nccl-bench-wandb.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Init"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import matplotlib\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import wandb\n",
    "import qgrid\n",
    "\n",
    "import sys, os\n",
    "sys.path.insert(0, os.environ['HOME']+'/git0/aws-network-benchmarks')\n",
    "import parse_nccltest_output"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# NCCL-tests EFA performance "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "import wandb\n",
    "import qgrid\n",
    "api = wandb.Api()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "69\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>busbw_256M</th>\n",
       "      <th>zone</th>\n",
       "      <th>network</th>\n",
       "      <th>ofi_patch_fixed</th>\n",
       "      <th>num_tasks</th>\n",
       "      <th>NCCL_VERSION_TAG</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100.40</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>98.72</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>49.04</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>99.28</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>97.60</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>17.76</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>old</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>17.76</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>old</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>17.76</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>old</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>17.76</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>old</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>17.76</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>old</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>17.76</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>old</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>99.52</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>99.76</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>100.72</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>97.36</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>100.40</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>99.44</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>45.92</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>100.32</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>97.04</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>99.52</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>47.76</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>False</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>45.44</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>False</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>49.04</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>False</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>89.52</td>\n",
       "      <td>us-west-2c</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>101.60</td>\n",
       "      <td>us-west-2a</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>17.84</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>52.00</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>37.36</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>17.68</td>\n",
       "      <td>us-east-1b</td>\n",
       "      <td>efa</td>\n",
       "      <td>True</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>48.32</td>\n",
       "      <td>us-west-2a</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>8</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>49.04</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>49.04</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>45.92</td>\n",
       "      <td>us-east-1c</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>38.08</td>\n",
       "      <td>us-west-2c</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>38.64</td>\n",
       "      <td>us-west-2c</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>43.84</td>\n",
       "      <td>us-west-2a</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>43.76</td>\n",
       "      <td>us-west-2c</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>21.84</td>\n",
       "      <td>us-east-1d</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>45.20</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>49.84</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>18.88</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>old</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>18.88</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>old</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>25.12</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>efa</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>25.12</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>undef</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>25.12</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>undef</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>25.12</td>\n",
       "      <td>us-west-2b</td>\n",
       "      <td>undef</td>\n",
       "      <td>undef</td>\n",
       "      <td>2</td>\n",
       "      <td>2.4.6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    busbw_256M        zone network ofi_patch_fixed  num_tasks NCCL_VERSION_TAG\n",
       "0       100.40  us-east-1b     efa            True          2            2.4.6\n",
       "1        98.72  us-east-1c     efa            True          2            2.4.6\n",
       "2        49.04  us-east-1b     efa            True          2            2.4.6\n",
       "3        99.28  us-east-1c     efa            True          2            2.4.6\n",
       "4        97.60  us-west-2b     efa            True          2            2.4.6\n",
       "5        17.76  us-east-1c     old            True          2            2.4.6\n",
       "6        17.76  us-east-1c     old            True          2            2.4.6\n",
       "7        17.76  us-east-1b     old            True          2            2.4.6\n",
       "8        17.76  us-east-1b     old            True          2            2.4.6\n",
       "9        17.76  us-east-1c     old            True          2            2.4.6\n",
       "10       17.76  us-east-1b     old            True          2            2.4.6\n",
       "11       99.52  us-east-1c     efa            True          2            2.4.6\n",
       "12       99.76  us-east-1b     efa            True          2            2.4.6\n",
       "13      100.72  us-east-1c     efa            True          2            2.4.6\n",
