OSU_benchmarks/analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "88611c77-f911-43ce-9d5e-75b9a33ce186",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas\n",
    "import glob\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1ab186f5-24bc-4ffd-8067-a518fd2c0182",
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_meta_dataframe(log_data):\n",
    "    files = glob.glob(log_data)\n",
    "    df = pandas.DataFrame()\n",
    "    for file in files:\n",
    "        df_tmp = pandas.read_csv(file, delimiter=';', header=None, converters={0:str.strip, 1:str.strip}).transpose()\n",
    "        df_tmp.rename(columns=df_tmp.iloc[0], inplace=True)\n",
    "        df_tmp.drop([0], inplace=True)\n",
    "        df_tmp['logfile'] = file.split('.')[0]\n",
    "        df = df.append(df_tmp, ignore_index=True)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "882d5e64-8ad3-4f27-a81c-f7d59537f377",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_meta = load_meta_dataframe('nesh_logs/*.meta')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "958c6d90-80f2-4db2-a7bf-a9cea62cd12e",
   "metadata": {},
   "outputs": [
    {
     "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>timestamp</th>\n",
       "      <th>osu_exe</th>\n",
       "      <th>ntasks</th>\n",
       "      <th>nodes</th>\n",
       "      <th>tasks-per-node</th>\n",
       "      <th>constraint</th>\n",
       "      <th>nodelist</th>\n",
       "      <th>slurm_log</th>\n",
       "      <th>runtime</th>\n",
       "      <th>mpi_api</th>\n",
       "      <th>logfile</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2021-06-21T19:11:53+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_openmpi-3.1.6</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:11:53+02:00_a4b780e4-8...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2021-06-21T19:14:01+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:01+02:00_c57ff681-d...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2021-06-21T19:13:34+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>4(x2)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[173-174]</td>\n",
       "      <td>slurm-1333635.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2021-06-21T19:12:03+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>4(x2)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[173-174]</td>\n",
       "      <td>slurm-1333635.out</td>\n",
       "      <td>native_openmpi-3.1.6</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:12:03+02:00_b0e94454-6...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2021-06-21T19:12:18+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_openmpi-3.1.6</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:12:18+02:00_d48f4352-3...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>235</th>\n",
       "      <td>2021-06-21T19:11:47+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_openmpi-3.1.6</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:11:47+02:00_7281264c-0...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>236</th>\n",
       "      <td>2021-06-21T19:14:34+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>singularity_mpich-3.4.1</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:34+02:00_0791fd01-5...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>237</th>\n",
       "      <td>2021-06-21T19:15:22+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>singularity_mpich-3.4.1</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:15:22+02:00_5ae3e82d-b...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>238</th>\n",
       "      <td>2021-06-21T19:14:03+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "      <td>2(x4)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[268-271]</td>\n",
       "      <td>slurm-1333634.out</td>\n",
       "      <td>singularity_mpich-3.4.1</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:03+02:00_e26bf8c0-9...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>239</th>\n",
       "      <td>2021-06-21T19:14:08+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>240 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     timestamp                       osu_exe ntasks nodes  \\\n",
       "0    2021-06-21T19:11:53+02:00          mpi/startup/osu_init      8     8   \n",
       "1    2021-06-21T19:14:01+02:00          mpi/startup/osu_init      8     8   \n",
       "2    2021-06-21T19:13:34+02:00  mpi/collective/osu_allgather      8     2   \n",
       "3    2021-06-21T19:12:03+02:00  mpi/collective/osu_allgather      8     2   \n",
       "4    2021-06-21T19:12:18+02:00  mpi/collective/osu_allgather      8     8   \n",
       "..                         ...                           ...    ...   ...   \n",
       "235  2021-06-21T19:11:47+02:00          mpi/startup/osu_init      8     8   \n",
       "236  2021-06-21T19:14:34+02:00          mpi/startup/osu_init      8     8   \n",
       "237  2021-06-21T19:15:22+02:00  mpi/collective/osu_allgather      8     8   \n",
       "238  2021-06-21T19:14:03+02:00          mpi/startup/osu_init      8     4   \n",
       "239  2021-06-21T19:14:08+02:00  mpi/collective/osu_allgather      8     8   \n",
       "\n",
       "    tasks-per-node      constraint                 nodelist  \\\n",
       "0            1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "1            1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "2            4(x2)  cascade,mem192          neshcl[173-174]   \n",
       "3            4(x2)  cascade,mem192          neshcl[173-174]   \n",
       "4            1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "..             ...             ...                      ...   \n",
