notebooks/SoftwareAnalysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PyTorch software benchmark analysis\n",
    "\n",
    "Single node throughput comparisons between installation versions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import re\n",
    "import glob\n",
    "\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams.update({'font.size': 14})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[0m\u001b[01;34mgpu-1.7.0-nccl-n1\u001b[0m/         \u001b[01;34mgpu-v1.2.0-nccl-n1\u001b[0m/  \u001b[01;34mhsw-1.7.1-mpi-n4\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-gloo-n1\u001b[0m/         \u001b[01;34mgpu-v1.3.1-nccl-n1\u001b[0m/  \u001b[01;34mhsw-1.7.1-mpi-n64\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-gloo-n16\u001b[0m/        \u001b[01;34mgpu-v1.4.0-nccl-n1\u001b[0m/  \u001b[01;34mhsw-1.7.1-mpi-n8\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-gloo-n2\u001b[0m/         \u001b[01;34mgpu-v1.5.1-nccl-n1\u001b[0m/  \u001b[01;34mhsw-v1.2.0-mpi-n1\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-gloo-n32\u001b[0m/        \u001b[01;34mgpu-v1.6.0-nccl-n1\u001b[0m/  \u001b[01;34mhsw-v1.3.1-mpi-n1\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-gloo-n4\u001b[0m/         \u001b[01;34mhsw-1.7.1-gloo-n1\u001b[0m/   \u001b[01;34mhsw-v1.4.0-mpi-n1\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-gloo-n64\u001b[0m/        \u001b[01;34mhsw-1.7.1-gloo-n16\u001b[0m/  \u001b[01;34mhsw-v1.5.0-mpi-n1\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-gloo-n8\u001b[0m/         \u001b[01;34mhsw-1.7.1-gloo-n2\u001b[0m/   \u001b[01;34mhsw-v1.6.0-mpi-n1\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-nccl-n1\u001b[0m/         \u001b[01;34mhsw-1.7.1-gloo-n32\u001b[0m/  \u001b[01;34mknl-1.7.1-mpi-n1\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-nccl-n16\u001b[0m/        \u001b[01;34mhsw-1.7.1-gloo-n4\u001b[0m/   \u001b[01;34mknl-1.7.1-mpi-n16\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-nccl-n2\u001b[0m/         \u001b[01;34mhsw-1.7.1-gloo-n64\u001b[0m/  \u001b[01;34mknl-1.7.1-mpi-n2\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-nccl-n32\u001b[0m/        \u001b[01;34mhsw-1.7.1-gloo-n8\u001b[0m/   \u001b[01;34mknl-1.7.1-mpi-n32\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-nccl-n4\u001b[0m/         \u001b[01;34mhsw-1.7.1-mpi-n1\u001b[0m/    \u001b[01;34mknl-1.7.1-mpi-n4\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-nccl-n64\u001b[0m/        \u001b[01;34mhsw-1.7.1-mpi-n16\u001b[0m/   \u001b[01;34mknl-1.7.1-mpi-n64\u001b[0m/\n",
      "\u001b[01;34mgpu-1.7.1-nccl-n8\u001b[0m/         \u001b[01;34mhsw-1.7.1-mpi-n2\u001b[0m/    \u001b[01;34mknl-1.7.1-mpi-n8\u001b[0m/\n",
      "\u001b[01;34mgpu-ngc-20.10-v0-nccl-n1\u001b[0m/  \u001b[01;34mhsw-1.7.1-mpi-n32\u001b[0m/\n"
     ]
    }
   ],
   "source": [
    "ls /global/cscratch1/sd/sfarrell/pytorch-benchmarks/results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['/global/cscratch1/sd/sfarrell/pytorch-benchmarks/results/hsw-v1.3.1-mpi-n1/results.txt',\n",
       " '/global/cscratch1/sd/sfarrell/pytorch-benchmarks/results/hsw-v1.6.0-mpi-n1/results.txt',\n",
       " '/global/cscratch1/sd/sfarrell/pytorch-benchmarks/results/hsw-v1.2.0-mpi-n1/results.txt',\n",
       " '/global/cscratch1/sd/sfarrell/pytorch-benchmarks/results/hsw-1.7.1-mpi-n1/results.txt']"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_files = glob.glob('/global/cscratch1/sd/sfarrell/pytorch-benchmarks/results/hsw-*-n1/results.txt')\n",
    "\n",
    "results_files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>train_rate</th>\n",
       "      <th>inference_rate</th>\n",
       "      <th>ranks</th>\n",
       "      <th>hardware</th>\n",
       "      <th>version</th>\n",
       "      <th>backend</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>23.892666</td>\n",
       "      <td>46.485593</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.2.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>cnn3d</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>14.632251</td>\n",
       "      <td>37.498390</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.2.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>resnet50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>213.417174</td>\n",
       "      <td>682.692711</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.2.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>alexnet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>650.044671</td>\n",
       "      <td>2734.242050</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.2.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>lstm</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>7.608130</td>\n",
       "      <td>28.118234</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.2.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>transformer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>7.745227</td>\n",
       "      <td>27.457707</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.3.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>transformer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>644.677564</td>\n",
       "      <td>2610.069286</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.3.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>lstm</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>198.531888</td>\n",
       "      <td>684.354039</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.3.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>alexnet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>14.996481</td>\n",
       "      <td>46.225346</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.3.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>resnet50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>24.243660</td>\n",