       "14       97.36  us-west-2b     efa            True          2            2.4.6\n",
       "15      100.40  us-east-1c     efa            True          2            2.4.6\n",
       "16       99.44  us-east-1c     efa            True          2            2.4.6\n",
       "17       45.92  us-east-1b     efa            True          2            2.4.6\n",
       "18      100.32  us-east-1b     efa            True          2            2.4.6\n",
       "19       97.04  us-west-2b     efa            True          2            2.4.6\n",
       "20       99.52  us-east-1b     efa            True          2            2.4.6\n",
       "21       47.76  us-east-1c     efa           False          2            2.4.6\n",
       "22       45.44  us-east-1b     efa           False          2            2.4.6\n",
       "23       49.04  us-west-2b     efa           False          2            2.4.6\n",
       "24       89.52  us-west-2c     efa            True          2            2.4.6\n",
       "25      101.60  us-west-2a     efa            True          2            2.4.6\n",
       "26       17.84  us-east-1b     efa            True          2            2.4.6\n",
       "27       52.00  us-east-1c     efa            True          2            2.4.6\n",
       "28       37.36  us-west-2b     efa            True          2            2.4.6\n",
       "29       17.68  us-east-1b     efa            True          2            2.4.6\n",
       "30       48.32  us-west-2a     efa           undef          8            2.4.6\n",
       "31       49.04  us-east-1c     efa           undef          2            2.4.6\n",
       "32       49.04  us-west-2b     efa           undef          2            2.4.6\n",
       "33       45.92  us-east-1c     efa           undef          2            2.4.6\n",
       "34       38.08  us-west-2c     efa           undef          2            2.4.6\n",
       "35       38.64  us-west-2c     efa           undef          2            2.4.6\n",
       "36       43.84  us-west-2a     efa           undef          2            2.4.6\n",
       "37       43.76  us-west-2c     efa           undef          2            2.4.6\n",
       "38       21.84  us-east-1d     efa           undef          2            2.4.6\n",
       "39       45.20  us-west-2b     efa           undef          2            2.4.6\n",
       "40       49.84  us-west-2b     efa           undef          2            2.4.6\n",
       "41       18.88  us-west-2b     old           undef          2            2.4.6\n",
       "42       18.88  us-west-2b     old           undef          2            2.4.6\n",
       "43       25.12  us-west-2b     efa           undef          2            2.4.6\n",
       "44       25.12  us-west-2b   undef           undef          2            2.4.6\n",
       "45       25.12  us-west-2b   undef           undef          2            2.4.6\n",
       "46       25.12  us-west-2b   undef           undef          2            2.4.6"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "keys = ['zone', 'network', 'ofi_patch_fixed', 'num_tasks', 'NCCL_VERSION_TAG']\n",
    "\n",
    "runs = api.runs(\"yaroslavvb/nccl_bench\")\n",
    "stat='busbw_256M'\n",
    "def stats():\n",
    "    print(f\"Summarizing {len(runs)} runs\")\n",
    "    cols = {}\n",
    "    \n",
    "    for i in range(len(runs)):\n",
    "        run = runs[i]\n",
    "        hist = run.history()\n",
    "        if stat in hist:\n",
    "            series = hist[stat]\n",
    "            val = series[series>0][1]\n",
    "            cols.setdefault(stat, []).append(val)\n",
    "            for k in keys:\n",
    "                val = run.config.get(k, 'undef')\n",
    "                try:\n",
    "                    if 'value' in val:\n",
    "                        val = val['value']\n",
    "                except:\n",
    "                    pass\n",
    "                cols.setdefault(k, []).append(val)\n",
    "    return cols\n",
    "cols = stats()\n",
    "df = pd.DataFrame(cols)\n",
    "df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dfcaef7a3ce24af9a8f2204fb769d8d7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "QgridWidget(grid_options={'fullWidthRows': True, 'syncColumnCellResize': True, 'forceFitColumns': True, 'defau…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "qgrid_widget=qgrid.show_grid(df, show_toolbar=True)\n",
    "qgrid_widget"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "        Notebook configured with <a href=\"https://wandb.com\" target=\"_blank\">W&B</a>. You can <a href=\"https://app.wandb.ai/yaroslavvb/testtest/runs/ye9ffodz\" target=\"_blank\">open</a> the run page, or call <code>%%wandb</code>\n",
       "        in a cell containing your training loop to display live results.  Learn more in our <a href=\"https://docs.wandb.com/docs/integrations/jupyter.html\" target=\"_blank\">docs</a>.\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "W&B Run: https://app.wandb.ai/yaroslavvb/testtest/runs/ye9ffodz"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import wandb\n",
    "wandb.init(project='testtest')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Ring vs Tree cutoff"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "metadata": {},
   "outputs": [],
   "source": [
    "import importlib\n",
    "importlib.reload(parse_nccltest_output)\n",
    "\n",
    "def getvals(name):\n",
    "    x = list(results[name].duration.keys())\n",
    "    y = list(results[name].duration.values())\n",
    "    return x, y\n",
    "\n",
    "def getvals2(name):\n",
    "    \"\"\"Returns log(size), bus bandwidth\"\"\"\n",
    "    x = np.array(list(results[name].duration.keys()))\n",
    "    y = np.array(list(results[name].duration.values()))\n",
    "    return np.log(x)/np.log(2), 2*8/1000*x/y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_output(run, fn='output'):\n",
    "    \"\"\"Download output file from given run\"\"\"\n",
    "    if len(run.files()) == 0:\n",
    "        return\n",
    "    for f in run.files():\n",
    "        if f.name == fn:\n",
    "            local_f = f.download(root='/tmp', replace=True)\n",
    "            return local_f.name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "found 21 runs\n",
      "found ring-efa-32\n",
      "found tree-efa-32\n",
      "found tree-efa-2\n",
      "found ring-efa-24\n",
      "found tree-efa-24\n",
      "found tree-efa-8\n",
      "found ring-efa-8\n",
      "found ring-efa-16\n",
      "found tree-efa-16\n",
      "found tree-sockets-8\n",
      "found ring-efa-2\n"
     ]
    }
   ],
   "source": [
