benchmark_svd.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting matplotlib\n",
      "  Using cached https://files.pythonhosted.org/packages/da/83/d989ee20c78117c737ab40e0318ea221f1aed4e3f5a40b4f93541b369b93/matplotlib-3.1.0-cp36-cp36m-manylinux1_x86_64.whl\n",
      "Collecting cycler>=0.10 (from matplotlib)\n",
      "  Using cached https://files.pythonhosted.org/packages/f7/d2/e07d3ebb2bd7af696440ce7e754c59dd546ffe1bbe732c8ab68b9c834e61/cycler-0.10.0-py2.py3-none-any.whl\n",
      "Requirement already satisfied: python-dateutil>=2.1 in /home/filter/anaconda3/envs/hy3/lib/python3.6/site-packages (from matplotlib) (2.8.0)\n",
      "Collecting kiwisolver>=1.0.1 (from matplotlib)\n",
      "  Using cached https://files.pythonhosted.org/packages/f8/a1/5742b56282449b1c0968197f63eae486eca2c35dcd334bab75ad524e0de1/kiwisolver-1.1.0-cp36-cp36m-manylinux1_x86_64.whl\n",
      "Collecting pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 (from matplotlib)\n",
      "  Using cached https://files.pythonhosted.org/packages/dd/d9/3ec19e966301a6e25769976999bd7bbe552016f0d32b577dc9d63d2e0c49/pyparsing-2.4.0-py2.py3-none-any.whl\n",
      "Requirement already satisfied: numpy>=1.11 in /home/filter/anaconda3/envs/hy3/lib/python3.6/site-packages (from matplotlib) (1.16.4)\n",
      "Requirement already satisfied: six in /home/filter/anaconda3/envs/hy3/lib/python3.6/site-packages (from cycler>=0.10->matplotlib) (1.12.0)\n",
      "Requirement already satisfied: setuptools in /home/filter/anaconda3/envs/hy3/lib/python3.6/site-packages (from kiwisolver>=1.0.1->matplotlib) (41.0.1)\n",
      "Installing collected packages: cycler, kiwisolver, pyparsing, matplotlib\n",
      "Successfully installed cycler-0.10.0 kiwisolver-1.1.0 matplotlib-3.1.0 pyparsing-2.4.0\n"
     ]
    }
   ],
   "source": [
    "! pip install matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "blas_mkl_info:\n",
      "  NOT AVAILABLE\n",
      "blis_info:\n",
      "    libraries = ['blas', 'cblas', 'lapack', 'blas', 'cblas', 'lapack']\n",
      "    library_dirs = ['/home/filter/anaconda3/envs/hy3/lib']\n",
      "    define_macros = [('HAVE_CBLAS', None)]\n",
      "    include_dirs = ['/home/filter/anaconda3/envs/hy3/include']\n",
      "    language = c\n",
      "blas_opt_info:\n",
      "    libraries = ['blas', 'cblas', 'lapack', 'blas', 'cblas', 'lapack', 'blas', 'cblas', 'lapack']\n",
      "    library_dirs = ['/home/filter/anaconda3/envs/hy3/lib']\n",
      "    define_macros = [('HAVE_CBLAS', None)]\n",
      "    include_dirs = ['/home/filter/anaconda3/envs/hy3/include']\n",
      "    language = c\n",
      "lapack_mkl_info:\n",
      "  NOT AVAILABLE\n",
      "openblas_lapack_info:\n",
      "    libraries = ['blas', 'cblas', 'lapack', 'blas', 'cblas', 'lapack']\n",
      "    library_dirs = ['/home/filter/anaconda3/envs/hy3/lib']\n",
      "    language = c\n",
      "    define_macros = [('HAVE_CBLAS', None)]\n",
      "lapack_opt_info:\n",
      "    libraries = ['blas', 'cblas', 'lapack', 'blas', 'cblas', 'lapack', 'blas', 'cblas', 'lapack']\n",
      "    library_dirs = ['/home/filter/anaconda3/envs/hy3/lib']\n",
      "    language = c\n",
      "    define_macros = [('HAVE_CBLAS', None)]\n"
     ]
    }
   ],
   "source": [
    "# https://simplyml.com/benchmarking-the-singular-value-decomposition/\n",
    "# https://github.com/charanpald/tyre-hug/blob/master/tyrehug/exp/svdbenchmark.py\n",
    "\n",
    "import time\n",
    "import numpy\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import scipy.sparse\n",
    "import numpy as np\n",
    "\n",
    "from scipy.sparse.linalg import svds\n",
    "# from sparsesvd import sparsesvd\n",
    "from sklearn.utils.extmath import randomized_svd\n",
    "from gensim.models.lsimodel import stochastic_svd\n",
    "\n",
    "np.show_config()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def time_reps(func, params, reps):\n",
    "    start = time.time()\n",
    "    for s in range(reps):\n",
    "        func(*params)\n",
    "    print(\"Executed \" + str(func))\n",
    "    return (time.time() - start) / reps\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "Compare and contrast different methods of computing the SVD\n",
    "\"\"\"\n",
    "\n",
    "k = 50\n",
    "p = 0\n",
    "reps = 5\n",
    "density = 10**-3\n",
    "n_iter = 5\n",
    "var_range = numpy.arange(1, 11)\n",
    "\n",
    "\n",
    "def gensim_svd(A, k):\n",
    "    return stochastic_svd(A, k, A.shape[0], power_iters=n_iter, extra_dims=10)\n",