       "235          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "236          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "237          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "238          2(x4)  cascade,mem192          neshcl[268-271]   \n",
       "239          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "\n",
       "             slurm_log                  runtime mpi_api  \\\n",
       "0    slurm-1333633.out     native_openmpi-3.1.6    pmi2   \n",
       "1    slurm-1333633.out   native_intelmpi-20.0.4    pmi2   \n",
       "2    slurm-1333635.out   native_intelmpi-20.0.4    pmi2   \n",
       "3    slurm-1333635.out     native_openmpi-3.1.6    pmi2   \n",
       "4    slurm-1333633.out     native_openmpi-3.1.6    pmi2   \n",
       "..                 ...                      ...     ...   \n",
       "235  slurm-1333633.out     native_openmpi-3.1.6    pmi2   \n",
       "236  slurm-1333633.out  singularity_mpich-3.4.1    pmi2   \n",
       "237  slurm-1333633.out  singularity_mpich-3.4.1    pmi2   \n",
       "238  slurm-1333634.out  singularity_mpich-3.4.1    pmi2   \n",
       "239  slurm-1333633.out   native_intelmpi-20.0.4    pmi2   \n",
       "\n",
       "                                               logfile  \n",
       "0    nesh_logs/2021-06-21T19:11:53+02:00_a4b780e4-8...  \n",
       "1    nesh_logs/2021-06-21T19:14:01+02:00_c57ff681-d...  \n",
       "2    nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...  \n",
       "3    nesh_logs/2021-06-21T19:12:03+02:00_b0e94454-6...  \n",
       "4    nesh_logs/2021-06-21T19:12:18+02:00_d48f4352-3...  \n",
       "..                                                 ...  \n",
       "235  nesh_logs/2021-06-21T19:11:47+02:00_7281264c-0...  \n",
       "236  nesh_logs/2021-06-21T19:14:34+02:00_0791fd01-5...  \n",
       "237  nesh_logs/2021-06-21T19:15:22+02:00_5ae3e82d-b...  \n",
       "238  nesh_logs/2021-06-21T19:14:03+02:00_e26bf8c0-9...  \n",
       "239  nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...  \n",
       "\n",
       "[240 rows x 11 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_meta"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d41859c3-7ec4-46b6-b823-c8bced4f3b27",
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_osu_mpi_init(files):\n",
    "    converters = {\n",
    "        0: lambda x: int(x.split(':')[-1]),\n",
    "        1: lambda x: float(x.split(':')[-1].split('ms')[0]),\n",
    "        2: lambda x: float(x.split(':')[-1].split('ms')[0]),\n",
    "        3: lambda x: float(x.split(':')[-1].split('ms')[0]),\n",
    "    }\n",
    "    df = pandas.DataFrame()\n",
    "    for file in files:\n",
    "        df_tmp = pandas.read_csv(file+'.log', comment='#', header=None, converters=converters, names=['nprocs', 'min', 'max', 'avg'])\n",
    "        df_tmp['logfile'] = file\n",
    "        df = df.append(df_tmp, ignore_index=True)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "94d9eb19-0619-47a3-9003-9f7e98da51e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "mask = df_meta['osu_exe'].str.contains('mpi/startup/osu_init')\n",
    "files = df_meta[mask]['logfile'].to_list()\n",
    "df_init = read_osu_mpi_init(files) # 120 experiments! 3-runtimes * 4-node-configs * 10-experiments\n",
    "df_init = pandas.merge(df_meta, df_init, on='logfile', how='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e50d85f6-35e5-4365-9016-e7070872e831",
   "metadata": {},
   "outputs": [
    {
     "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>timestamp</th>\n",
       "      <th>osu_exe</th>\n",
       "      <th>ntasks</th>\n",
       "      <th>nodes</th>\n",
       "      <th>tasks-per-node</th>\n",
       "      <th>constraint</th>\n",
       "      <th>nodelist</th>\n",
       "      <th>slurm_log</th>\n",
       "      <th>runtime</th>\n",
       "      <th>mpi_api</th>\n",
       "      <th>logfile</th>\n",
       "      <th>nprocs</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>avg</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2021-06-21T19:11:53+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_openmpi-3.1.6</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:11:53+02:00_a4b780e4-8...</td>\n",
       "      <td>8</td>\n",
       "      <td>109.0</td>\n",
       "      <td>121.0</td>\n",
       "      <td>116.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2021-06-21T19:14:01+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:01+02:00_c57ff681-d...</td>\n",
       "      <td>8</td>\n",
       "      <td>876.0</td>\n",
       "      <td>895.0</td>\n",
       "      <td>888.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2021-06-21T19:13:37+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "      <td>2(x4)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[268-271]</td>\n",
       "      <td>slurm-1333634.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:13:37+02:00_eb9b4751-4...</td>\n",
       "      <td>8</td>\n",
       "      <td>354.0</td>\n",
       "      <td>357.0</td>\n",
       "      <td>356.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2021-06-21T19:12:55+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl126</td>\n",
       "      <td>slurm-1333636.out</td>\n",
       "      <td>singularity_mpich-3.4.1</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:12:55+02:00_be78d802-8...</td>\n",
       "      <td>8</td>\n",
       "      <td>33.0</td>\n",
       "      <td>110.0</td>\n",
       "      <td>67.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2021-06-21T19:13:55+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>4(x2)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[173-174]</td>\n",
