtuZp8GNgPa8yMiaeGFZeX86vW1fO6kEUw/YUjU5XQqiQTED8ysF/BV\ngvMfegJfSWpVIiIJWFtZw9fmLmHSsN7cdv5hn78rbUjkehDPhHffB84AMDMFhIhEak9DE196dCHZ\nmcYvr5xCbpYG4Wtvh3t64X8cfBYRkeRwd7791FJWb93Fzy4/gSG9u0VdUqd0uAGh48dEJDK/f3sj\nTy7axFfOOpZPHKshd5LlcAMi7kB6IiLJ9k7pTr739ApOH1PIjWceHXU5ndqBxmLaRfwgMEDtORFJ\nuR27G/jSowspLMjlpzMnk5GhzoxkOtBQGzqYWETSRnOz85XZi6ncVc/c606mT35O1CV1ehoDV0Q6\nhJ//ZQ2vvVvJ7Rccz6RhvaMup0tQQIhI2nvt3Up+9vK7zJgyhCtOHB51OV2GAkJE0lrZjlpuenwR\nY4oKuGP6BA3Cl0IKCBFJW/WNTdzw6EKampz7r5pKtxydDJdKiQy1ISISie8/s4J3yt7nv6+ayqj+\n+VGX0+WoBSEiaempRWU88o+NfPG00UwbPzDqcrokBYSIpJ1VW6r55pNLOXFUX77+qTFRl9NlKSBE\nJK3sqtvL9Y8spCAvm19ccQJZmdpMRUX7IEQkbbg7X5+7hI3ba3ns305iQEHewReSpFE0i0jaePCN\ndbywfAu3TjuOE0fpumRRU0CISFp4e912Zr2wimnjBvKvHx8VdTmCAkJE0kDFrjpu+P1Chvftzt2X\nTtTJcGlC+yBEJFKbd+7hS48upKaukUeu+SgFedlRlyQhBYSIRMLdmVNSyg+eWUljs3PPzEmMGahB\npNOJAkJEUm7Tzj3c+sQS3nhvGyeN7stdF09ieL/uUZclrSggRCRl3J3H55dyx7MraXbn+xeO48qP\njtCFf9KUAkJEUqJsRy3ffHIpb7y3jZNH9+OuSyYyrK9aDelMASEiSeXuPPZ2KT98biXuzg+mj+eK\nE4er1dABKCBEJGnKdtRy6xNLeXPNNj52dD9mzVCroSNRQIhIu2tudn7/9kbufG4lAHdcFLQadH5D\nx5LygDCzYcDvgCLAgQfc/V4z6wvMBkYC64GZ7r4j1fWJyJEp3V7LN55Ywt/+WcWpR/dn1sUTGNpH\nrYaOKIoWRCPwVXdfaGYFwAIzewn4PPCyu88ys1uBW4FvRFCfiByG5mbn0bc2cOfzq8gw484ZE7j8\nI8PUaujAUh4Q7l4OlIf3d5nZSmAIcCFwejjbw8CrKCBEOoTS7bXcMm8Jf19bxceP6c+siycypHe3\nqMuSIxTpPggzGwmcALwFFIXhAbCFoAsq3jLXAtcCDB8+PPlFikibmpudR97awKyw1TBrxgQuU6uh\n04gsIMysB/AE8BV3r479Qrm7m5nHW87dHwAeACguLo47j4gk34aq3dwybwlvrdvOaccWMmvGBAar\n1dCpRBIQZpZNEA6PuvuT4eStZjbI3cvNbBBQEUVtInJgzc3O7/6+nh+9sJqsDOOuiydyafFQtRo6\noSiOYjLg18BKd78n5qmngauBWeHtH1Ndm4gc2Pptu7nliSW8vW47nzi2kFkXT2BQL7UaOqsoWhAf\nAz4HLDWzxeG0bxEEwxwzuwbYAMyMoDYRiaO52Xn47+v50QuryM7M4K5LJnLpVLUaOrsojmJ6E2jr\nW3VWKmsRkYNbt20335i3hLfXb+eMMYX8cIZaDV2FzqQWkbiamp2H/raeu18MWg0/vnQSF08ZolZD\nF6KAEJEPWVtZwy3zllCyYQdnHjeAH140gYG98qIuS1JMASEi+zQ1O7/96zrufnE1uVkZ3DNzEhed\noFZDV6WAEBEA/hm2GhZs2MFZxw3ghzMmUNRTrYauTAEh0sU1NTu/eXMdP/7zavKyM/npZZOYPlmt\nBlFAiHRpaypq+Pq8d1i0cSefHFvEDy8azwC1GiSkgBDpgpqanQffWMtPXnqX7jmZ3Hv5ZC6YNFit\nBtmPAkKki1lTsYuvzV3C4tKdnHN8ET+4aDwDCtRqkA9TQIh0EZt27mFuSSm/fPWfajVIQhQQIp1Y\n6fZaXli2hWeXlrO4dCcA08YN5PvTx1NYkBtxdZLuFBAinczGqlqeW1bO80vLeafsfQDGD+nJLdPG\ncN74QYzsnx9xhdJRKCBEOoH123bz3LJynltazrJN1QBMGtqLW889jvPGD2J4P10TWg6dAkKkg1pb\nWcPzy7bw7JJyVpQHoTB5WG++fd5Ypo0fyLC+CgU5MgoIkQ5kTUUNzy0NWgqrtuwCYMrw3tz26bGc\nO2GQrgMt7UoBIZLm3tu6i2eXlvP80i2s3hqEQvGIPnzn/OM5d8JADb0tSaOAEEkz7s67W2t4Nmwp\nrKmowQw+MrIv3/3M8UwbP0gjq0pKKCBE0oC7s7J8F88vK+fZpeWsrdxNhsGJo/py9cnj+NS4gRoC\nQ1JOASESEXdn+eZqnl9WznNLt7BuWxAKJ43uxxc+NopPjRuocxUkUgoIkRRyd5Ztqg72KSwrZ0NV\nLZkZxilH9ePfPj6aT40rol8PhYKkBwWESJK5O0vK3g+OPlpWTun2PWRlGKcc3Z/rP3EU54wbSN/8\nnKjLFPkQBYRIErg7i0p38vzSoPto084gFE49pj83nnEMZx9fRB+FgqQ5BYTIYarb28SmnXvYtGMP\nZTv2ULajdt/thqpaqnY3kJ1pfPyYQm4++1jOHltEr+7ZUZctkjAFhEgb6vY2UbZjD5t2xm78P7hf\nuat+v/mzMozBvbsxtE83zjxuACcf1Y+zxhbRq5tCQTomBYR0WXsamti0M3bDv38QbKvZPwCyM2MC\nYMwAhvQJ7g/t052hfbpR1DOPzAwNnS2dhwJCOq3ahsag+2fnhzf+m3bUsq2mYb/5szONIb2DDf4n\nxw5gaJ9uYQgEATCgQAEgXYsCQjokd6d6TyNbquvY3EYXUNXu/QMgJzNj36/+448v2rfhH9qnG0N6\nd2dAQS4ZCgCRfRQQklbcnZr6RrZW11NRXcfWXXVUVNeztbo+vF8X3K+uo76xeb9lc7IyGNo7+NV/\nzuBe+zb+Q/t0Z1ifbvTvoQAQORQKCEmZ2obGcGNfx9ZdYQDEbPArdgW3tQ1NH1q2R24WA3rmUlSQ\nx5ThvSnqmceAnnkMKMjd1yron68AEGlPCgg5YnV7m6jctf9GPrYF0BIAu+oaP7RsXnYGRT3zKCrI\nY9zgnpx53ACKeuYGAVCQR1HPXAb0zKNHrr6qIqmmvzqJy93ZVd9IVU0DVTX1bKup32/jv7U67PrZ\nVcfO2r0fWj4nMyP4xd8zj2OLenDq0f3DjX4wrWXD3zMvCzP96hdJR2kXEGY2DbgXyAQedPdZEZfU\naTQ0NrN9dwPbauqp2t3Atl31VO2up6qmgW01DfvuV9XUs213Aw2t+vgBMjOMAQXBxn1Ev+6cOKrv\nvo19y4a/qCCP3t2zteEX6eDSKiDMLBO4DzgbKAPmm9nT7r4i2srSU8uRPNt214cb+5Zf+7Eb+yAQ\nttXUUx2niweCnbuFPXLp1yOH/j1yGDOwILifn0v/ghz65efSNz+Hop559MvPUT+/SBeRVgEBnAis\ncfe1AGb2OHAh0OkCwt1xBwea9913mpqdHbV7qaqp37dxb9nwV9U0UBnetgRAY7N/6LXNoE/3HPrl\n59CvRw5jB/ekf34O/Xrk0j8mCPrlB/d75KqbR0Q+LN0CYghQGvO4DPhoe69k1ZZqvvz7RcFGGoKN\nc3h/38Y63jTa2LC3er453Njve53wfnPM+g5VXnZGuHHPZVCvPMYP6bnvcezGvl+PHPp2zyErM6Md\nPzER6YrSLSAOysyuBa4FGD58+GG9Rl5WJscW9cAwzMDMMCAj5r5Z8FyGsf98RjjvB/c/mG7ha+w/\nLe7rQDj9g/tmRoYZfbpn06/ll37YzdM9p8P9V4lIB5duW51NwLCYx0PDafu4+wPAAwDFxcWH8Vsc\nRvbP55dXTj3cGkVEuoR064eYDxxjZqPMLAe4HHg64ppERLqktGpBuHujmX0ZeJHgMNffuPvyiMsS\nEemS0irDYwWTAAAKPUlEQVQgANz9OeC5qOsQEenq0q2LSURE0oQCQkRE4lJAiIhIXAoIERGJSwEh\nIiJxmR/OuA9pwswqgQ1H8BL9gW3tVE57Ul2HRnUdGtV1aDpjXSPcvfBgM3XogDhSZlbi7sVR19Ga\n6jo0quvQqK5D05XrUheTiIjEpYAQEZG4unpAPBB1AW1QXYdGdR0a1XVoumxdXXofhIiItK2rtyBE\nRKQNnTYgzOw3ZlZhZstipr1qZh/a629m681sqZktDm8vTGFND5lZrZkVxEz7mZm5mfUPHzeFtb1j\nZgvN7JQk1dLXzF4ys/fC2z7h9NPDev41Zt7J4bSvxbyPdWGdq8zs9sOoaZiZvWJmK8xsuZndlCZ1\n5ZnZ2+Hnv9zMvhdOvzt8zSVm9pSZ9U5lXeHr9DazeeFrrDSzk2Oe+2qr71FS62rjO9XWZ5RtZg+H\nf28rzeybMcusN7M3Wr324pbXDd/H++G0JWb2f2Y2oI2a2vpOfdfMNoWvsdjMzotZZqKZ/T2cf6mZ\n5bV3XTGv17LdKQmnXRqut9litlVmdraZLQjnX2BmZybj82qt0wYE8BAw7RDmP8PdJwOXAP+VlIra\nrmkNwbW3MbMM4Ez2v1DSHnef7O6TgG8CdyaplluBl939GODl8HGLZcDMmMefBd5ptfzXw89wMnC1\nmY06xJoaga+6+/HAScANZnZ8GtRVD5wZfv6TgWlmdhLwEjDe3ScC7xL836SyLoB7gRfc/ThgErAS\ngg0jcA6wsdX8yazrIT78nWrrM7oUyHX3CcBU4ItmNjJmuYLwPWBmY+Os643wb2IiwXVkbmijpra+\nUwA/DV9jcjiKNGaWBTwCXOfu44DTgb1JqKvFGeH8LWGwDJgBvN5qvm3AZ8LP62rgf1s93951AZ04\nINz9dWB7vOfMLCP8pfSDOE/3BHakuKbHgcvC+6cDfyX4YsfTLvW1UcuFwMPh/YeB6THPbQDyzKzI\nzIxgQ/B8Gy+fF97uPsSayt19YXh/F8HGbkga1OXuXhM+zA7/ubv/2d1b/p/+QXAFxJTVZWa9gNOA\nX4d1Nrj7zvDpnwK3EFwqPVbS6or3nTrAZ+RAfrhB7gY0ANUxi87hg7+JzwKPxVtn+B4KaONv4gDf\nqbacAyxx93fCZarcvam962qLu69099Vxpi9y983hw+VANzPLTXZdnTYgDiALeBR4z91vi5n+Stgk\new24Le6SyfMuUGhB18lnCQIjVreWJj/wIPD9JNVR5O7l4f0tQFGr5+cR/PI7BVhI8Ms61t1mthgo\nAx5394rDLST8NXkC8FY61GVmmeFrVAAvuftbrWb5Ah/e0Ca7rlFAJfBbM1tkZg+aWb4FXaSbWjZy\ncaTs/7GV2M9oHkHwlBO0cn7s7rHh8gTBL2mAzwB/avVaHw9r3Ah8EvjNwVbe6jsFcGPY5fKb8G8P\n4FjAzexFC7pzb2n1Mu1ZlwP/F3YZXXuw+mNcDCx099j/t3b/vKBrBsSvgGXufker6We4+3hgAvAL\nM+uR4rqeJLjE6keBN1o919LFdBzBL77fhb8EksaDw9ta//qcQ7BhaesXSkvXxEDgLDvMfSXhZ/8E\n8BV3j/1VGVld7t4UvsZQ4EQzGx9T77cJWnyPpriuLGAKcL+7n0Cwwf0u8C3gOwdYLiX/j7HifEYn\nAk3AYIKg+6qZjY5ZpArYYWaXE/zqr231ki1dJsOA3wJ3HWT9rb9T9wOjCbrRyoGfhLNmAacCV4a3\nF5nZWUmq69Twcz6XoOvrtAO9h/B9jAN+BHyx1VPt+nm16IoB8TfgjJYdT625+z+BrcDx8Z5PotkE\nLYOX3L25rZnc/e8EY7AcdByVw7DVzAYBhLf7/XJ09y0E/bFnE+wLaKvGGuBVgj+wQ2Jm2QR/yI+6\n+5PpUlfMa+wEXiHsazezzwPnA1d6q2PGU1BXGVAW05qZRxAYo4B3zGw9QaAtNLOBKaxrP218RlcQ\n7DvZG7ZQ/gq0PoBkNnAfbXSXxHiaoKutrfV/