{
 "cells": [
  {
   "metadata": {
    "trusted": true,
    "collapsed": false
   },
   "cell_type": "code",
   "source": "from IPython.parallel import Client\nc = Client(profile='parallel_python2.7')\ndview = c.direct_view()\nlview = c.load_balanced_view()",
   "execution_count": 16,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": false,
    "scrolled": true
   },
   "cell_type": "code",
   "source": "from matplotlib import pyplot as plt\n%matplotlib inline\nwith dview.sync_imports():\n    import numpy",
   "execution_count": 21,
   "outputs": [
    {
     "output_type": "stream",
     "text": "importing numpy on engine(s)\n",
     "name": "stdout"
    }
   ]
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": false
   },
   "cell_type": "code",
   "source": "%%px\nprint 'Hello from engines'",
   "execution_count": 18,
   "outputs": [
    {
     "output_type": "stream",
     "text": "[stdout:0] Hello from engines\n[stdout:1] Hello from engines\n[stdout:2] Hello from engines\n[stdout:3] Hello from engines\n",
     "name": "stdout"
    }
   ]
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": false
   },
   "cell_type": "code",
   "source": "%%px --local\nprint 'Hello from engines AND client'",
   "execution_count": 5,
   "outputs": [
    {
     "output_type": "stream",
     "text": "Hello from engines AND client\n[stdout:0] Hello from engines AND client\n[stdout:1] Hello from engines AND client\n[stdout:2] Hello from engines AND client\n[stdout:3] Hello from engines AND client\n",
     "name": "stdout"
    }
   ]
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": false
   },
   "cell_type": "code",
   "source": "# create a work array\nw_full = numpy.random.rand(10000)\ndview.scatter('w', w_full)",
   "execution_count": 19,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": "<AsyncResult: finished>"
     },
     "metadata": {},
     "execution_count": 19
    }
   ]
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": false
   },
   "cell_type": "code",
   "source": "%%px\nprint len(w)\n# work on the array on the engines\nwsq = numpy.square(w)",
   "execution_count": 22,
   "outputs": [
    {
     "output_type": "stream",
     "text": "[stdout:0] 2500\n[stdout:1] 2500\n[stdout:2] 2500\n[stdout:3] 2500\n",
     "name": "stdout"
    }
   ]
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": false
   },
   "cell_type": "code",
   "source": "# get the result back on the client\nwsq_full = dview.gather('wsq').get()\n# check result\nnumpy.testing.assert_allclose(w_full**2, wsq_full)",
   "execution_count": 23,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": true
   },
   "cell_type": "code",
   "source": "",
   "execution_count": null,
   "outputs": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "name": "python2",
   "display_name": "Python 2.7",
   "language": "python"
  },
  "language_info": {
   "mimetype": "text/x-python",
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "name": "python",
   "file_extension": ".py",
   "pygments_lexer": "ipython2",
   "nbconvert_exporter": "python",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}