       "      <td>slurm-1333635.out</td>\n",
       "      <td>singularity_mpich-3.4.1</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:13:55+02:00_a2c9ed3d-4...</td>\n",
       "      <td>8</td>\n",
       "      <td>37.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>55.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115</th>\n",
       "      <td>2021-06-21T19:11:54+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_openmpi-3.1.6</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:11:54+02:00_5c2eda68-3...</td>\n",
       "      <td>8</td>\n",
       "      <td>74.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>78.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116</th>\n",
       "      <td>2021-06-21T19:11:51+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_openmpi-3.1.6</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:11:51+02:00_c6fa3090-1...</td>\n",
       "      <td>8</td>\n",
       "      <td>68.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>73.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117</th>\n",
       "      <td>2021-06-21T19:11:47+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_openmpi-3.1.6</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:11:47+02:00_7281264c-0...</td>\n",
       "      <td>8</td>\n",
       "      <td>124.0</td>\n",
       "      <td>330.0</td>\n",
       "      <td>174.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118</th>\n",
       "      <td>2021-06-21T19:14:34+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>singularity_mpich-3.4.1</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:34+02:00_0791fd01-5...</td>\n",
       "      <td>8</td>\n",
       "      <td>25.0</td>\n",
       "      <td>55.0</td>\n",
       "      <td>41.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>2021-06-21T19:14:03+02:00</td>\n",
       "      <td>mpi/startup/osu_init</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "      <td>2(x4)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[268-271]</td>\n",
       "      <td>slurm-1333634.out</td>\n",
       "      <td>singularity_mpich-3.4.1</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:03+02:00_e26bf8c0-9...</td>\n",
       "      <td>8</td>\n",
       "      <td>35.0</td>\n",
       "      <td>58.0</td>\n",
       "      <td>46.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>120 rows × 15 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     timestamp               osu_exe ntasks nodes  \\\n",
       "0    2021-06-21T19:11:53+02:00  mpi/startup/osu_init      8     8   \n",
       "1    2021-06-21T19:14:01+02:00  mpi/startup/osu_init      8     8   \n",
       "2    2021-06-21T19:13:37+02:00  mpi/startup/osu_init      8     4   \n",
       "3    2021-06-21T19:12:55+02:00  mpi/startup/osu_init      8     1   \n",
       "4    2021-06-21T19:13:55+02:00  mpi/startup/osu_init      8     2   \n",
       "..                         ...                   ...    ...   ...   \n",
       "115  2021-06-21T19:11:54+02:00  mpi/startup/osu_init      8     8   \n",
       "116  2021-06-21T19:11:51+02:00  mpi/startup/osu_init      8     8   \n",
       "117  2021-06-21T19:11:47+02:00  mpi/startup/osu_init      8     8   \n",
       "118  2021-06-21T19:14:34+02:00  mpi/startup/osu_init      8     8   \n",
       "119  2021-06-21T19:14:03+02:00  mpi/startup/osu_init      8     4   \n",
       "\n",
       "    tasks-per-node      constraint                 nodelist  \\\n",
       "0            1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "1            1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "2            2(x4)  cascade,mem192          neshcl[268-271]   \n",
       "3                8  cascade,mem192                neshcl126   \n",
       "4            4(x2)  cascade,mem192          neshcl[173-174]   \n",
       "..             ...             ...                      ...   \n",
       "115          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "116          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "117          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "118          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "119          2(x4)  cascade,mem192          neshcl[268-271]   \n",
       "\n",
       "             slurm_log                  runtime mpi_api  \\\n",
       "0    slurm-1333633.out     native_openmpi-3.1.6    pmi2   \n",
       "1    slurm-1333633.out   native_intelmpi-20.0.4    pmi2   \n",
       "2    slurm-1333634.out   native_intelmpi-20.0.4    pmi2   \n",
       "3    slurm-1333636.out  singularity_mpich-3.4.1    pmi2   \n",
       "4    slurm-1333635.out  singularity_mpich-3.4.1    pmi2   \n",
       "..                 ...                      ...     ...   \n",
       "115  slurm-1333633.out     native_openmpi-3.1.6    pmi2   \n",
       "116  slurm-1333633.out     native_openmpi-3.1.6    pmi2   \n",
       "117  slurm-1333633.out     native_openmpi-3.1.6    pmi2   \n",
       "118  slurm-1333633.out  singularity_mpich-3.4.1    pmi2   \n",
       "119  slurm-1333634.out  singularity_mpich-3.4.1    pmi2   \n",
       "\n",
       "                                               logfile  nprocs    min    max  \\\n",
       "0    nesh_logs/2021-06-21T19:11:53+02:00_a4b780e4-8...       8  109.0  121.0   \n",
       "1    nesh_logs/2021-06-21T19:14:01+02:00_c57ff681-d...       8  876.0  895.0   \n",
       "2    nesh_logs/2021-06-21T19:13:37+02:00_eb9b4751-4...       8  354.0  357.0   \n",
       "3    nesh_logs/2021-06-21T19:12:55+02:00_be78d802-8...       8   33.0  110.0   \n",
       "4    nesh_logs/2021-06-21T19:13:55+02:00_a2c9ed3d-4...       8   37.0   73.0   \n",