6Dvl7lvD0G8G/ocgsCD4TF93923uXktwdcspyajL\n3TeFtxXAUzE1tPU+hobz/b9wO9Vau9QVqysGxK8J/tPnhP2f+wn37o/iyAYBPGTuvgH4NvDLA81n\nZscRXK+7KgllPE2wA4zw9o9x5vkO8I1W/bL7CT/XjwLxvsRtCltFvwZWuvs9aVRXoX1w9E03gg3r\nKjObRtDPf0G4MYknaXWFG/pSMxsTTjqLoOthgLuPdPeRBBu8KeG8Kamr1Wu09RltJDgYAzPLJ9iB\nvKrV4k8R/NJ98SCrObWtGtv6TrX84AhdRLBzmHBdE8yse/j+PwGsSEJd+RYeuRi+/3Niaog3f2/g\nWeBWd/9rG7MdcV2tpd01qduLmT1GsMO3v5mVAfsO13P3e8IdfP9rZleGk18xsyaCHZC3uvvWVNYU\n1vWrNhbtFvYfAhhw9YH+sI+gllkEwXkNQUDObL2cu//tAC97t5ndBuQQ/DJ98gDzxvMx4HPA0pj3\n+600qGsQ8LCZZRL8qJrj7s+Y2RogF3gp7PH7h7tfl8K6AG4EHjWzHGAt8C+JLJSMutr4Tn2T+J/R\nfQT7TpYTfKd/6+5LWtW4i6A7hTg9qi196ga8D/xr6xlCbX2nPmtmkwm6K9cTdtm4+w4zu4fgSB8H\nnnP3Z5NQVxHwVLh8FvB7d3/BzC4Cfk7QQ/CsmS12908BXwaOBr5jZi3dh+fE7h9qp7r2ozOpRUQk\nrq7YxSQiIglQQIiISFwKCBERiUsBISIicSkgREQkLgWEdHn2wWi5yy0YsfWrFgyaeKBlRprZFe20\n/gftgwHkRNKGDnOVLs/Maty9R3h/APB74K/u3uZQ12Z2OvA1dz8/NVWKpJ5aECIxwhOPrgW+bIGR\nZvZGOHBb7LU4ZhGefGRmNx9gvn3Cs2efDVspy8zssnD6q2ZWbGYX2AfXJ1htZuvC56ea2WsWDOr2\nYquzgEWSptOeSS1yuNx9bXjW9ACCcZ/Odvc6MzuGYJybYoJrUuxrQZhZ9zbmizUN2Ozunw6X6dVq\nvU8TDCuCmc0BXgvHEfo5cKG7V4ahcgfByKgiSaWAEDmwbILRfScTjD567BHMtxT4iZn9CHjG3VuP\n2guABUNM73H3+ywYNXY8HwxVkUkw+qhI0ikgRFqxYNjpJoLWw+0Eo/tOIuiSrWtjsZsPNp+7v2tm\nU4DzgB+Y2cvu/p+t1v1JgqG4W0bbNGC5u5+MSIppH4RIDDMrBP4b+EU4NHUvoDwcFvpzBL/gAXYR\nXJmrRVvzxb72YKDW3R8B7qbVMNJmNoJgELtL3X1POHk1wcWkTg7nybbgmgAiSacWhMgHo+VmE1zU\n5n+BlqGhfwk8YWb/D3iBDy6/uQRoMrN3CK7F3NZ8sSYQjJTaTHA9hutbPf95oB/wh7A7abO7n2dm\nlwD/Fe6zyAJ+RnDZSZGk0mGuIiISl7qYREQkLgWEiIjEpYAQEZG4FBAiIhKXAkJEROJSQIiISFwK\nCBERiUsBISIicf1/ZBOXrqE0D1cAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10e8cff50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8 \n",
    "## Perpetual - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512538401_perpetual_encoder_binary8_0_8_67108864'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512534085_perpetual_encoder_binary8_0_8_33554432'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512531773_perpetual_encoder_binary8_0_8_16777216'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512530465_perpetual_encoder_binary8_0_8_8388608'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512529659_perpetual_encoder_binary8_0_8_4194304'\n",
    "twenty_mb = 'results/encoder/thousand/1512529105_perpetual_encoder_binary8_0_8_2621440'\n",
    "ten_mb = 'results/encoder/thousand/1512528650_perpetual_encoder_binary8_0_8_1310720'\n",
    "one_mb = 'results/encoder/thousand/1512528271_perpetual_encoder_binary8_0_8_131072'\n",
    "one_kb = 'results/encoder/thousand/1512527960_perpetual_encoder_binary8_0_8_128'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.legend(['G=8', 'G=8(T)'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/perpetual_gen_8.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wdkIsXxb8RjrAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x113016f90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 16 \n",
