fld_model_figures.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# FlexDriver performance model\n",
    "\n",
    "Run the performance model of FlexDriver to recreate the charts of the paper."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%config InlineBackend.figure_format = 'retina'\n",
    "\n",
    "from common import *\n",
    "from fld_model import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "fld_model_df = pd.concat(fld_model(**kwargs) for kwargs in [\n",
    "    { 'model': '25GbE/PCIe x8', 'linerate': 25, 'pci_lanes': 8, }, \n",
    "    { 'model': '50GbE/PCIe x8', 'linerate': 50, 'pci_lanes': 8, }, \n",
    "    { 'model': '100GbE/PCIe x16', 'linerate': 100, 'pci_lanes': 16 },\n",
    "])\n",
    "fld_model_df.model = pd.Categorical(fld_model_df.model).reorder_categories(\n",
    "    ['25GbE/PCIe x8', '50GbE/PCIe x8', '100GbE/PCIe x16'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>pci_outbound_control_overhead</th>\n",
       "      <th>model</th>\n",
       "      <th>pci_expected_bandwidth</th>\n",
       "      <th>pci_packet_rate</th>\n",
       "      <th>ethernet_linerate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>88</td>\n",
       "      <td>25GbE/PCIe x8</td>\n",
       "      <td>46.556279</td>\n",
       "      <td>5.348837</td>\n",
       "      <td>24.548736</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>88</td>\n",
       "      <td>50GbE/PCIe x8</td>\n",
       "      <td>46.556279</td>\n",
       "      <td>5.348837</td>\n",
       "      <td>49.097473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>88</td>\n",
       "      <td>100GbE/PCIe x16</td>\n",
       "      <td>93.891091</td>\n",
       "      <td>10.787120</td>\n",
       "      <td>98.194946</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     pci_outbound_control_overhead            model  pci_expected_bandwidth  \\\n",
       "271                             88    25GbE/PCIe x8               46.556279   \n",
       "271                             88    50GbE/PCIe x8               46.556279   \n",
       "271                             88  100GbE/PCIe x16               93.891091   \n",
       "\n",
       "     pci_packet_rate  ethernet_linerate  \n",
       "271         5.348837          24.548736  \n",
       "271         5.348837          49.097473  \n",
       "271        10.787120          98.194946  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fld_model_df[fld_model_df['size'] == 1088][[\n",
    "    'pci_outbound_control_overhead', 'model', 'pci_expected_bandwidth',\n",
    "    'pci_packet_rate', 'ethernet_linerate']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/haggai/miniconda3/envs/flc-model/lib/python3.9/site-packages/plotnine/geoms/geom_path.py:75: PlotnineWarning: geom_path: Removed 1994 rows containing missing values.\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 287.566x130 with 3 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 170,
       "width": 282
      }
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (8770106130978)>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "value_vars = ['ethernet_linerate', 'pci_expected_bandwidth']\n",
    "value_var_names = ['Eth', 'PCIe']\n",
    "melted = (fld_model_df\n",
    "          .melt(id_vars=['size', 'model', 'linerate'], value_vars=value_vars)\n",
    "          .sort_values('linerate')\n",
    "         )\n",
    "melted.variable.replace(dict(zip(value_vars, value_var_names)), inplace=True)\n",
    "melted.variable = melted.variable.astype(pd.CategoricalDtype()).cat.reorder_categories(value_var_names[:min(len(value_vars), len(value_var_names))])\n",
    "plt = p9.ggplot(melted,\n",
    "                p9.aes(x='size', y='value', color='variable', linetype='variable')) + \\\n",
    "    p9.geom_line() + \\\n",
    "    p9.scale_x_continuous(name='packet size [byte]', limits=(4, 1024), breaks=[512 * i for i in range(1, 5)]) + \\\n",
    "    p9.scale_y_continuous(name='bandwidth [Gbps]', limits=(0, None)) + \\\n",
    "    top_legend + \\\n",
    "    p9.theme(figure_size=(fig_width-0.5, performance_model_height),\n",
    "             subplots_adjust={'top': 0.8}) + \\\n",
    "    p9.coord_cartesian(ylim=(0, None)) + \\\n",
    "    p9.facet_wrap('model') + \\\n",
    "    eth_pci_scale\n",
    "plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/haggai/miniconda3/envs/flc-model/lib/python3.9/site-packages/plotnine/ggplot.py:719: PlotnineWarning: Saving 2.87566 x 1.3 in image.\n",
      "/home/haggai/miniconda3/envs/flc-model/lib/python3.9/site-packages/plotnine/ggplot.py:722: PlotnineWarning: Filename: fld_model_breakdown.pdf\n",
      "/home/haggai/miniconda3/envs/flc-model/lib/python3.9/site-packages/plotnine/geoms/geom_path.py:75: PlotnineWarning: geom_path: Removed 1994 rows containing missing values.\n"
     ]
    }
   ],
   "source": [
    "plt.save('fld_model_breakdown.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = fld_model_df[fld_model_df['size'] == 512][['linerate', 'bandwidth', 'ethernet_linerate']].set_index('linerate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "linerate\n",
       "25     1.000000\n",
       "50     0.907016\n",
       "100    0.914600\n",
       "dtype: float64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['bandwidth'] / df['ethernet_linerate']"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "c6e5e40b5d548b459b1e0d7764dd601243dd92169d54f8c16faf73bff9192545"
  },
  "kernelspec": {
   "display_name": "Python 3.8.8 64-bit ('base': conda)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}