       "..                                                 ...     ...    ...    ...   \n",
       "115  nesh_logs/2021-06-21T19:11:54+02:00_5c2eda68-3...       8   74.0   87.0   \n",
       "116  nesh_logs/2021-06-21T19:11:51+02:00_c6fa3090-1...       8   68.0   81.0   \n",
       "117  nesh_logs/2021-06-21T19:11:47+02:00_7281264c-0...       8  124.0  330.0   \n",
       "118  nesh_logs/2021-06-21T19:14:34+02:00_0791fd01-5...       8   25.0   55.0   \n",
       "119  nesh_logs/2021-06-21T19:14:03+02:00_e26bf8c0-9...       8   35.0   58.0   \n",
       "\n",
       "       avg  \n",
       "0    116.0  \n",
       "1    888.0  \n",
       "2    356.0  \n",
       "3     67.0  \n",
       "4     55.0  \n",
       "..     ...  \n",
       "115   78.0  \n",
       "116   73.0  \n",
       "117  174.0  \n",
       "118   41.0  \n",
       "119   46.0  \n",
       "\n",
       "[120 rows x 15 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_init"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b29feb35-aabb-4c31-b681-c88ff5fbbeec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8, 4))\n",
    "df_init[['runtime', 'avg']].pivot(columns='runtime', values='avg').boxplot();\n",
    "plt.ylabel('Latency in ms'); plt.title('startup/osu_init');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "8b087cd2-b81c-4345-916b-925fbb0f4a27",
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_osu_collective_allgather(files):\n",
    "    df = pandas.DataFrame()\n",
    "    for file in files:\n",
    "        df_tmp = pandas.read_csv(file+'.log', comment='#', header=None, delim_whitespace=True, names=['message_size', 'latency_avg'])\n",
    "        df_tmp['logfile'] = file\n",
    "        df = df.append(df_tmp, ignore_index=True)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "eb0ba1d5-61e0-46c6-9037-b77108043f25",
   "metadata": {},
   "outputs": [],
   "source": [
    "mask = df_meta['osu_exe'].str.contains('mpi/collective/osu_allgather')\n",
    "files = df_meta[mask]['logfile'].to_list()\n",
    "df_allgather = read_osu_collective_allgather(files)\n",
    "df_allgather = pandas.merge(df_meta, df_allgather, on='logfile', how='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "63421286-a03d-47fc-940c-b6ff62caf6d0",
   "metadata": {},
   "outputs": [
    {
     "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>timestamp</th>\n",
       "      <th>osu_exe</th>\n",
       "      <th>ntasks</th>\n",
       "      <th>nodes</th>\n",
       "      <th>tasks-per-node</th>\n",
       "      <th>constraint</th>\n",
       "      <th>nodelist</th>\n",
       "      <th>slurm_log</th>\n",
       "      <th>runtime</th>\n",
       "      <th>mpi_api</th>\n",
       "      <th>logfile</th>\n",
       "      <th>message_size</th>\n",
       "      <th>latency_avg</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2021-06-21T19:13:34+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>4(x2)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[173-174]</td>\n",
       "      <td>slurm-1333635.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...</td>\n",
       "      <td>1</td>\n",
       "      <td>3.52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2021-06-21T19:13:34+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>4(x2)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[173-174]</td>\n",
       "      <td>slurm-1333635.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...</td>\n",
       "      <td>2</td>\n",
       "      <td>2.67</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2021-06-21T19:13:34+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>4(x2)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[173-174]</td>\n",
       "      <td>slurm-1333635.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...</td>\n",
       "      <td>4</td>\n",
       "      <td>2.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2021-06-21T19:13:34+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>4(x2)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[173-174]</td>\n",
       "      <td>slurm-1333635.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...</td>\n",
       "      <td>8</td>\n",
       "      <td>3.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2021-06-21T19:13:34+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>4(x2)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[173-174]</td>\n",
       "      <td>slurm-1333635.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...</td>\n",
       "      <td>16</td>\n",
       "      <td>2.68</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2515</th>\n",
       "      <td>2021-06-21T19:14:08+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...</td>\n",
       "      <td>65536</td>\n",
       "      <td>134.53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2516</th>\n",
       "      <td>2021-06-21T19:14:08+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...</td>\n",
       "      <td>131072</td>\n",
       "      <td>259.65</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2517</th>\n",
       "      <td>2021-06-21T19:14:08+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...</td>\n",
       "      <td>262144</td>\n",
       "      <td>415.27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2518</th>\n",
       "      <td>2021-06-21T19:14:08+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...</td>\n",
       "      <td>524288</td>\n",
       "      <td>712.87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2519</th>\n",
       "      <td>2021-06-21T19:14:08+02:00</td>\n",
       "      <td>mpi/collective/osu_allgather</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1(x8)</td>\n",
       "      <td>cascade,mem192</td>\n",