    "## Perpetual - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512549406_perpetual_encoder_binary8_0_16_33554432'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512544868_perpetual_encoder_binary8_0_16_16777216'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512542532_perpetual_encoder_binary8_0_16_8388608'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512541214_perpetual_encoder_binary8_0_16_4194304'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512540406_perpetual_encoder_binary8_0_16_2097152'\n",
    "twenty_mb = 'results/encoder/thousand/1512539851_perpetual_encoder_binary8_0_16_1310720'\n",
    "ten_mb = 'results/encoder/thousand/1512539396_perpetual_encoder_binary8_0_16_655360'\n",
    "one_mb = 'results/encoder/thousand/1512539017_perpetual_encoder_binary8_0_16_65536'\n",
    "one_kb = 'results/encoder/thousand/1512538705_perpetual_encoder_binary8_0_16_64'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.legend(['G=16', 'G=16(T)'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/perpetual_gen_16.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "282.708894\n",
      "2147483648\n",
      "[8192, 8388608, 83886080, 167772160, 268435456, 536870912, 1073741824, 2147483648, 4294967296]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSFT/HzauKK9m/fapAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x113224bd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 32 \n",
    "## Perpetual - latency \n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def data_size_in_bytes(data_size, mb=True):\n",
    "    kb = 1024\n",
    "    if mb:\n",
    "        return data_size * kb * kb\n",
    "    else:\n",
    "        return data_size * kb\n",
    "\n",
    "def  calulate_theorectical(gen, data_size, redundancy=0):\n",
    "    symbols = data_size / gen\n",
    "    return symbols * gen * (gen + redundancy)\n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds\n",
    "    return np.average(latency)\n",
    "\n",
    "five_hundre_and_twelve_mb = 'results/encoder/thousand/1512560472_perpetual_encoder_binary8_0_32_16777216'\n",
    "two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512555987_perpetual_encoder_binary8_0_32_8388608'\n",
    "one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512553603_perpetual_encoder_binary8_0_32_4194304'\n",
    "sixty_four_mb = 'results/encoder/thousand/1512552231_perpetual_encoder_binary8_0_32_2097152'\n",
    "thirdy_two_mb = 'results/encoder/thousand/1512551417_perpetual_encoder_binary8_0_32_1048576'\n",
    "twenty_mb = 'results/encoder/thousand/1512550859_perpetual_encoder_binary8_0_32_655360'\n",
    "ten_mb = 'results/encoder/thousand/1512550401_perpetual_encoder_binary8_0_32_327680'\n",
    "one_mb = 'results/encoder/thousand/1512550021_perpetual_encoder_binary8_0_32_32768'\n",
    "one_kb = 'results/encoder/thousand/1512549710_perpetual_encoder_binary8_0_32_32'\n",
    "\n",
    "latencies = {}\n",
    "\n",
    "latencies[0] = get_latency(one_kb)\n",
    "latencies[1] = get_latency(one_mb)\n",
    "latencies[2] = get_latency(ten_mb)\n",
    "latencies[3] = get_latency(twenty_mb)\n",
    "latencies[4] = get_latency(thirdy_two_mb)\n",
    "latencies[5] = get_latency(sixty_four_mb)\n",