    "results = {}\n",
    "urls = {}\n",
    "runs = api.runs(\"yaroslavvb/nccl_bench\")\n",
    "api = wandb.Api()\n",
    "\n",
    "i = -1\n",
    "runs = api.runs(\"yaroslavvb/nccl_bench\", {'description': 'nccl_bench2'})\n",
    "print(f'found {len(runs)} runs')\n",
    "\n",
    "for run in runs:\n",
    "    i+=1\n",
    "    nccltest_output = get_output(run)\n",
    "    if nccltest_output:\n",
    "        perf = parse_nccltest_output.parse(nccltest_output)\n",
    "        key = f\"{run.config['aggregation']}-{run.config['network']}-{run.config['num_tasks']}\"\n",
    "        if key in results:\n",
    "            continue  # don't store duplicate results\n",
    "        results[key] = perf\n",
    "        urls[key] = run.url\n",
    "        print(f'found {key}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot(num_machines):\n",
    "    fig = plt.figure()\n",
    "    matplotlib.rc('figure', figsize=[10,5])\n",
    "    ax = plt.subplot(111)\n",
    "    plt.plot(*getvals2(f'tree-efa-{num_machines}'), label='tree')\n",
    "    plt.plot(*getvals2(f'ring-efa-{num_machines}'), label='ring')\n",
    "    plt.xlabel(\"log(data)\")\n",
    "    plt.ylabel(\"Gbps\")\n",
    "    ax.legend()\n",
    "    plt.title(f'{num_machines}-machine EFA')\n",
    "    plt.show()\n",
    "    print(f'{num_machines}-machine tree: ', urls[f'tree-efa-{num_machines}']+'/logs')\n",
    "    print(f'{num_machines}-machine ring: ', urls[f'ring-efa-{num_machines}']+'/logs')          "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2-machine tree:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/xey7bzxc/logs\n",
      "2-machine ring:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/8ghilchv/logs\n"
     ]
    }
   ],
   "source": [
    "plot(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8-machine tree:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/4gxbcaiy/logs\n",
      "8-machine ring:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/6ntnyd3b/logs\n"
     ]
    }
   ],
   "source": [
    "plot(8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "16-machine tree:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/y6ica2li/logs\n",
      "16-machine ring:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/q5rgn137/logs\n"
     ]
    }
   ],
   "source": [
    "plot(16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24-machine tree:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/wwj2s6ng/logs\n",
      "24-machine ring:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/j31hhw08/logs\n"
     ]
    }
   ],
   "source": [
    "plot(24)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32-machine tree:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/sg4cboyq/logs\n",
      "32-machine ring:  https://app.wandb.ai/yaroslavvb/nccl_bench/runs/3xq522ok/logs\n"
     ]
    }
   ],
   "source": [
    "plot(32)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Scratch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2.0"
      ]
     },
     "execution_count": 174,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.log(4)/np.log(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'https://app.wandb.ai/yaroslavvb/nccl_bench/runs/vswufwtx'"
      ]
     },
     "execution_count": 209,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "run.url"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'bench-16-efa-stock\\nbench-16-efa-stock\\n'"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "run.description"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "runs = api.runs(\"username/project\", {\"$or\": [{\"config.experiment_name\": \"foo\"}, {\"config.experiment_name\": \"bar\"}]})\n",
    "print(\"Found %i\" % len(runs))\n",
    "Calling api.runs(."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "found 9 runs\n"
     ]
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Summarizing 196 runs\n"
     ]
    },
    {
     "ename": "IndexError",
     "evalue": "index out of bounds",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/pandas/core/indexes/base.py\u001b[0m in \u001b[0;36mget_value\u001b[0;34m(self, series, key)\u001b[0m\n\u001b[1;32m   4374\u001b[0m             return self._engine.get_value(s, k,\n\u001b[0;32m-> 4375\u001b[0;31m                                           tz=getattr(series.dtype, 'tz', None))\n\u001b[0m\u001b[1;32m   4376\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mKeyError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_value\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_value\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_loc\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/hashtable_class_helper.pxi\u001b[0m in \u001b[0;36mpandas._libs.hashtable.Int64HashTable.get_item\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/hashtable_class_helper.pxi\u001b[0m in \u001b[0;36mpandas._libs.hashtable.Int64HashTable.get_item\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;31mKeyError\u001b[0m: 1",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mIndexError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-137-25365cd99ce4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     23\u001b[0m                 \u001b[0mcols\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msetdefault\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     24\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mcols\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 25\u001b[0;31m \u001b[0mcols\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstats\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     26\u001b[0m \u001b[0mdf\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mDataFrame\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcols\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     27\u001b[0m \u001b[0mdf\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-137-25365cd99ce4>\u001b[0m in \u001b[0;36mstats\u001b[0;34m()\u001b[0m\n\u001b[1;32m     12\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mstat\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mhist\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m             \u001b[0mseries\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhist\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mstat\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m             \u001b[0mval\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mseries\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mseries\u001b[0m\u001b[0;34m>\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     15\u001b[0m             \u001b[0mcols\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msetdefault\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstat\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mkeys\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/pandas/core/series.py\u001b[0m in \u001b[0;36m__getitem__\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m    866\u001b[0m         \u001b[0mkey\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mapply_if_callable\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    867\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 868\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_value\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    869\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    870\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mis_scalar\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mresult\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/pandas/core/indexes/base.py\u001b[0m in \u001b[0;36mget_value\u001b[0;34m(self, series, key)\u001b[0m\n\u001b[1;32m   4379\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   4380\u001b[0m             \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 4381\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mlibindex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_value_box\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   4382\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mIndexError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   4383\u001b[0m                 \u001b[0;32mraise\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.get_value_box\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.get_value_at\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/util.pxd\u001b[0m in \u001b[0;36mpandas._libs.util.get_value_at\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/util.pxd\u001b[0m in \u001b[0;36mpandas._libs.util.validate_indexer\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;31mIndexError\u001b[0m: index out of bounds"
     ]
    }
   ],
   "source": [
    "keys = ['zone', 'network', 'ofi_patch_fixed', 'num_tasks', 'NCCL_VERSION_TAG']\n",
    "\n",
    "runs = api.runs(\"yaroslavvb/nccl_bench\")\n",
    "stat='busbw_256M'\n",
    "def stats():\n",
    "    print(f\"Summarizing {len(runs)} runs\")\n",
    "    cols = {}\n",
    "    \n",
    "    for i in range(len(runs)):\n",
    "        run = runs[i]\n",
    "        hist = run.history()\n",
    "        if stat in hist:\n",
    "            series = hist[stat]\n",
    "            val = series[series>0][1]\n",
    "            cols.setdefault(stat, []).append(val)\n",
    "            for k in keys:\n",
    "                val = run.config.get(k, 'undef')\n",
    "                try:\n",
    "                    if 'value' in val:\n",
    "                        val = val['value']\n",
    "                except:\n",
    "                    pass\n",
    "                cols.setdefault(k, []).append(val)\n",
    "    return cols\n",
    "cols = stats()\n",
    "df = pd.DataFrame(cols)\n",
    "df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x120345048>,\n",
       " <matplotlib.lines.Line2D at 0x120345198>]"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEJCAYAAABohnsfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3Xd8FHX+x/HXNwm9hRogCYTekRKKHUSqSLGAeip63nm/n73d6XmWs5566ql3nieWn3LeSVFBEJCiIFiQhC6IEJqbAEkgEBJC2ub7+2PGM3pAlpBkdjfv5+ORB7vfnd35DIG8MzPf+Yyx1iIiIhKICK8LEBGR0KHQEBGRgCk0REQkYAoNEREJmEJDREQCptAQEZGAKTRERCRgCg0REQmYQkNERAIW5XUBFa1Zs2Y2ISHB6zJERELKmjVrDlhrm5e1XNiFRkJCAsnJyV6XISISUowxewJZToenREQkYAoNEREJmEJDREQCptAQEZGAKTRERCRgCg0REQmYQkNERAKm0BARCXV5WbDwXsjPrvRVKTREREKVtbBxFvxtACS9Dnu+qvRVht0V4SIi1cLh7+GjuyBlCcQmwriXIKZHpa9WoSEiEkpK/LD6NfjkUef5qKdh4K8hIrJKVq/QEBEJFelbYO6tkJYMHYfD2Ochuk2VlqDQEBEJdsUFsOJZ+PwvULshXPI69LoMjKnyUhQaIiLBbM9XMO82OLANel8BI5+Eek09K6fM2VPGmHhjzDJjzBZjzGZjzO3ueBNjzBJjzHb3z8buuDHGvGSMSTHGbDTG9Cv1WVPc5bcbY6aUGu9vjNnkvuclY5z4PNE6RETCXv4R50T3/42Cony4+n245FVPAwMCm3JbDNxtre0ODAZuNsZ0B+4DPrHWdgI+cZ8DjAY6uV83Aq+AEwDAw8AgYCDwcKkQeAX4dan3jXLHT7QOEZHwtXUBvDwI1vwfDL4ZbvoKOl7odVVAAKFhrd1nrV3rPs4BvgVigfHA2+5ibwMT3MfjgWnWsQqINsa0AkYCS6y1WdbaQ8ASYJT7WkNr7SprrQWm/eyzjrcOEZHwk5MOM6fA9CuhTmO4YSmMehJq1fe6sv84pXMaxpgEoC/wNRBjrd3nvrQfiHEfxwK+Um9LdcdONp56nHFOso6f13Ujzl4NbdpU7UwCEZHTZi2sewcW/8E5FHXBg3D27RBZw+vK/kvAoWGMqQ+8D9xhrT1iSp21t9ZaY4ythPoCWoe1diowFSAxMbFS6xARqVAHd8BHd8CuFdD2bLj4RWjWyeuqTiig0DDG1MAJjH9Zaz9wh9ONMa2stfvcQ0wZ7ngaEF/q7XHuWBow5Gfjy93xuOMsf7J1iIiENn8xfPU3WP4niKwJY1+AflMgIri7OwUye8oAbwDfWmufL/XSXOCHGVBTgA9LjV/rzqIaDGS7h5gWASOMMY3dE+AjgEXua0eMMYPddV37s8863jpERELX3vXw2lBY+rBzgvvm1ZB4fdAHBgS2p3E2cA2wyRiz3h27H3gKmGmMuQHYA0xyX1sAjAFSgDzgegBrbZYx5jEgyV3uUWttlvv4JuAtoA6w0P3iJOsQEQk9hXnOnsVXL0O9ZjDpn9B9nNdVnRLjTFgKH4mJiTY5OdnrMkREfmrncph3Oxza7RyGGv4o1In2uqr/MMassdYmlrWcrggXEalMeVmw+EFY/w406QBTPoJ253pdVbkpNEREKoO1sHk2LPydExzn3AXn/w5q1PG6stOi0BARqWjZaTD/bti2EFr3hWtmQ8teXldVIRQaIiIVpaQEkt+ApY+A9cOIJ2DQ/0Bk+PyoDZ8tERHxUuZ3zr0ufF9D+6Fw8QvQOMHrqiqcQkNE5HQUFzr3uVj5LNSsBxP+AWdc4cm9LqqCQkNEpLx8q2HubZD5LfS8DEY9BfWbe11VpVJoiIicqoIc+OQxWD0VGsbCVTOh80ivq6oSCg0RkVOxbZFzc6QjaTDwRhj2INRq4HVVVUahISISiNxM+Pg++OY9aN4VblgM8QO9rqrKKTRERE7GWtgwHRb9HgqPwpD74Zw7Iaqm15V5QqEhInIih3bDvDtg5zKIHwzjXoLmXbyuylMKDRGRn/MXw9f/gGVPgImEi56D/r8MidbllU2hISJS2v5NzkV6e9dB59Fw0bPQKK7s91UTCg0REYCiY/DZM/DFi1C3CVz+FnSfELYX6ZWXQkNEZPfnzkV6WTug79Uw/DEnOOS/KDREpPo6dhiWPARr33b6RF37IbQf4nFRwU2hISLV05a5sOAeOHoAzr4dzr8Patb1uqqgp9AQkerlyD4nLLZ+BC17Oy1AWvfxuqqQodAQkeqhpMQ5DLXkIfAXOvfoHnxzWN3roirob0tEwt+B7TDvdtjzBbQ7D8a+AE07eF1VSFJoiEj48hfBFy/AZ3+GGrVh/MvQ5xeaRnsaFBoiEp5S1zgX6WVshh4TYdTT0CDG66pCnkJDRMJLQa7T/uPrf0D9lnDFu9B1jNdVhQ2FhoiEj5SlMO9OyP4eBvwKhj0MtRt6XVVYUWiISOg7ehAW3Q8bp0OzznD9x9D2TK+rCksKDREJXdbCplnOzZHyj8D598K5d0NULa8rC1sKDREJTYe/h4/udA5JxSbCuL9CTHevqwp7Cg0RCS0lflg9FT55zHk++hnn/EVEpLd1VRMKDREJHelbnGm0acnQcTiM/QtEx3tdVbWi0BCR4FeUDyufg8+fh9qN4NI3oOelukjPAwoNEQlue7507nVxcDuccSWMeALqNfW6qmpLoSEiwSk/G5b+EZLfhOg2cPUH0HGY11VVewoNEQk+W+fD/Hsgdz+ceQsMvR9q1vO6KkGhISLBJCcdFv4WtnwIMT3hincgtr/XVUkpCg0R8Z61sO6fsPgB56T3sIfgrNsgsobXlcnPKDRExFsHdzj3uti9EtqeAxe/CM06el2VnIBCQ0S84S+Cr/4Gy5+CyFpOWPS9FiIivK5MTqLM744x5k1jTIYx5ptSY380xqQZY9a7X2NKvfZ7Y0yKMeY7Y8zIUuOj3LEUY8x9pcbbGWO+dsdnGGNquuO13Ocp7usJFbXRIuKxvevgtaHO7KhOw+Hmr6H/dQqMEBDId+gtYNRxxv9ire3jfi0AMMZ0B64Aerjv+bsxJtIYEwm8DIwGugNXussCPO1+VkfgEHCDO34DcMgd/4u7nIiEssI857zFaxdAbiZM+idMfgcatvK6MglQmaFhrV0BZAX4eeOB6dbaAmvtLiAFGOh+pVhrd1prC4HpwHhjjAEuAN5z3/82MKHUZ73tPn4PGOYuLyKhaMcyeOVM+PKv0O9aZ++i+zivq5JTdDr7grcYYza6h68au2OxgK/UMqnu2InGmwKHrbXFPxv/yWe5r2e7y/8XY8yNxphkY0xyZmbmaWySiFS4vCyYcxP8cwKYSLhuvnP+ok6015VJOZQ3NF4BOgB9gH3AcxVWUTlYa6daaxOttYnNmzf3shQR+YG18M378PJA2DjDuc/F/34JCed4XZmchnLNnrLWpv/w2BjzGvCR+zQNKN1yMs4d4wTjB4FoY0yUuzdRevkfPivVGBMFNHKXF5Fgl50K8++GbR9D635wzRxo2dPrqqQClGtPwxhT+qzVROCHmVVzgSvcmU/tgE7AaiAJ6OTOlKqJc7J8rrXWAsuAy9z3TwE+LPVZU9zHlwGfusuLSLAqKYHVr8HLg2DXChj5JPxqqQIjjJS5p2GMeRcYAjQzxqQCDwNDjDF9AAvsBn4DYK3dbIyZCWwBioGbrbV+93NuARYBkcCb1trN7iruBaYbYx4H1gFvuONvAP80xqTgnIi/4rS3VkQqT8ZWmHcb+L6GDhc497ponOB1VVLBTLj98p6YmGiTk5O9LkOk+igugM//AiuehVoNYNRT0HuS7nURYowxa6y1iWUtpyvCRaT8fKudO+llboVelzuBUa+Z11VJJVJoiMipK8iBTx51zl80ioNfvOdc2S1hT6EhIqdm2yL46E44shcG/Q9c8ADUqu91VVJFFBoiEpjcTPj4XufaixbdYdI0iCvzELiEGYWGiJyctbDhXVh0PxQehaEPwNm3Q1RNrysTDyg0ROTEsnbBR3fAzuXQ5ky4+CVo3tnrqsRDCg0R+W/+Yvj6Ffj0CYiIgoueg/6/VOtyUWiIyM/s2+hMo923HrqMgTHPQqPYst8n1YJCQ0QcRcfgs6fhi5egblO4/G3oPl4X6clPKDRExOkTNe92yNoJfa+BEY9BncZlv0+qHYWGSHV27BAseQjWToPG7eDaudD+fK+rkiCm0BCpjqyFb+fCgt/C0QNw9h0w5D6oUcfryuQUWWtZ+/1hZiX7uGtEZ1o0qF2p61NoiFQ3R/Y6YbH1I2h1BvxilvOnhJQDuQXMXpvGjGQfKRm51K0ZycieLWnRRaEhIhWhpATWvgVLHgZ/EQx/DAbfBJH6MRAq/CWWFdsymZHkY+m36RSXWPq2iebpS3txUe/W1K9V+d9L/WsRqQ4ObIe5t8H3X0K78+HiF6BJe6+rkgB9fzCPWWt8zEpOZf+RfJrUq8l1ZyUweUA8nWIaVGktCg2RcFZcCF++CJ89AzXqwvi/Q5+rNI02BOQX+Vm0eT8zknx8ueMgEQbO69ychy/uzrBuMdSM8uZCS4WGSLhKXeNcpJexGXpcAqOfhvotvK5KyvBNWjYzk33MWZfGkfxi4hrX4a7hnbmsfxyto72fqKDQEAk3Bbmw7AlY9Qo0bA1XTocuo72uSk4i+1gRc9c7J7W/STtCzagIRvVoyeQB8ZzZvikREcGzZ6jQEAkn25c697rI9sGAX8Gwh6B2Q6+rkuMoKbGs2nWQmUk+Fn6zn4LiErq1asgj43owvk9rousGZxdhhYZIODh6ED6+DzbNhGZd4JcfQ5vBXlclx7E/O5/31viYmZzK91l5NKgdxeWJcUxObEPP2IaYID/fpNAQCWXWwqZZTmDkH4Hz74Nz74KoWl5XJqUU+Uv45NsMZib7WP5dBiUWBrdvwp3DOzGqRyvq1Iz0usSAKTREQtWhPc6hqB2fQNwAGPdXaNHN66qklJSMXGYm+/hgbSoHcgtp0aAW/3N+ByYlxpPQrJ7X5ZWLQkMk1JT44etX4dPHwETA6D/DgBsgInR+Ww1nRwuKmb9pHzOSfKzZc4ioCMMFXVsweUA853duTlRkaN+TRKEhEkrSNzvTaNPWQKcRcNHzEB3vdVXVnrWWdb7DzEzyMW/DXo4W+mnfvB6/H92VS/rF0bxB+BwuVGiIhIKifFjxZ/jiBagdDZe+AT0v1UV6HjuYW8DsdWnMSPKxPSOXOjUiuah3KyYPiCexbeOgP6ldHgoNkWC3+wuYdxscTIEzroKRT0DdJl5XVW35Sywrt2cyM9nHki3pFPktfeKj+dMlvRjbuxUNatfwusRKpdAQCVb52U5zwTX/B9Ft4ZrZ0OECr6uqtnxZecxK9vHemlT2Zjv9n649M4FJifF0aVm1/Z+8pNAQCUbffgQL7oHcdDjzFhh6P9QMzdk2oSy/yM/iLenMSPqeL1IOYgyc26k5D4ztzoUe9n/ykkJDJJjk7HfudfHtXIjpBVf8G2L7eV1VtbNl7xFmJvuYvS6N7GNFxEbX4c4LO3NZYhyxQdD/yUsKDZFgYK1zy9XFD0JxPgx7GM66FSLD+/h4MMk+VsTcDXuZmeRjU1o2NSMjGNmzJZMT4zmrQ3D1f/KSQkPEawd3wLzbYfdKSDgXLn4RmnbwuqpqwVrL17uymJnkY/6mfRQUl9C1ZQMevrg7E/rE0rhecPZ/8pJCQ8Qr/iL48q/w2dMQWQsufgn6XatptFUg/Ug+761JZVayj90H82hQK4rL+scxeUA8vWIbheVU2Yqi0BDxQtpa50566Zug2zgY82do0NLrqsJakb+ET7dmMDPJx/JtmfhLLIPaNeG2YZ0Y3TO0+j95SaEhUpUKj8KyJ2HV36F+DEz+F3Qb63VVYW1HZi4zk3y8vzaNA7kFtGhQixvPa8+kxHjahWj/Jy8pNESqyo5PYd4dcHgP9L8ehj8CtRt5XVVYyissZv7GfcxM9pG0+xCRP/R/SoxnSJfQ7//kJYWGSGXLy4LFD8D6f0HTjnDdAkg42+uqwo61lvW+w8xM9jFvwz5yC4pp36we943uyiX9YmnRoLbXJYYFhYZIZdryIcy/B/IOwrl3w3m/gxr64VWRso4Wuv2fvmdbutP/aUwvp//TgITw7P/kJYWGSGXISXeu6P52LrTsDVe/D616e11V2PCXWD5POcDMJB+Lt+ynyG85Iz6aJyf24uIzwr//k5fKDA1jzJvAWCDDWtvTHWsCzAASgN3AJGvtIeNE+ovAGCAPuM5au9Z9zxTgAfdjH7fWvu2O9wfeAuoAC4DbrbX2ROs47S0WqUzWwvp/w6L7oegYXPhHOPNWiNTvZxXBl5XHrDWpvJfsY292PtF1a3D14LZMHhBP15a6F3pVMNbaky9gzHlALjCtVGg8A2RZa58yxtwHNLbW3muMGQPcihMag4AXrbWD3ABIBhIBC6wB+rtBsxq4DfgaJzRestYuPNE6ytqgxMREm5ycXJ6/C5HTc2gPfHSHc8K7zZnOnfSadfK6qpBXUOxn8eZ0Zib7+DzlAADndGzG5AHxDO8eQ60oTZWtCMaYNdbaxLKWK/PXH2vtCmNMws+GxwND3MdvA8uBe93xadZJolXGmGhjTCt32SXW2iy3uCXAKGPMcqChtXaVOz4NmAAsPMk6RIJLSQkkvQZLH3EuzBvzLCTeABGaoXM6vt13hBlJPuasT+NwntP/6fZhnbisfxxxjet6XV61Vd595hhr7T738X4gxn0cC/hKLZfqjp1sPPU44ydbh0jwyNzm3EnPtwo6XghjX9Cd9E7Dkfwi5q7fy8xkHxtTnf5Pw3vEMDkxnrM7NiNS/Z88d9oHWt3zDyc/xlXJ6zDG3AjcCNCmTZvKLEXE4S+CL150WoDUrAcTX4Xek9UCpBystazelcWMZB8LNu0jv8jp//TQ2O5M7Kv+T8GmvKGRboxpZa3d5x5+ynDH04DSv2bFuWNp/Hio6Yfx5e543HGWP9k6/ou1diowFZxzGuXcJpHA7F0Pc2+B/Zugx0QY/QzUb+F1VSEn40g+761NZVZyKrsOHKV+rSgm9o3jigHx9I5T/6dgVd7QmAtMAZ5y//yw1PgtxpjpOCfCs90f+ouAJ40xjd3lRgC/t9ZmGWOOGGMG45wIvxb4axnrEPFG0TFnz+KLl6BeM7UAKYdifwnLvstkRtL3LPvO6f80MKEJNw/tyJheLalbU7PMgl0gU27fxdlLaGaMSQUexvlBPtMYcwOwB5jkLr4AZ+ZUCs6U2+sB3HB4DEhyl3v0h5PiwE38OOV2ofvFSdYhUvX2fOXsXRxMgb7XwIjHoE7jst8nAOzMzGVmcirvr00lM6eA5g1q8etz2zMpMY72zet7XZ6cgjKn3IYaTbmVClWQ48yKSnoNots47cs7DPW6qpCQV1jMgk37mZnkY/XuLCIjDEO7NGdSYjxDu7aghvo/BZUKm3IrUm1tX+pcd5GdCoNvggse0H26y2CtZWNqNjOSfcxdv5fcgmISmtbld6O6cFm/OFo0VAuVUKfQEPm5vCzniu4N70KzLnDDYogf6HVVQe2Q2/9pZrKPrftzqF0jwun/lBjPwHZNdFI7jCg0RH5grdNgcME9cOyQ01zwvHsgqpbXlQWlkhLLFzsOMD3Jx5LN6RT6S+gd14jHJ/RkXJ/WNFT/p7Ck0BAByNkP8++GrR9Bqz5wzWxo2cvrqoJS2uFjzEr2MSs5lbTDx4iuW4OrBrVh8oB4urVS/6dwp9CQ6s1aWPcOLPoD+Atg+KMw+GY1GPyZgmI/S7akMyPJ6f9krdP/6b7RXRnePYbaNdT/qbrQ/wypvg7thnm3w87l0PZsp8Fg0w5eVxVUvtufw4wkH7PXpXIor4jWjWpz6wWduLx/HPFN1P+pOlJoSPVT4ofVU+GTR8FEwkXPO7dfVYNBAHLyi5i3YR8zkn1s8B2mRqRhRPeWTBoQzznq/1TtKTSkesnY6jQYTF0NnUbA2L9Ao7iy3xfmrLUk7T7EjCSn/9OxIj+dY+rzwEXdmNg3lqb1NRlAHAoNqR78RfD5C7DiGahZHy55DXpdXu0bDGbk5PP+mjRmJfvYeeAo9WpGMqFvayYlxtMnPlpTZeW/KDQk/O1dBx/eAunfQM9LYdTTUL+511V5pthfwvLvMpmR7OPTrRn4SyyJbRvzv0M6cFHvVur/JCelfx0SvoqOwfI/wZd/hfoxcMW70HWM11V5ZteBo8xM9vH+mlQycgpoVr8mvzqnHZcnxtOxhfo/SWAUGhKedn8Oc2+DrB3Qb4ozlbZOtNdVVbljhX4WfrOP6Uk+Vu/KIsLA0C4tmDQgngvU/0nKQaEh4SX/CCx9GJLfhMYJcO1caH++11VVKWstm9KymZHk9H/KKSimbdO6/HZkFy7tF0fLRur/JOWn0JDwsW2x02AwZx+ceQsMvb9aNRg8nOf0f5qR5PR/qhXl9H+alBjPoHZNiNBUWakACg0JfUcPwsf3waaZ0LwrTJoGcWV2eA4LJSWWL3ccZEayj0Xf7KfQX0Kv2EY8NqEn485oTaM66v8kFUuhIaHLWtj8ASz4HeQfhvPvg3PvqhYNBvcePsas5FRmrfGReugYjerU4MqB8UwaEE+P1o28Lk/CmEJDQtORfTD/LvhuAbTuB+PnQkwPr6uqVAXFfpZuyWBGso+V2zOxFs7u2JTfjuzCyB4t1f9JqoRCQ0KLtbB2Gix+EPyFMOJx5wZJEeH7A3Nb+g/9n9LIOlpIq0a1uXVoRy5PjFf/J6lyCg0JHVk7nQaDu1ZAwrlw8Yth22Awt6CYeRv2MiPJx3q3/9Pw7jFMSozn3E7N1f9JPKPQkOBX4odVr8Cnj0NkDRj7gnPtRZg1GLTWsmbPIaYn+Zi/0en/1KmF+j9JcFFoSHBL3wJzb4G0NdB5lNORtlGs11VVqMycAj5Ym8qMZB87M53+T+P7tGbSgHj6qv+TBBmFhgSn4kL4/HlY8SzUbgiXvuH0jQqTH6DF/hI+25bJjCSn/1NxiaV/28Y8c6nT/6leLf3XlOCkf5kSfNLWOA0GM7Y4nWhHPQX1mnldVYXYc9Dp//TemlTSjzj9n355TjsmJcbRsUUDr8sTKZNCQ4JHYR4sewJW/R3qt4Qrp0OX0V5XddpyC4pZuGkf769NZdVOp//TkC4teGRcPMO6qf+ThBaFhgSHXSudmyMd2uXcRW/4I1A7dC9SK/KX8Pn2A3ywLo0lW/aTX1RCmyZ1uWdEZy7rH6/+TxKyFBrirfxsWPIQrHkLGreDKfOg3XleV1Uu1lo2pmYze10a8zbs5eDRQqLr1uCy/nFM7BtHvzY6qS2hT6Eh3vluIXx0F+Tuh7NuhSH3Q83Qu1jNl5XHnHVpzF6fxs7Mo9SMiuDCbi2Y0CeWIV1aUDNKh58kfCg0pOodPQAL74Vv3oMWPeCKdyC2v9dVnZLDeYXM37SP2WvTSN5zCIBB7Zpw47ntGd2rlRoFSthSaEjVsRY2vQcLfwcFOc6exTl3QlRNrysLSEGxn2VbM5i9Lo1lWzMp9JfQsUV9fjuyC+P7tCaucejtJYmcKoWGVI3sNKfB4LaPITYRxv8NWnTzuqoylZRYkvccYva6NOZv3MuR/GKa1a/FNWe2ZWLfWHq0bqjzFFKtKDSkcpWUwNq3YPFDUFIMI5+EQf8T9A0GUzJymbMujTnr00g9dIw6NSIZ2SOGif3iOLtDU6I0TVaqKYWGVJ6DO5wGg7tXOjOiLn4JmrTzuqoTOpBbwLwNe5m9Lo2NqdlEGDi7YzPuGt6ZkT1a6iptERQaUhn8xc4FesuegMiaTlj0uzYoW4AcK/SzeMt+Zq9LY+X2A/hLLD1aN+SBi7ox7ozWtGio6ylESlNoSMVK3+y0ANm7FrqMgYueg4atva7qJ/wllq92HGT2ujQ+/mYfRwv9tG5UmxvPa8/EvrF0jlE7D5ETUWhIxSgugJXPOV+1o+GyN6HHJUG1d7Fl7xHmrE/jw/VppB8poEGtKMb2bs2EvrEMateECN2jQqRMCg05fb4kp3155lboPRlG/gnqNfW6KgD2ZR/jw/V7mbMuja37c4iKMAzp0oKHxsYyrFsL3SJV5BQpNKT8Co/Cp26DwYat4aqZ0Hmk11WRk1/Ewm/2M2ddGl/tPIi10LdNNI+N78FFvVvTpF5oXBciEowUGlI+KUth/t1waDck3gAX/tG574VHivwlrNyeyQdr01iyJZ2C4hLaNq3LbRd0YmLfWBKa1fOsNpFwclqhYYzZDeQAfqDYWptojGkCzAASgN3AJGvtIeNcAfUiMAbIA66z1q51P2cK8ID7sY9ba992x/sDbwF1gAXA7dZaezo1y2nK2gWL/gDfzYcmHeC6+ZBwjielWGvZkJrN7LWpzNu4j6yjhTSuW4NJifFM7Beru96JVIKK2NMYaq09UOr5fcAn1tqnjDH3uc/vBUYDndyvQcArwCA3ZB4GEgELrDHGzLXWHnKX+TXwNU5ojAIWVkDNcqoK8+Dzv8AXL0JEFAx7GM68GaKq/r7V3x/MY876NOasS2PnAadB4PBuMUzoG8v5nZurQaBIJaqMw1PjgSHu47eB5TihMR6Y5u4prDLGRBtjWrnLLrHWZgEYY5YAo4wxy4GG1tpV7vg0YAIKjaplLWz5EBY/ANk+6HkZjHisyqfRHs4r5KON+5iz7scGgYPbN+E357dnVE81CBSpKqcbGhZYbIyxwKvW2qlAjLV2n/v6fiDGfRwL+Eq9N9UdO9l46nHGpapkfOs0F9y1AmJ6wsRXIeHsKlt9QbGfT791GwR+l0GR39KpRX1+N6oL4/vEEhtdp8pqERHH6YbGOdbaNGNMC2CJMWZr6RettdYNlEpljLkRuBGgTZs2lb268JefDcufgq9fhVoNYMyzzt30Iit/3sSPDQJTmb9xH0fyi2neoBZTzkxgghoEinjutH4KWGvT3D8zjDGzgYFAujGmlbV2n3v4KcNdPA2IL/X2OHcsjR8PZ/0wvtwdjzvO8serYyowFSAxMVEnysurpAR3SNAOAAAPVklEQVQ2/BuW/tG550X/KXDBQ1VyzUVKRi6z16UyZ91e0g47DQJH9WzJxL6xnKUGgSJBo9yhYYypB0RYa3PcxyOAR4G5wBTgKffPD923zAVuMcZMxzkRnu0GyyLgSWNMY3e5EcDvrbVZxpgjxpjBOCfCrwX+Wt56pQxpa2DB7yAtGeIGwi9mQeu+lbrKzJwfGwRuSnMaBJ7TqTm/HdmF4d1j1CBQJAidzv/KGGC2e6ggCvi3tfZjY0wSMNMYcwOwB5jkLr8AZ7ptCs6U2+sB3HB4DEhyl3v0h5PiwE38OOV2IToJXvFyM+GTR2DdO1CvOUz4h3NVd0Tl/GafV1jMki3pfLA2jc9TnAaBPWPdBoF9WtOigRoEigQzE26XPSQmJtrk5GSvywh+/mJIeh2WPQlFR517XJx/b6VcoOcvsXy54wCz16Wx6Jv9HC30Extdh/F9WjOxbyyd1CBQxHPGmDXW2sSyltP+f3W0a6UzKypjC7QfCqOfhuZdKnQV1lq27DvCnHVpfLh+Lxk5BTSoHcXFZzgNAgcmqEGgSChSaFQn2anO9RabZ0N0G5j8DnQdW6GdaPdlH2POOqdB4HfpOdSIdBoETuwbywVd1SBQJNQpNKqDonz46q+w8nmwJTDk93D27VCjYq5z+KFB4Oy1aaza5TQI7Ncmmscm9GRsr1Y0VoNAkbCh0Ahn1sK2j+Hj+5zGgt0uhhFPQOO2p/3Rxf4SPtuWyQfr0ljqNghMaFqXO4Z1ZkLf1rRtqgaBIuFIoRGuDqQ4YZGyBJp1gWvmQIehp/2xuQXFzEjy8ebnu0g7fIzGdWsweUA8E/vG0kcNAkXCnkIj3BTkwoo/w1cvQ1RtZ89i0G8g8vR6M2Ucyef/vtzNv1bt4Uh+MQMTmvDg2O4M69aCGrrwTqTaUGiEC2th03uw5EHI2QdnXOXc46JBTFnvPKnt6TlMXbGTD9fvpaikhFE9WnLjee3p26Zx2W8WkbCj0AgH+zc5V3N//yW06gOTpkH8wHJ/nLWWVTuzeG3lTj7dmkHtGhFMHhDPr85tp3MVItWcQiOU5WU5F+clvwF1GsPFL0LfayCifNNai/0lfLx5P1NX7GRjajZN69Xkzgs7c82ZbXWLVBEBFBqhqcQPa6fBJ49C/mEY8CsYer8THOWQV1jMzCQfb3yxC1/WMdo1q8cTE3tyab84XVchIj+h0Ag1vtWw4B7YtwHang2jn4GWPcv1UZk5Bbz95W7+uWoP2ceK6N+2MQ9c1J0Lu8UQqau1ReQ4FBqhIicdlj4MG96FBq3h0jeg56Xlupo7JSOX11fu5IN1aRT5SxjRPYYbz2tP/7ZNKqFwEQknCo1gV1wIq1+F5U+DvwDOuQvOvRtq1T+lj7HWkrT7EFNX7GDptxnUiorg8v5x3HBOO9o3P7XPEpHqS6ERrKyF7Yth0f1wMAU6jYRRf4KmHU7pY/wllkWb9/Pqip1s8B2mcd0a3D6sE9ec2ZZm9WtVUvEiEq4UGsEoY6sTFjs+gaYd4aqZ0HnkKX3EsUI/s9b4eH3lLr7PyqNt07o8NqEnl/WLo05NndwWkfJRaASTvCxY/idIesM5/DTyT87MqKjAp7seyC1gmnty+1BeEX3io/n96K6M6NFSJ7dF5LQpNIKBvwiS33SuuSg4Av2vh6F/OKV7c+/MzOX1z3fx/ppUCopLuLBbDL85vz2JbRurH5SIVBiFhte2L3UORR34DtoPcfYuYroH/Pbk3VlMXbGTJd+mUyMygkv7xfKrc9vTQSe3RaQSKDS8cmC7ExbbF0OT9nDFu9BldEBTaP0lliVb0pm6Ygdrvz9MdN0a3DK0I9eemUDzBjq5LSKVR6FR1Y4dgs+egdVToUZdGPE4DLwRosr+YZ9f5Oe9Nam88fkudh04SnyTOjwyrgeXJ8ZRt6a+lSJS+fSTpqr4i2HtW/DpE07rj37XwtAHoH7zMt+adbSQaV/tZtpXe8g6WsgZcY14+ap+jOwRQ5TakotIFVJoVIUdy5xDURlbIOFc53qLlr3KfFtmTgGvLN/Bv1fvIb+ohGFdW/Dr89ozqF0TndwWEU8oNCrTwR2w+AH4bgE0ToDJ70DXsWWetzicV8irK3by1he7KfSXMLFvLL85rz2dYhpUTd0iIieg0KgM+UdgxTOw6h/OuYoL/wiDbyrzvEVOfhFvfr6b11fuJLewmIt7t+aOCzupzYeIBA2FRkU7tAf+dTkc2AZ9r4YLHizz7nnHCv1M+2o3//hsB4fyihjRPYa7RnSma8uGVVOziEiAFBoVKXUNvDsZ/IUwZS60O++kixcU+5m+2sfflqWQmVPA+Z2bc/eIzvSOi66igkVETo1Co6J8Ow/e/zXUbwHXLYDmnU+4aJG/hA/WpvLSJymkHT7GwHZNePmqfgxsp9bkIhLcFBqny1pY9XdY9AeI7Q9XTj/hNFp/iWXehr28sHQbuw/mcUZ8NE9f2puzOzbVbCgRCQkKjdPhL4aP74Ok16DbOLhkKtSo81+LWeu0J39+yTa2pefStWUDXr82kWHdWigsRCSkKDTKqyAX3r8Btn0MZ90KFz4KET+90M5ay/JtmTy/eBub0rJp37wef7uqL2N6tiJCHWdFJAQpNMrjyD749yRI/wYues5pX/4zX+04yHOLvyN5zyHiGtfh2cvPYEKf1rqCW0RCmkLjVKVvhn9NcnpIXTkDOo/4ycurd2XxwtJtfLnjIDENa/H4hJ5MSoynZpTCQkRCn0LjVKSthWnjoWY9+OVCaHXGf15K2u2ExRcpB2lWvxYPXNSNqwe3pXYN3SVPRMKHQiNQ6VvgnUugTrQzpTY6Hjh+WPxiUFvdUlVEwpJCIxAHd8A/J0BkLbh2LkTHk7w7ixeWbufzlAMKCxGpNhQaZclOdQ5JlRTDdQtIPtKIFz742g2LmgoLEalWFBonk5vhBEZ+NrvGTuexBTl8unWHwkJEqi2FxonsWgFzb6MkZz9/i/0zf/n3YRrUiuLeUV2ZclZb3SlPRKqloP/JZ4wZBbwIRAKvW2ufqtQV5mXB4gdh/Ttk1Yrl5vx7WbujGTeel8BN53ekUd0albp6EZFgFtShYYyJBF4GhgOpQJIxZq61dkulrHDvOkreuRzysnjdTuCFnImM69+e5y/sRKtG/90eRESkugnq0AAGAinW2p0AxpjpwHigwkOjOD+X3H/9kvy8Eq4veJwOvQbz0fDOugGSiEgpwR4asYCv1PNUYFBlrCjqqViigUcaPcrTl1+le1qIiBxHWPS2MMbcaIxJNsYkZ2ZmluszUtpMBuDB229TYIiInECwh0YaEF/qeZw79hPW2qnW2kRrbWLz5se/l0VZOv5yKvwxW91nRUROIthDIwnoZIxpZ4ypCVwBzPW4JhGRaiuoz2lYa4uNMbcAi3Cm3L5prd3scVkiItVWUIcGgLV2AbDA6zpERCT4D0+JiEgQUWiIiEjAFBoiIhIwhYaIiARMoSEiIgEz1lqva6hQxphMYE85394MOFCB5QQjbWPoC/ftA22jF9paa8u8OjrsQuN0GGOSrbWJXtdRmbSNoS/ctw+0jcFMh6dERCRgCg0REQmYQuOnpnpdQBXQNoa+cN8+0DYGLZ3TEBGRgGlPQ0REAlYtQ8MYM8oY850xJsUYc99xXq9ljJnhvv61MSah6qs8PQFs413GmC3GmI3GmE+MMW29qLO8ytq+UstdaoyxxpiQm6USyDYaYya538fNxph/V3WNpyuAf6dtjDHLjDHr3H+rY7yos7yMMW8aYzKMMd+c4HVjjHnJ3f6Nxph+VV3jKbPWVqsvnBbrO4D2QE1gA9D9Z8vcBPzDfXwFMMPruithG4cCdd3H/xtK2xjI9rnLNQBWAKuARK/rroTvYSdgHdDYfd7C67orYRunAv/rPu4O7Pa67lPcxvOAfsA3J3h9DLAQMMBg4Guvay7rqzruaQwEUqy1O621hcB0YPzPlhkPvO0+fg8YZowJpVv6lbmN1tpl1to89+kqnLsihopAvocAjwFPA/lVWVwFCWQbfw28bK09BGCtzajiGk9XINtogYbu40bA3iqs77RZa1cAWSdZZDwwzTpWAdHGmFZVU135VMfQiAV8pZ6numPHXcZaWwxkA02rpLqKEcg2lnYDzm87oaLM7XN38+OttfOrsrAKFMj3sDPQ2RjzhTFmlTFmVJVVVzEC2cY/AlcbY1Jx7qtza9WUVmVO9f+q54L+JkxSuYwxVwOJwPle11JRjDERwPPAdR6XUtmicA5RDcHZU1xhjOllrT3saVUV60rgLWvtc8aYM4F/GmN6WmtLvC6suqqOexppQHyp53Hu2HGXMcZE4ewWH6yS6ipGINuIMeZC4A/AOGttQRXVVhHK2r4GQE9guTFmN86x4rkhdjI8kO9hKjDXWltkrd0FbMMJkVARyDbeAMwEsNZ+BdTG6dkULgL6vxpMqmNoJAGdjDHtjDE1cU50z/3ZMnOBKe7jy4BPrXvWKkSUuY3GmL7AqziBEWrHwk+6fdbabGttM2ttgrU2AeeczThrbbI35ZZLIP9O5+DsZWCMaYZzuGpnVRZ5mgLZxu+BYQDGmG44oZFZpVVWrrnAte4sqsFAtrV2n9dFnUy1OzxlrS02xtwCLMKZvfGmtXazMeZRINlaOxd4A2c3OAXnJNYV3lV86gLcxj8D9YFZ7jn+76214zwr+hQEuH0hLcBtXASMMMZsAfzAb621IbNHHOA23g28Zoy5E+ek+HWh9AucMeZdnGBv5p6XeRioAWCt/QfOeZoxQAqQB1zvTaWB0xXhIiISsOp4eEpERMpJoSEiIgFTaIiISMAUGiIiEjCFhohICCurKeLPlm3rNijdaIxZbow55fZBCg0RkdD2FhBoC5lncXpd9QYeBf50qitTaIiIhLDjNUU0xnQwxnxsjFljjFlpjOnqvtQd+NR9vIzjN/o8KYWGiEj4mQrcaq3tD9wD/N0d3wBc4j6eCDQwxpxSM9Zqd0W4iEg4M8bUB87ix24PALXcP+8B/maMuQ7nXjNpON0EAqbQEBEJLxHAYWttn5+/YK3di7un4YbLpafaFVmHp0REwoi19giwyxhzOfznlrJnuI+bubcOAPg98Oapfr5CQ0QkhLlNEb8CuhhjUo0xNwC/AG4wxmwANvPjCe8hwHfGmG1ADPDEKa9PDQtFRCRQ2tMQEZGAKTRERCRgCg0REQmYQkNERAKm0BARkYApNEREJGAKDRERCZhCQ0REAvb/B+M/u3b1AdAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def getvals(name):\n",
    "    x = list(results[name].duration.keys())\n",
    "    y = list(results[name].duration.values())\n",
    "    return x, y\n",
    "\n",
    "\n",
    "name = '2-tree-2'\n",
    "x = list(results[name].duration.keys())\n",
    "y = list(results[name].duration.values())\n",
    "plt.plot(*getvals('2-tree-2'), *getvals('1-ring-2'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "ax = plt.subplot(111)\n",