       "      <td>neshcl[142-143,268-273]</td>\n",
       "      <td>slurm-1333633.out</td>\n",
       "      <td>native_intelmpi-20.0.4</td>\n",
       "      <td>pmi2</td>\n",
       "      <td>nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...</td>\n",
       "      <td>1048576</td>\n",
       "      <td>1324.90</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2520 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                      timestamp                       osu_exe ntasks nodes  \\\n",
       "0     2021-06-21T19:13:34+02:00  mpi/collective/osu_allgather      8     2   \n",
       "1     2021-06-21T19:13:34+02:00  mpi/collective/osu_allgather      8     2   \n",
       "2     2021-06-21T19:13:34+02:00  mpi/collective/osu_allgather      8     2   \n",
       "3     2021-06-21T19:13:34+02:00  mpi/collective/osu_allgather      8     2   \n",
       "4     2021-06-21T19:13:34+02:00  mpi/collective/osu_allgather      8     2   \n",
       "...                         ...                           ...    ...   ...   \n",
       "2515  2021-06-21T19:14:08+02:00  mpi/collective/osu_allgather      8     8   \n",
       "2516  2021-06-21T19:14:08+02:00  mpi/collective/osu_allgather      8     8   \n",
       "2517  2021-06-21T19:14:08+02:00  mpi/collective/osu_allgather      8     8   \n",
       "2518  2021-06-21T19:14:08+02:00  mpi/collective/osu_allgather      8     8   \n",
       "2519  2021-06-21T19:14:08+02:00  mpi/collective/osu_allgather      8     8   \n",
       "\n",
       "     tasks-per-node      constraint                 nodelist  \\\n",
       "0             4(x2)  cascade,mem192          neshcl[173-174]   \n",
       "1             4(x2)  cascade,mem192          neshcl[173-174]   \n",
       "2             4(x2)  cascade,mem192          neshcl[173-174]   \n",
       "3             4(x2)  cascade,mem192          neshcl[173-174]   \n",
       "4             4(x2)  cascade,mem192          neshcl[173-174]   \n",
       "...             ...             ...                      ...   \n",
       "2515          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "2516          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "2517          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "2518          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "2519          1(x8)  cascade,mem192  neshcl[142-143,268-273]   \n",
       "\n",
       "              slurm_log                 runtime mpi_api  \\\n",
       "0     slurm-1333635.out  native_intelmpi-20.0.4    pmi2   \n",
       "1     slurm-1333635.out  native_intelmpi-20.0.4    pmi2   \n",
       "2     slurm-1333635.out  native_intelmpi-20.0.4    pmi2   \n",
       "3     slurm-1333635.out  native_intelmpi-20.0.4    pmi2   \n",
       "4     slurm-1333635.out  native_intelmpi-20.0.4    pmi2   \n",
       "...                 ...                     ...     ...   \n",
       "2515  slurm-1333633.out  native_intelmpi-20.0.4    pmi2   \n",
       "2516  slurm-1333633.out  native_intelmpi-20.0.4    pmi2   \n",
       "2517  slurm-1333633.out  native_intelmpi-20.0.4    pmi2   \n",
       "2518  slurm-1333633.out  native_intelmpi-20.0.4    pmi2   \n",
       "2519  slurm-1333633.out  native_intelmpi-20.0.4    pmi2   \n",
       "\n",
       "                                                logfile  message_size  \\\n",
       "0     nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...             1   \n",
       "1     nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...             2   \n",
       "2     nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...             4   \n",
       "3     nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...             8   \n",
       "4     nesh_logs/2021-06-21T19:13:34+02:00_deb80a52-4...            16   \n",
       "...                                                 ...           ...   \n",
       "2515  nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...         65536   \n",
       "2516  nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...        131072   \n",
       "2517  nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...        262144   \n",
       "2518  nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...        524288   \n",
       "2519  nesh_logs/2021-06-21T19:14:08+02:00_7e4f1da4-c...       1048576   \n",
       "\n",
       "      latency_avg  \n",
       "0            3.52  \n",
       "1            2.67  \n",
       "2            2.64  \n",
       "3            3.33  \n",
       "4            2.68  \n",
       "...           ...  \n",
       "2515       134.53  \n",
       "2516       259.65  \n",
       "2517       415.27  \n",
       "2518       712.87  \n",
       "2519      1324.90  \n",
       "\n",
       "[2520 rows x 13 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_allgather"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "de3af88a-f342-4580-ad1c-0bbb21471caa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "for i,runtime in enumerate(df_allgather['runtime'].unique()):\n",
    "    plt.subplot(1,3,i+1)\n",
    "    df_allgather.groupby('runtime').get_group(runtime).pivot(columns='message_size', values='latency_avg').boxplot(rot=90)\n",
    "    plt.yscale('log'); plt.title(runtime); plt.ylabel('collective/osu_allgather\\nlatency in ms')\n",
    "    plt.ylim(1e-1, 1e4); plt.xlabel('message size')\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "abde098b-f0e2-4592-880f-189e9589c54d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['4(x2)', '1(x8)', '2(x4)', '8'], dtype=object)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_allgather['tasks-per-node'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "204f99f2-b1bf-4033-9f0c-e29f8832da8e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "for i,runtime in enumerate(df_allgather['runtime'].unique()):\n",