    "\n",
    "\n",
    "def time_densities():\n",
    "    n = 10**4\n",
    "    densities = var_range * 2 * 10**-3\n",
    "    times = numpy.zeros((4, densities.shape[0]))\n",
    "\n",
    "    for i, density in enumerate(densities):\n",
    "        # Generate random sparse matrix\n",
    "        inds = numpy.random.randint(int(n), size=(2, int(n * n * density)))\n",
    "        data = numpy.random.rand(int(n * n * density))\n",
    "        A = scipy.sparse.csc_matrix((data, inds), (n, n))\n",
    "        print(A.shape, A.nnz)\n",
    "        \n",
    "        times[3, i] = time_reps(gensim_svd, (A, k), reps)\n",
    "        times[2, i] = time_reps(randomized_svd, (A, k, 10, n_iter), reps)\n",
    "        times[0, i] = time_reps(svds, (A, k), reps)\n",
    "#         times[1, i] = time_reps(sparsesvd, (A, k), reps)\n",
    "        print(n, density, times[:, i])\n",
    "\n",
    "    plt.figure(0)\n",
    "#     plt.plot(densities, times[1, :], 'k-', label=\"sparsesvd\")\n",
    "    plt.plot(densities, times[0, :], 'b-', label=\"scipy\")\n",
    "    plt.plot(densities, times[2, :], 'b--', label=\"sklearn (random)\")\n",
    "    plt.plot(densities, times[3, :], 'k--', label=\"Gensim SVD (random)\")\n",
    "\n",
    "    plt.legend(loc=\"upper left\")\n",
    "    plt.xlabel(\"density\")\n",
    "    plt.ylabel(\"time (s)\")\n",
    "    plt.savefig(\"time_densities.png\", format=\"png\")\n",
    "\n",
    "\n",
    "# Next, vary the matrix size\n",
    "def time_ns():\n",
    "    density = 10**-3\n",
    "    ns = var_range * 10**4\n",
    "    times = numpy.zeros((4, ns.shape[0]))\n",
    "\n",
    "    for i, n in enumerate(ns):\n",
    "        # Generate random sparse matrix\n",
    "        inds = numpy.random.randint(int(n), size=(2, int(n * n * density)))\n",
    "        data = numpy.random.rand(int(n * n * density))\n",
    "        A = scipy.sparse.csc_matrix((data, inds), (n, n))\n",
    "        print(A.shape, A.nnz)\n",
    "        \n",
    "        times[3, i] = time_reps(gensim_svd, (A, k), reps)\n",
    "        times[0, i] = time_reps(svds, (A, k), reps)\n",
    "#         times[1, i] = time_reps(sparsesvd, (A, k), reps)\n",
    "        times[2, i] = time_reps(randomized_svd, (A, k, 10, n_iter), reps)\n",
    "        print(n, density, times[:, i])\n",
    "\n",
    "    plt.figure(1)\n",
    "#     plt.plot(ns, times[1, :], 'k-', label=\"sparsesvd\")\n",
    "    plt.plot(ns, times[0, :], 'b-', label=\"scipy\")\n",
    "    plt.plot(ns, times[2, :], 'b--', label=\"sklearn (random)\")\n",
    "    plt.plot(ns, times[3, :], 'k--', label=\"Gensim SVD (random)\")\n",
    "\n",
    "    plt.legend(loc=\"upper left\")\n",
    "    plt.xlabel(\"n\")\n",
    "    plt.ylabel(\"time (s)\")\n",
    "    plt.savefig(\"time_ns.png\", format=\"png\")\n",
    "\n",
    "\n",
    "def time_ks():\n",
    "    n = 10**4\n",
    "    density = 10**-3\n",
    "    ks = var_range * 20\n",
    "    times = numpy.zeros((4, ks.shape[0]))\n",
    "\n",
    "    for i, k in enumerate(ks):\n",
    "        # Generate random sparse matrix\n",
    "        inds = numpy.random.randint(int(n), size=(2, int(n * n * density)))\n",
    "        data = numpy.random.rand(int(n * n * density))\n",
    "        A = scipy.sparse.csc_matrix((data, inds), (n, n))\n",
    "        print(A.shape, A.nnz)\n",
    "        \n",
    "        times[3, i] = time_reps(gensim_svd, (A, k), reps)\n",
    "        times[0, i] = time_reps(svds, (A, k), reps)\n",
    "#         times[1, i] = time_reps(sparsesvd, (A, k), reps)\n",
    "        times[2, i] = time_reps(randomized_svd, (A, k, 10, n_iter), reps)\n",
    "        print(n, density, times[:, i])\n",
    "\n",
    "    plt.figure(2)\n",
    "#     plt.plot(ks, times[1, :], 'k-', label=\"sparsesvd\")\n",
    "    plt.plot(ks, times[0, :], 'b-', label=\"scipy\")\n",
    "    plt.plot(ks, times[2, :], 'b--', label=\"sklearn (random)\")\n",
    "    plt.plot(ks, times[3, :], 'k--', label=\"Gensim SVD (random)\")\n",
    "\n",
    "    plt.legend(loc=\"upper left\")\n",
    "    plt.xlabel(\"k\")\n",
    "    plt.ylabel(\"time (s)\")\n",
    "    plt.savefig(\"time_ks.png\", format=\"png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10000, 10000) 199823\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.002 [0.72477307 0.         0.2160953  0.22866769]\n",
      "(10000, 10000) 399242\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.004 [1.04061861 0.         0.2940403  0.28254018]\n",