    "plt.plot(*getvals('2-tree-2'), label='tree')\n",
    "plt.plot(*getvals('1-ring-2'), label='ring')\n",
    "ax.legend()\n",
    "plt.title('2-machine EFA')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Automatic pdb calling has been turned ON\n"
     ]
    },
    {
     "ename": "TypeError",
     "evalue": "string indices must be integers",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/wandb/meta.py\u001b[0m in \u001b[0;36msetup\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m     46\u001b[0m             \u001b[0;32mimport\u001b[0m \u001b[0m__main__\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 47\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"program\"\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m__main__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__file__\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     48\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mImportError\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mAttributeError\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: module '__main__' has no attribute '__file__'",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-16-396ff2e31a0d>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mget_ipython\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_line_magic\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'pdb'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m''\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mwandb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'/tmp/nccl.txt'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/wandb/__init__.py\u001b[0m in \u001b[0;36msave\u001b[0;34m(glob_str, base_path, policy)\u001b[0m\n\u001b[1;32m    402\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    403\u001b[0m     run.send_message(\n\u001b[0;32m--> 404\u001b[0;31m         {\"save_policy\": {\"glob\": wandb_glob_str, \"policy\": policy}})\n\u001b[0m\u001b[1;32m    405\u001b[0m     \u001b[0mfiles\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    406\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0mpath\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mglob\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mglob\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mglob_str\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/wandb/wandb_run.py\u001b[0m in \u001b[0;36msend_message\u001b[0;34m(self, options)\u001b[0m\n\u001b[1;32m    161\u001b[0m         \u001b[0;32melif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jupyter_agent\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    162\u001b[0m             \u001b[0;31m# Running in jupyter\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 163\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_jupyter_agent\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstart\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    164\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0moptions\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"save_policy\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    165\u001b[0m                 self._jupyter_agent.rm.update_user_file_policy(\n",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/wandb/jupyter.py\u001b[0m in \u001b[0;36mstart\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    110\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mstart\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    111\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpaused\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 112\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrm\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mRunManager\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mwandb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutput\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    113\u001b[0m             \u001b[0mwandb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mapi\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_file_stream_api\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    114\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmirror_stdout_stderr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/wandb/run_manager.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, run, project, tags, cloud, output, port)\u001b[0m\n\u001b[1;32m    454\u001b[0m         \u001b[0;31m# Calling .start() on _meta and _system_stats will spin a thread that reports system stats every 30 seconds\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    455\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_system_stats\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstats\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSystemStats\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_api\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 456\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_meta\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmeta\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mMeta\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_api\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_run\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdir\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    457\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_meta\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"jobType\"\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_run\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjob_type\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    458\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_meta\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"mode\"\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_run\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmode\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/wandb/meta.