    "latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "\n",
    "print(latencies[8])\n",
    "\n",
    "theory = {}\n",
    "\n",
    "print calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[0] = calulate_theorectical(8, data_size_in_bytes(1, False)) \n",
    "theory[1] = calulate_theorectical(8, data_size_in_bytes(1))\n",
    "theory[2] = calulate_theorectical(8, data_size_in_bytes(10))\n",
    "theory[3] = calulate_theorectical(8, data_size_in_bytes(20))\n",
    "theory[4] = calulate_theorectical(8, data_size_in_bytes(32))\n",
    "theory[5] = calulate_theorectical(8, data_size_in_bytes(64))\n",
    "theory[6] = calulate_theorectical(8, data_size_in_bytes(128))\n",
    "theory[7] = calulate_theorectical(8, data_size_in_bytes(256))\n",
    "theory[8] = calulate_theorectical(8, data_size_in_bytes(512))\n",
    "\n",
    "print theory.values()\n",
    "\n",
    "data_size = np.array(latencies.keys())\n",
    "latency = np.array(latencies.values())\n",
    "theory = np.array(theory.values())\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "\n",
    "plt.plot(data_size, latency)\n",
    "#plt.plot(data_size, theory)\n",
    "plt.legend(['G=32', 'G=32(T)'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/perpetual_gen_32.eps')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.91212903226\n",
      "1.27071235951\n",
      "1.25558240778\n",
      "1.29021013085\n",
      "1.31753387673\n",
      "1.39421829787\n",
      "1.51208027594\n",
      "1.604826946\n",
      "1.60925885973\n",
      "t\n",
      "2.64005668399\n",
      "1.66656081542\n",
      "1.51829363177\n",
      "1.48414861656\n",
      "1.4836006879\n",
      "1.46127695356\n",
      "1.45133425873\n",
      "1.62795882318\n",
      "1.76269253482\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x111ba4990>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8,16,32 \n",
    "## Actual On The Fly - latency comparison\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512538401_perpetual_encoder_binary8_0_8_67108864'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512534085_perpetual_encoder_binary8_0_8_33554432'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512531773_perpetual_encoder_binary8_0_8_16777216'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512530465_perpetual_encoder_binary8_0_8_8388608'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512529659_perpetual_encoder_binary8_0_8_4194304'\n",
    "    twenty_mb = 'results/encoder/thousand/1512529105_perpetual_encoder_binary8_0_8_2621440'\n",
    "    ten_mb = 'results/encoder/thousand/1512528650_perpetual_encoder_binary8_0_8_1310720'\n",
    "    one_mb = 'results/encoder/thousand/1512528271_perpetual_encoder_binary8_0_8_131072'\n",