       "      <td>48.183499</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.3.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>cnn3d</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>14.944277</td>\n",
       "      <td>27.766081</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.6.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>cnn3d</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>16.685013</td>\n",
       "      <td>48.906670</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.6.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>resnet50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>210.820440</td>\n",
       "      <td>710.939807</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.6.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>alexnet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>834.144656</td>\n",
       "      <td>1953.373295</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.6.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>lstm</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>8.068550</td>\n",
       "      <td>34.283509</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.6.0</td>\n",
       "      <td>mpi</td>\n",
       "      <td>transformer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>8.667265</td>\n",
       "      <td>34.391664</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.7.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>transformer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>823.987637</td>\n",
       "      <td>1929.792356</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.7.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>lstm</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>209.991957</td>\n",
       "      <td>753.882725</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.7.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>alexnet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>16.304242</td>\n",
       "      <td>55.684349</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.7.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>resnet50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>46.387229</td>\n",
       "      <td>98.640427</td>\n",
       "      <td>1</td>\n",
       "      <td>hsw</td>\n",
       "      <td>1.7.1</td>\n",
       "      <td>mpi</td>\n",
       "      <td>cnn3d</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    train_rate  inference_rate  ranks hardware version backend        model\n",
       "14   23.892666       46.485593      1      hsw   1.2.0     mpi        cnn3d\n",
       "13   14.632251       37.498390      1      hsw   1.2.0     mpi     resnet50\n",
       "12  213.417174      682.692711      1      hsw   1.2.0     mpi      alexnet\n",
       "11  650.044671     2734.242050      1      hsw   1.2.0     mpi         lstm\n",
       "10    7.608130       28.118234      1      hsw   1.2.0     mpi  transformer\n",
       "0     7.745227       27.457707      1      hsw   1.3.1     mpi  transformer\n",
       "1   644.677564     2610.069286      1      hsw   1.3.1     mpi         lstm\n",
       "2   198.531888      684.354039      1      hsw   1.3.1     mpi      alexnet\n",
       "3    14.996481       46.225346      1      hsw   1.3.1     mpi     resnet50\n",
       "4    24.243660       48.183499      1      hsw   1.3.1     mpi        cnn3d\n",
       "9    14.944277       27.766081      1      hsw   1.6.0     mpi        cnn3d\n",
       "8    16.685013       48.906670      1      hsw   1.6.0     mpi     resnet50\n",
       "7   210.820440      710.939807      1      hsw   1.6.0     mpi      alexnet\n",
       "6   834.144656     1953.373295      1      hsw   1.6.0     mpi         lstm\n",
       "5     8.068550       34.283509      1      hsw   1.6.0     mpi  transformer\n",
       "15    8.667265       34.391664      1      hsw   1.7.1     mpi  transformer\n",
       "16  823.987637     1929.792356      1      hsw   1.7.1     mpi         lstm\n",
       "17  209.991957      753.882725      1      hsw   1.7.1     mpi      alexnet\n",
       "18   16.304242       55.684349      1      hsw   1.7.1     mpi     resnet50\n",
       "19   46.387229       98.640427      1      hsw   1.7.1     mpi        cnn3d"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = [pd.read_csv(f, delim_whitespace=True) for f in results_files]\n",
    "results = pd.concat(results, ignore_index=True)\n",
    "results.version = results.version.apply(lambda s: s.lstrip('v'))\n",
    "results = results.sort_values('version')\n",
    "\n",
    "results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training throughput comparisons"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "models = results.model.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(nrows=2, ncols=3, figsize=(18, 10))\n",
    "axs = axs.flatten()\n",
    "\n",
    "for i, model in enumerate(models):\n",
    "    ax = axs[i]\n",
    "    model_results = results[results.model == model].sort_values('version')\n",
    "    ax.bar(model_results.version, model_results.train_rate)\n",
    "    ax.set_title(model)\n",
    "    ax.set_ylabel('samples / s')\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inference throughput comparisons"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(nrows=2, ncols=3, figsize=(18, 10))\n",
    "axs = axs.flatten()\n",
    "\n",
    "for i, model in enumerate(models):\n",
    "    ax = axs[i]\n",
    "    model_results = results[results.model == model].sort_values('version')\n",
    "    ax.bar(model_results.version, model_results.inference_rate)\n",
    "    ax.set_title(model)\n",
    "    ax.set_ylabel('samples / s')\n",
    "\n",
    "plt.tight_layout()"
   ]
  }
 ],
 "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": "3.7.4"
  },
  "toc-autonumbering": false
 },
 "nbformat": 4,
 "nbformat_minor": 4
}