    "    plt.subplot(1,3,i+1)\n",
    "    df_allgather.groupby('runtime').get_group(runtime).groupby('tasks-per-node').get_group('1(x8)').pivot(columns='message_size', values='latency_avg').boxplot(rot=90)\n",
    "    plt.yscale('log'); plt.title(runtime); plt.ylabel('collective/osu_allgather\\nlatency in ms')\n",
    "    plt.ylim(1e-1, 1e4); plt.xlabel('message size')\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c2dab865-57f5-44cc-9fc7-eb5fbde5ca70",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "for i,runtime in enumerate(df_allgather['runtime'].unique()):\n",
    "    plt.subplot(1,3,i+1)\n",
    "    df_allgather.groupby('runtime').get_group(runtime).groupby('tasks-per-node').get_group('8').pivot(columns='message_size', values='latency_avg').boxplot(rot=90)\n",
    "    plt.yscale('log'); plt.title(runtime); plt.ylabel('collective/osu_allgather\\nlatency in ms')\n",
    "    plt.ylim(1e-1, 1e4); plt.xlabel('message size')\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "075fff3e-3588-473a-a921-29767c534dae",
   "metadata": {},
   "source": [
    "Python environment,"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "edd3b566-5794-4016-85b2-107512bce8b2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Package                       Version\n",
      "----------------------------- -------------------\n",
      "anyio                         3.1.0\n",
      "argon2-cffi                   20.1.0\n",
      "async-generator               1.10\n",
      "attrs                         21.2.0\n",
      "Babel                         2.9.1\n",
      "backcall                      0.2.0\n",
      "backports.functools-lru-cache 1.6.4\n",
      "bleach                        3.3.0\n",
      "brotlipy                      0.7.0\n",
      "certifi                       2020.12.5\n",
      "cffi                          1.14.4\n",
      "chardet                       3.0.4\n",
      "conda                         4.10.1\n",
      "conda-package-handling        1.7.2\n",
      "cryptography                  3.3.1\n",
      "cycler                        0.10.0\n",
      "decorator                     5.0.9\n",
      "defusedxml                    0.7.1\n",
      "entrypoints                   0.3\n",
      "idna                          2.10\n",
      "importlib-metadata            4.2.0\n",
      "ipykernel                     5.5.5\n",
      "ipython                       7.23.1\n",
      "ipython-genutils              0.2.0\n",
      "jedi                          0.18.0\n",
      "Jinja2                        2.11.3\n",
      "json5                         0.9.5\n",
      "jsonschema                    3.2.0\n",
      "jupyter-client                6.1.12\n",
      "jupyter-core                  4.7.1\n",
      "jupyter-server                1.8.0\n",
      "jupyterlab                    3.0.16\n",
      "jupyterlab-pygments           0.1.2\n",
      "jupyterlab-server             2.5.2\n",
      "kiwisolver                    1.3.1\n",
      "MarkupSafe                    1.1.1\n",
      "matplotlib                    3.3.4\n",
      "matplotlib-inline             0.1.2\n",
      "mistune                       0.8.4\n",
      "nbclassic                     0.3.1\n",
      "nbclient                      0.5.3\n",
      "nbconvert                     6.0.7\n",
      "nbformat                      5.1.3\n",
      "nest-asyncio                  1.5.1\n",
      "notebook                      6.4.0\n",
      "numpy                         1.19.4\n",
      "olefile                       0.46\n",
      "packaging                     20.9\n",
      "pandas                        1.1.4\n",
      "pandocfilters                 1.4.2\n",
      "parso                         0.8.2\n",
      "pexpect                       4.8.0\n",
      "pickleshare                   0.7.5\n",
      "Pillow                        7.2.0\n",
      "pip                           20.3.1\n",
      "prometheus-client             0.10.1\n",
      "prompt-toolkit                3.0.18\n",
      "ptyprocess                    0.7.0\n",
      "pycosat                       0.6.3\n",
      "pycparser                     2.20\n",
      "Pygments                      2.9.0\n",
      "pyOpenSSL                     20.0.0\n",
      "pyparsing                     2.4.7\n",
      "pyrsistent                    0.17.3\n",
      "PySocks                       1.7.1\n",
      "python-dateutil               2.8.1\n",
      "pytz                          2021.1\n",
      "pyzmq                         19.0.2\n",
      "requests                      2.25.0\n",
      "ruamel-yaml-conda             0.15.80\n",
      "Send2Trash                    1.5.0\n",
      "setuptools                    51.0.0.post20201207\n",
      "sip                           4.19.13\n",
      "six                           1.15.0\n",
      "sniffio                       1.2.0\n",
      "terminado                     0.10.0\n",
      "testpath                      0.5.0\n",
      "tornado                       6.1\n",
      "tqdm                          4.54.1\n",
      "traitlets                     5.0.5\n",
      "urllib3                       1.25.11\n",
      "wcwidth                       0.2.5\n",
      "webencodings                  0.5.1\n",
      "websocket-client              0.57.0\n",
      "wheel                         0.36.1\n",
      "zipp                          3.4.1\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "ff9e9ee9-6efb-4187-b8e0-411b565f098a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# This file may be used to create an environment using:\n",