    "    one_kb = 'results/encoder/thousand/1512527960_perpetual_encoder_binary8_0_8_128'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512549406_perpetual_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512544868_perpetual_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512542532_perpetual_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512541214_perpetual_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512540406_perpetual_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512539851_perpetual_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512539396_perpetual_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512539017_perpetual_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512538705_perpetual_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_thirdy_two():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512560472_perpetual_encoder_binary8_0_32_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512555987_perpetual_encoder_binary8_0_32_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512553603_perpetual_encoder_binary8_0_32_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512552231_perpetual_encoder_binary8_0_32_2097152'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512551417_perpetual_encoder_binary8_0_32_1048576'\n",
    "    twenty_mb = 'results/encoder/thousand/1512550859_perpetual_encoder_binary8_0_32_655360'\n",
    "    ten_mb = 'results/encoder/thousand/1512550401_perpetual_encoder_binary8_0_32_327680'\n",
    "    one_mb = 'results/encoder/thousand/1512550021_perpetual_encoder_binary8_0_32_32768'\n",
    "    one_kb = 'results/encoder/thousand/1512549710_perpetual_encoder_binary8_0_32_32'\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def calc_diff(x,y):\n",
    "    print (y/float(x))\n",
    "    \n",
    "def diff_gen(g, gg):\n",
    "    for x, y in zip(g,gg):\n",
    "        calc_diff(g[x],gg[y])\n",
    "\n",
    "eight_latencies = read_gen_eight()\n",
    "eight_data_size = np.array(eight_latencies.keys())\n",
    "eight_latency = np.array(eight_latencies.values())\n",
    "\n",
    "\n",
    "sixteen_latencies = read_gen_sixteen()\n",
    "diff_gen(eight_latencies,sixteen_latencies)\n",
    "sixteen_data_size = np.array(sixteen_latencies.keys())\n",
    "sixteen_latency = np.array(sixteen_latencies.values())\n",
    "\n",
    "print('t')\n",
    "thridy_two_latencies = read_gen_thirdy_two()\n",
    "diff_gen(sixteen_latencies,thridy_two_latencies)\n",
    "thridy_two_data_size = np.array(thridy_two_latencies.keys())\n",
    "thridy_two_latency = np.array(thridy_two_latencies.values())\n",
    "\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(data_size, my_xticks)\n",
    "plt.plot(eight_data_size, eight_latency)\n",
    "\n",
    "plt.xticks(sixteen_data_size, my_xticks)\n",
    "plt.plot(sixteen_data_size, sixteen_latency)\n",
    "\n",
    "plt.xticks(thridy_two_data_size, my_xticks)\n",
    "plt.plot(thridy_two_data_size, thridy_two_latency)\n",
    "\n",
    "plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['G=8', 'G=16', 'G=32'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/perpetual_gen_comparison.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n7AQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x106177550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Generation Size: 8,16,32 \n",
    "## Actual On The Fly - latency comparison\n",
    "import pandas as pd\n",
