TensorFlow with GPU

tensorflow-with-gpu.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "TensorFlow with GPU",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 2",
      "name": "python2"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "[View in Colaboratory](https://colab.research.google.com/gist/danielpcox/36b5f234de24b57c65bac86108d3b076/tensorflow-with-gpu.ipynb)"
      ]
    },
    {
      "metadata": {
        "id": "BlmQIFSLZDdc",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Confirm TensorFlow can see the GPU\n",
        "\n",
        "Simply select \"GPU\" in the Accelerator drop-down in Notebook Settings (either through the Edit menu or the command palette at cmd/ctrl-shift-P)."
      ]
    },
    {
      "metadata": {
        "id": "3IEVK-KFxi5Z",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "0940e8b1-5283-4b2d-e3e6-8853f3deb804"
      },
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "device_name = tf.test.gpu_device_name()\n",
        "if device_name != '/device:GPU:0':\n",
        "  raise SystemError('GPU device not found')\n",
        "print('Found GPU at: {}'.format(device_name))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found GPU at: /device:GPU:0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "QXRh0DPiZRyG",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Observe TensorFlow speedup on GPU relative to CPU\n",
        "\n",
        "This example constructs a typical convolutional neural network layer over a\n",
        "random image and manually places the resulting ops on either the CPU or the GPU\n",
        "to compare execution speed."
      ]
    },
    {
      "metadata": {
        "id": "t9ALbbpmY9rm",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        },
        "outputId": "c5ee74e3-f88b-4660-dcba-354d09fa5659"
      },
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import timeit\n",
        "\n",
        "# See https://www.tensorflow.org/tutorials/using_gpu#allowing_gpu_memory_growth\n",
        "config = tf.ConfigProto()\n",
        "config.gpu_options.allow_growth = True\n",
        "\n",
        "with tf.device('/cpu:0'):\n",
        "  random_image_cpu = tf.random_normal((100, 100, 100, 3))\n",
        "  net_cpu = tf.layers.conv2d(random_image_cpu, 32, 7)\n",
        "  net_cpu = tf.reduce_sum(net_cpu)\n",
        "\n",
        "with tf.device('/gpu:0'):\n",
        "  random_image_gpu = tf.random_normal((100, 100, 100, 3))\n",
        "  net_gpu = tf.layers.conv2d(random_image_gpu, 32, 7)\n",
        "  net_gpu = tf.reduce_sum(net_gpu)\n",
        "\n",
        "sess = tf.Session(config=config)\n",
        "\n",
        "# Test execution once to detect errors early.\n",
        "try:\n",
        "  sess.run(tf.global_variables_initializer())\n",
        "except tf.errors.InvalidArgumentError:\n",
        "  print(\n",
        "      '\\n\\nThis error most likely means that this notebook is not '\n",
        "      'configured to use a GPU.  Change this in Notebook Settings via the '\n",
        "      'command palette (cmd/ctrl-shift-P) or the Edit menu.\\n\\n')\n",
        "  raise\n",
        "\n",
        "def cpu():\n",
        "  sess.run(net_cpu)\n",
        "  \n",
        "def gpu():\n",
        "  sess.run(net_gpu)\n",
        "  \n",
        "# Runs the op several times.\n",
        "print('Time (s) to convolve 32x7x7x3 filter over random 100x100x100x3 images '\n",
        "      '(batch x height x width x channel). Sum of ten runs.')\n",
        "print('CPU (s):')\n",
        "cpu_time = timeit.timeit('cpu()', number=10, setup=\"from __main__ import cpu\")\n",
        "print(cpu_time)\n",
        "print('GPU (s):')\n",
        "gpu_time = timeit.timeit('gpu()', number=10, setup=\"from __main__ import gpu\")\n",
        "print(gpu_time)\n",
        "print('GPU speedup over CPU: {}x'.format(int(cpu_time/gpu_time)))\n",
        "\n",
        "sess.close()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Time (s) to convolve 32x7x7x3 filter over random 100x100x100x3 images (batch x height x width x channel). Sum of ten runs.\n",
            "CPU (s):\n",
            "8.350230318000058\n",
            "GPU (s):\n",
            "0.1842791589999706\n",
            "GPU speedup over CPU: 45x\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}