numpy_savetxt_benchmark.ipynb

numpy_savetxt_benchmark.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "numpy_savetxt_benchmark.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMPxexvXmwpxY1xTOJoStQJ",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/rjodon/2f05a60825f9bf885faf359be66d9a73/numpy_savetxt_benchmark.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "syeOp7pHMtan"
      },
      "source": [
        "\n",
        "\"\"\"\n",
        "Small benchmark of numpy.savetxt() vs python io.open + io.write\n",
        "\n",
        "* I am using time plus an explicit loop to profile the snippet\n",
        "* Why? timeit can't control python caches here. Thus, results would not reflect reality.\n",
        "\n",
        "* Savetxt is highly inefficient for arrays of dim>500. It is browsing row per row,\n",
        "  formatting, then writing to a file.\n",
        "\n",
        "\"\"\"\n",
        "import numpy\n",
        "\n",
        "nc = 1; nb = 20; ni = 6; nc = 2; ia = 20; ib = 20; ic = 0\n",
        "scale_factor = 10000\n",
        "\n",
        "U1 = numpy.array(range(0,    scale_factor))\n",
        "U2 = numpy.array(range(scale_factor,   2*scale_factor))\n",
        "U3 = numpy.array(range(3*scale_factor, 4*scale_factor))\n",
        "U4 = numpy.array(range(4*scale_factor, 5*scale_factor))\n",
        "U5 = numpy.array(range(5*scale_factor, 6*scale_factor))\n",
        "\n",
        "a = nc*(nb*nc*ia+nc*ib+ic)+U1\n",
        "a2 = ia + U1\n",
        "b2 = ib + U3\n",
        "c2 = ic + U4\n",
        "b = nc * (nb * nc * a2 + nc * b2 + c2) + U2\n",
        "A = numpy.array((a, b, U5)).T"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "K02_PyvmMyMP",
        "outputId": "deb14a30-cadc-4f71-bb04-a38f84167519"
      },
      "source": [
        "# NUMPY SAVETXT\n",
        "%%time\n",
        "for i in range(100):\n",
        "  numpy.savetxt(\"savetxt.txt\", A)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 3.41 s, sys: 128 ms, total: 3.54 s\n",
            "Wall time: 3.61 s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1Jdsx3bcM3uw",
        "outputId": "cd49f477-4be2-4f13-d001-dfb467cecc00"
      },
      "source": [
        "# Builtin IO\n",
        "%%time\n",
        "with open(file=\"write.txt\", mode=\"w\") as b:\n",
        "  for i in range(100):\n",
        "    b.write(numpy.array2string(A))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 15.1 ms, sys: 825 µs, total: 16 ms\n",
            "Wall time: 19.5 ms\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "10dIHCkYNM_c",
        "outputId": "df6c88bc-03ae-422c-9fd2-a1971a3ef1c1"
      },
      "source": [
        "# Builtin IO + opening/closing each iteration\n",
        "%%time\n",
        "for i in range(100):\n",
        "  with open(file=\"write.txt\", mode=\"w\") as b:\n",
        "    b.write(numpy.array2string(A))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 22.7 ms, sys: 3.12 ms, total: 25.8 ms\n",
            "Wall time: 31.8 ms\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Bk1lB4dvN4g1"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}