      "# $ conda create --name <env> --file <this file>\n",
      "# platform: linux-64\n",
      "@EXPLICIT\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/_libgcc_mutex-0.1-main.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/ca-certificates-2020.12.5-ha878542_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/ld_impl_linux-64-2.33.1-h53a641e_7.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libgfortran4-7.5.0-h14aa051_19.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/libstdcxx-ng-9.1.0-hdf63c60_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/pandoc-2.13-h7f98852_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/noarch/tzdata-2020d-h14c3975_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/libgcc-ng-9.1.0-hdf63c60_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libgfortran-ng-7.5.0-h14aa051_19.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/expat-2.4.1-h2531618_2.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/icu-58.2-hf484d3e_1000.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/jpeg-9d-h36c2ea0_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/libffi-3.3-he6710b0_2.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libopenblas-0.3.12-pthreads_hb3c22a3_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libsodium-1.0.18-h36c2ea0_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libuuid-2.32.1-h14c3975_1000.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/lz4-c-1.9.2-he1b5a44_3.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/ncurses-6.2-he6710b0_1.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/openssl-1.1.1k-h27cfd23_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/pcre-8.44-he1b5a44_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/pthread-stubs-0.4-h36c2ea0_1001.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/xorg-libxau-1.0.9-h14c3975_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/xorg-libxdmcp-1.1.3-h516909a_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/xz-5.2.5-h7b6447c_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/yaml-0.2.5-h7b6447c_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/zlib-1.2.11-h7b6447c_3.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/glib-2.68.2-h36276a3_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libblas-3.9.0-8_openblas.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/libedit-3.1.20191231-h14c3975_1.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libpng-1.6.37-h21135ba_2.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libxcb-1.13-h14c3975_1002.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/libxml2-2.9.10-hb55368b_3.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/readline-8.0-h7b6447c_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/tk-8.6.10-hbc83047_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/zeromq-4.3.4-h2531618_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/zstd-1.4.5-h9ceee32_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/dbus-1.13.6-he372182_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/freetype-2.10.4-h7ca028e_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/gstreamer-1.14.0-h28cd5cc_2.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/libcblas-3.9.0-8_openblas.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/liblapack-3.9.0-8_openblas.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/libtiff-4.1.0-h2733197_1.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/sqlite-3.33.0-h62c20be_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/fontconfig-2.13.1-he4413a7_1000.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/gst-plugins-base-1.14.0-hbbd80ab_1.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/lcms2-2.11-hcbb858e_1.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/python-3.9.1-hdb3f193_2.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/async_generator-1.10-py_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/attrs-21.2.0-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/backcall-0.2.0-pyh9f0ad1d_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/backports-1.0-py_2.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/chardet-3.0.4-py39h06a4308_1003.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/decorator-5.0.9-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/defusedxml-0.7.1-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/entrypoints-0.3-pyhd8ed1ab_1003.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/noarch/idna-2.10-py_0.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/ipython_genutils-0.2.0-py_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/json5-0.9.5-pyh9f0ad1d_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/nest-asyncio-1.5.1-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/olefile-0.46-pyh9f0ad1d_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/pandocfilters-1.4.2-py_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/parso-0.8.2-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/pickleshare-0.7.5-py_1003.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/prometheus_client-0.10.1-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/ptyprocess-0.7.0-pyhd3deb0d_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/pycosat-0.6.3-py39h27cfd23_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/noarch/pycparser-2.20-py_2.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/pyparsing-2.4.7-pyh9f0ad1d_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/pysocks-1.7.1-py39h06a4308_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/python_abi-3.9-1_cp39.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/pytz-2021.1-pyhd8ed1ab_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/qt-5.9.7-h5867ecd_1.