    "import numpy as np \n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "def read_gen_eight():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512538401_perpetual_encoder_binary8_0_8_67108864'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512534085_perpetual_encoder_binary8_0_8_33554432'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512531773_perpetual_encoder_binary8_0_8_16777216'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512530465_perpetual_encoder_binary8_0_8_8388608'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512529659_perpetual_encoder_binary8_0_8_4194304'\n",
    "    twenty_mb = 'results/encoder/thousand/1512529105_perpetual_encoder_binary8_0_8_2621440'\n",
    "    ten_mb = 'results/encoder/thousand/1512528650_perpetual_encoder_binary8_0_8_1310720'\n",
    "    one_mb = 'results/encoder/thousand/1512528271_perpetual_encoder_binary8_0_8_131072'\n",
    "    one_kb = 'results/encoder/thousand/1512527960_perpetual_encoder_binary8_0_8_128'\n",
    "\n",
    "\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_sixteen():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512549406_perpetual_encoder_binary8_0_16_33554432'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512544868_perpetual_encoder_binary8_0_16_16777216'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512542532_perpetual_encoder_binary8_0_16_8388608'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512541214_perpetual_encoder_binary8_0_16_4194304'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512540406_perpetual_encoder_binary8_0_16_2097152'\n",
    "    twenty_mb = 'results/encoder/thousand/1512539851_perpetual_encoder_binary8_0_16_1310720'\n",
    "    ten_mb = 'results/encoder/thousand/1512539396_perpetual_encoder_binary8_0_16_655360'\n",
    "    one_mb = 'results/encoder/thousand/1512539017_perpetual_encoder_binary8_0_16_65536'\n",
    "    one_kb = 'results/encoder/thousand/1512538705_perpetual_encoder_binary8_0_16_64'\n",
    "    \n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "def read_gen_thirdy_two():\n",
    "    five_hundre_and_twelve_mb = 'results/encoder/thousand/1512560472_perpetual_encoder_binary8_0_32_16777216'\n",
    "    two_hundre_and_fifty_six_mb = 'results/encoder/thousand/1512555987_perpetual_encoder_binary8_0_32_8388608'\n",
    "    one_hundre_and_twenty_eight_mb = 'results/encoder/thousand/1512553603_perpetual_encoder_binary8_0_32_4194304'\n",
    "    sixty_four_mb = 'results/encoder/thousand/1512552231_perpetual_encoder_binary8_0_32_2097152'\n",
    "    thirdy_two_mb = 'results/encoder/thousand/1512551417_perpetual_encoder_binary8_0_32_1048576'\n",
    "    twenty_mb = 'results/encoder/thousand/1512550859_perpetual_encoder_binary8_0_32_655360'\n",
    "    ten_mb = 'results/encoder/thousand/1512550401_perpetual_encoder_binary8_0_32_327680'\n",
    "    one_mb = 'results/encoder/thousand/1512550021_perpetual_encoder_binary8_0_32_32768'\n",
    "    one_kb = 'results/encoder/thousand/1512549710_perpetual_encoder_binary8_0_32_32'\n",
    "    latencies = {}\n",