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, api, out_dir)\u001b[0m\n\u001b[1;32m     34\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlock\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mthreading\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mLock\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 36\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msetup\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     37\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_thread\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mthreading\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mThread\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtarget\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_thread_body\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     38\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_thread\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdaemon\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/wandb/meta.py\u001b[0m in \u001b[0;36msetup\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m     49\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"program\"\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m'<python with no main file>'\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     50\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mwandb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_get_python_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m\"python\"\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 51\u001b[0;31m                 \u001b[0mmeta\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mwandb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjupyter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnotebook_metadata\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     52\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mmeta\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"path\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     53\u001b[0m                     \u001b[0;32mif\u001b[0m \u001b[0;34m\"fileId=\"\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmeta\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"path\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/main/lib/python3.6/site-packages/wandb/jupyter.py\u001b[0m in \u001b[0;36mnotebook_metadata\u001b[0;34m()\u001b[0m\n\u001b[1;32m     98\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     99\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mnn\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mres\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 100\u001b[0;31m             \u001b[0;32mif\u001b[0m \u001b[0mnn\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'kernel'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'id'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mkernel_id\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    101\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m\"root\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0ms\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'notebook_dir'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"path\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mnn\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'notebook'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'path'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"name\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mnn\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'notebook'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'name'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    102\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mTypeError\u001b[0m: string indices must be integers"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "> \u001b[0;32m/Users/yaroslavvb/anaconda3/envs/main/lib/python3.6/site-packages/wandb/jupyter.py\u001b[0m(100)\u001b[0;36mnotebook_metadata\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32m     98 \u001b[0;31m            \u001b[0;32mreturn\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0m\u001b[0;32m     99 \u001b[0;31m        \u001b[0;32mfor\u001b[0m \u001b[0mnn\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mres\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0m\u001b[0;32m--> 100 \u001b[0;31m            \u001b[0;32mif\u001b[0m \u001b[0mnn\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'kernel'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'id'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mkernel_id\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0m\u001b[0;32m    101 \u001b[0;31m                \u001b[0;32mreturn\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m\"root\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0ms\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'notebook_dir'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"path\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mnn\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'notebook'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'path'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"name\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mnn\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'notebook'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'name'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0m\u001b[0;32m    102 \u001b[0;31m    \u001b[0;32mreturn\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0m\n",
      "ipdb> p nn\n",
      "'message'\n",
      "--KeyboardInterrupt--\n"
     ]
    }
   ],
   "source": [
    "%pdb\n",
    "wandb.save('/tmp/nccl.txt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%ipdb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'3.6.8 |Anaconda, Inc.| (default, Dec 29 2018, 19:04:46) \\n[GCC 4.2.1 Compatible Clang 4.0.1 (tags/RELEASE_401/final)]'"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import sys\n",
    "sys.version"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import jupyter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "module 'jupyter' has no attribute '__version__'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-5-f8179a769d28>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mjupyter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__version__\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m: module 'jupyter' has no attribute '__version__'"
     ]
    }
   ],
   "source": [
    "jupyter.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/bin/sh: conda: command not found\r\n"
     ]
    }
   ],
   "source": [
    "!conda list ipython"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4.4.0\r\n"
     ]
    }
   ],
   "source": [
    "!jupyter --version"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'api' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-f7f2754b99be>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mruns\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mapi\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mruns\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"yaroslavvb/testtest\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'api' is not defined"
     ]
    }
   ],
   "source": [
    "runs = api.runs(\"yaroslavvb/testtest\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "165px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}