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/ruamel_yaml-0.15.80-py39h27cfd23_0.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/send2trash-1.5.0-py_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/sip-4.19.13-py39h2531618_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/six-1.15.0-py39h06a4308_0.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/testpath-0.5.0-pyhd8ed1ab_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/noarch/tqdm-4.54.1-pyhd3eb1b0_0.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/webencodings-0.5.1-py_1.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/noarch/wheel-0.36.1-pyhd3eb1b0_0.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/zipp-3.4.1-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/babel-2.9.1-pyh44b312d_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/certifi-2020.12.5-py39hf3d152e_1.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/cffi-1.14.4-py39h261ae71_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/conda-package-handling-1.7.2-py39h27cfd23_1.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/cycler-0.10.0-py_2.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/importlib-metadata-4.2.0-py39hf3d152e_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/jedi-0.18.0-py39hf3d152e_2.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/kiwisolver-1.3.1-py39h081fc7a_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/markupsafe-1.1.1-py39h38d8fee_2.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/mistune-0.8.4-py39hbd71b63_1002.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/numpy-1.19.4-py39h57d35e7_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/packaging-20.9-pyh44b312d_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/pexpect-4.8.0-pyh9f0ad1d_2.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/pillow-7.2.0-py39h6f3857e_2.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/pyqt-5.9.2-py39h2531618_6.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/pyrsistent-0.17.3-py39hbd71b63_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/python-dateutil-2.8.1-py_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/pyzmq-19.0.2-py39hb69f2a1_2.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/sniffio-1.2.0-py39hf3d152e_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/tornado-6.1-py39hbd71b63_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/traitlets-5.0.5-py_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/websocket-client-0.57.0-py39hf3d152e_4.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/anyio-3.1.0-py39hf3d152e_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/argon2-cffi-20.1.0-py39hbd71b63_2.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/brotlipy-0.7.0-py39h27cfd23_1003.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/cryptography-3.3.1-py39h3c74f83_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/jupyter_core-4.7.1-py39hf3d152e_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/matplotlib-inline-0.1.2-pyhd8ed1ab_2.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/pandas-1.1.4-py39h9cfe711_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/setuptools-51.0.0-py39h06a4308_2.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/terminado-0.10.0-py39hf3d152e_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/backports.functools_lru_cache-1.6.4-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/bleach-3.3.0-pyh44b312d_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/jinja2-2.11.3-pyh44b312d_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/jsonschema-3.2.0-pyhd8ed1ab_3.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/jupyter_client-6.1.12-pyhd8ed1ab_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/matplotlib-base-3.3.4-py39h62a2d02_0.conda\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/pip-20.3.1-py39h06a4308_0.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/pygments-2.9.0-pyhd8ed1ab_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/noarch/pyopenssl-20.0.0-pyhd3eb1b0_1.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/jupyterlab_pygments-0.1.2-pyh9f0ad1d_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/linux-64/matplotlib-3.3.4-py39h06a4308_0.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/nbformat-5.1.3-pyhd8ed1ab_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/noarch/urllib3-1.25.11-py_0.conda\n",
      "https://conda.anaconda.org/conda-forge/noarch/wcwidth-0.2.5-pyh9f0ad1d_2.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/nbclient-0.5.3-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/prompt-toolkit-3.0.18-pyha770c72_0.tar.bz2\n",
      "https://repo.anaconda.com/pkgs/main/noarch/requests-2.25.0-pyhd3eb1b0_0.conda\n",
      "https://conda.anaconda.org/conda-forge/linux-64/conda-4.10.1-py39hf3d152e_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/ipython-7.23.1-py39hef51801_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/nbconvert-6.0.7-py39hf3d152e_3.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/linux-64/ipykernel-5.5.5-py39hef51801_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/jupyter_server-1.8.0-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/jupyterlab_server-2.5.2-pyhd8ed1ab_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/notebook-6.4.0-pyha770c72_0.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/nbclassic-0.3.1-pyhd8ed1ab_1.tar.bz2\n",
      "https://conda.anaconda.org/conda-forge/noarch/jupyterlab-3.0.16-pyhd8ed1ab_0.tar.bz2\n"
     ]
    }
   ],
   "source": [
    "!conda list --explicit"
   ]
  }
 ],
 "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": "3.9.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}