    "\n",
    "    latencies[0] = get_latency(one_kb)\n",
    "    latencies[1] = get_latency(one_mb)\n",
    "    latencies[2] = get_latency(ten_mb)\n",
    "    latencies[3] = get_latency(twenty_mb)\n",
    "    latencies[4] = get_latency(thirdy_two_mb)\n",
    "    latencies[5] = get_latency(sixty_four_mb)\n",
    "    latencies[6] = get_latency(one_hundre_and_twenty_eight_mb)\n",
    "    latencies[7] = get_latency(two_hundre_and_fifty_six_mb)\n",
    "    latencies[8] = get_latency(five_hundre_and_twelve_mb)\n",
    "    return latencies\n",
    "\n",
    "eight_latencies = read_gen_eight()\n",
    "eight_data_size = np.array(eight_latencies.keys())\n",
    "eight_latency = np.array(eight_latencies.values())\n",
    "\n",
    "sixteen_latencies = read_gen_sixteen()\n",
    "sixteen_data_size = np.array(sixteen_latencies.keys())\n",
    "sixteen_latency = np.array(sixteen_latencies.values())\n",
    "\n",
    "thridy_two_latencies = read_gen_thirdy_two()\n",
    "thridy_two_data_size = np.array(thridy_two_latencies.keys())\n",
    "thridy_two_latency = np.array(thridy_two_latencies.values())\n",
    "\n",
    "\n",
    "pd.DataFrame\n",
    "fig = plt.figure()\n",
    "my_xticks = ['1kB','1MB','10MB','20MB','32MB','64MB','128MB','256MB','512MB']\n",
    "plt.xticks(eight_data_size, my_xticks)\n",
    "plt.plot(eight_data_size, eight_latency)\n",
    "\n",
    "plt.xticks(sixteen_data_size, my_xticks)\n",
    "plt.plot(sixteen_data_size, sixteen_latency)\n",
    "\n",
    "plt.xticks(thridy_two_data_size, my_xticks)\n",
    "plt.plot(thridy_two_data_size, thridy_two_latency)\n",
    "\n",
    "# plt.yscale(\"log\", nonposy=\"clip\")\n",
    "plt.legend(['G=8', 'G=16', 'G=32'])\n",
    "plt.ylabel('Latency in ms')\n",
    "plt.xlabel('Data size')\n",
    "fig.savefig('results/graphs/perpetual_gen_simple_comparison.eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "150.652447552\n"
     ]
    }
   ],
   "source": [
    "def parseData(index, value):\n",
    "    if index == 0 or index == 1 or index == 2 or index == 3 or index == 4:\n",
    "        return int(value)\n",
    "    else:\n",
    "        return float(value)\n",
    "\n",
    "def get_latency(file_name):\n",
    "    file = open(file_name)\n",
    "    data = []\n",
    "    for line in file:\n",
    "        temp = line.split(',')\n",
    "        index = 0\n",
    "        for item in temp: \n",
    "            temp[index] = parseData(index, item)\n",
    "            index = index + 1\n",
    "        data.append(temp)\n",
    "    latency = [entry[6]/1000 for entry in data] # convert to milliseconds \n",
    "    return np.average(latency)\n",
    "\n",
    "one_kb = 'results/encoder/special/1515972020_full_vector_encoder_binary8_0_8_128'\n",
    "one_mb = 'results/encoder/special/1515972332_full_vector_encoder_binary8_0_8_128000'\n",
    "\n",
    "l1 = get_latency(one_kb)\n",
    "l2 = get_latency(one_mb)\n",
    "\n",
    "print l2/l1"
   ]
  },
  {
   "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": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}