      "(10000, 10000) 598251\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.006 [1.35831065 0.         0.36506567 0.34413681]\n",
      "(10000, 10000) 796816\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.008 [1.81681256 0.         0.44465284 0.44522762]\n",
      "(10000, 10000) 994973\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.01 [2.06747823 0.         0.53944216 0.51096578]\n",
      "(10000, 10000) 1192820\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.012 [2.24017401 0.         0.61882606 0.57838449]\n",
      "(10000, 10000) 1390118\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.014 [2.61627688 0.         0.68596401 0.65995069]\n",
      "(10000, 10000) 1587276\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.016 [2.94792542 0.         0.77650604 0.76668463]\n",
      "(10000, 10000) 1783829\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.018000000000000002 [3.27924042 0.         0.82667274 0.79936924]\n",
      "(10000, 10000) 1980475\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "10000 0.02 [3.53995748 0.         0.93633451 0.88032441]\n"
     ]
    },
    {
     "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": [
    "time_densities()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10000, 10000) 99963\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [0.56332397 0.         0.16327772 0.15186725]\n",
      "(20000, 20000) 399773\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "20000 0.001 [2.48298349 0.         0.60912576 0.61790543]\n",
      "(30000, 30000) 899527\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "30000 0.001 [5.60239406 0.         1.39338412 1.45393682]\n",
      "(40000, 40000) 1599206\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "40000 0.001 [11.30199447  0.          2.41360564  2.59306173]\n",
      "(50000, 50000) 2498782\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "50000 0.001 [18.23549981  0.          3.71365128  3.94062538]\n",
      "(60000, 60000) 3598270\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "60000 0.001 [27.46179085  0.          5.31994319  5.60144806]\n",
      "(70000, 70000) 4897497\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "70000 0.001 [39.44316697  0.          7.91870627  7.98339124]\n",
      "(80000, 80000) 6396757\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "80000 0.001 [52.91148567  0.          9.46508608  9.86497064]\n",
      "(90000, 90000) 8095889\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "90000 0.001 [68.63566403  0.         12.31226573 12.64846206]\n",
      "(100000, 100000) 9995052\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "100000 0.001 [87.18712888  0.         15.5757864  15.78205442]\n"
     ]
    },
    {
     "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": [
    "time_ns()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10000, 10000) 99950\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [0.1628334  0.         0.06912408 0.06188965]\n",
      "(10000, 10000) 99953\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [0.37151818 0.         0.12871561 0.11909833]\n",
      "(10000, 10000) 99951\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [0.77775002 0.         0.19969969 0.19150925]\n",
      "(10000, 10000) 99952\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [1.278546   0.         0.29540129 0.30497289]\n",
      "(10000, 10000) 99951\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [1.99425578 0.         0.40793977 0.45370545]\n",
      "(10000, 10000) 99957\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [2.51300855 0.         0.47944155 0.49271722]\n",
      "(10000, 10000) 99963\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [3.44795585 0.         0.60113568 0.680761  ]\n",
      "(10000, 10000) 99958\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [4.17465258 0.         0.66062264 0.69933505]\n",
      "(10000, 10000) 99953\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [5.27612076 0.         0.77981496 0.89439149]\n",
      "(10000, 10000) 99950\n",
      "Executed <function gensim_svd at 0x7f4e4239ff28>\n",
      "Executed <function svds at 0x7f4e443db598>\n",
      "Executed <function randomized_svd at 0x7f4e43efe048>\n",
      "10000 0.001 [5.92155252 0.         0.86180668 0.9026978 ]\n"
     ]
    },
    {
     "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": [
    "time_ks()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.show()"
   ]
  },
  {
   "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}