notebooks/03_Scaling_Python.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:a49ce555940008339be5252f7fd84f040fd839888e4dca98eccd5f88f59735ed"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "<style>\n",
      "    h1{\n",
      "        text-align:center;\n",
      "    }\n",
      "    h2{\n",
      "        text-align:center;\n",
      "        font-variant:small-caps;\n",
      "    }\n",
      "</style>\n",
      "# Scaling Python\n",
      "# Python in HPC Tutorial\n",
      "\n",
      "## Supercomputing 2014\n",
      "\n",
      "Matt Knepley and Aron Ahmadia\n",
      "\n",
      "[![Creative Commons License](../figures/creative_commons_logo.png)](http://creativecommons.org/licenses/by/3.0/deed.en_US)  "
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Acknowledgements\n",
      "\n",
      "* [mpi4py](http://code.google.com/p/mpi4py/) is a [Cythonized](http://www.cython.org/) wrapper around [MPI](http://www.mpi-forum.org/) originally developed by [Lisandro Dalcin](http://plus.google.com/107621373684536061961/about), CONICET\n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Slideshow Setup Code"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%matplotlib inline"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "-"
      }
     },
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## Connecting to MPI via IPython\n",
      "\n",
      "To run the examples in parallel, you must run the command:\n",
      "\n",
      "`ipcluster start --engines=MPI --n 4` \n",
      "\n",
      "Or use notebook Appendix_03_Launch_MPI_Engines \n",
      "(if you are using the VMs or have this configured for your environment).\n",
      "\n",
      "Note that `ipcluster` is implicitly taking `mpiexec` from your path. This\n",
      "can cause problems on machines which have MPI installed by default, like OSX.\n",
      "\n",
      "Then execute the next cell:"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from IPython.parallel import Client\n",
      "c = Client()\n",
      "view = c[:]\n",
      "\n",
      "%load_ext parallelmagic\n",
      "view.activate()\n",
      "view.block = True"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true
      }
     },
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "\n",
      "The Message Passing Model\n",
      "-------------------------------------\n",
      "\n",
      "* a process is (traditionally) a program counter for instructions and an address space for data\n",
      "* processes may have multiple threads (program counters and associated stacks) sharing a single address space\n",
      "* message passing is for communication among processes, which have separate address spaces\n",
      "* interprocess communication consists of\n",
      "  \n",
      "  - synchronization\n",
      "  - movement of data from one process's address space to another's\n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "\n",
      "Why MPI?\n",
      "-------------------------------------\n",
      "<!--====-->\n",
      "\n",
      "* **communicators** encapsulate communication spaces for library safety\n",
      "* **datatypes** reduce copying costs and permit heterogeneity\n",
      "* multiple **communication modes** allow more control of memory buffer management\n",
      "* extensive **collective operations** for scalable global communication\n",
      "* **process topologies** permit efficient process placement, user views of process layout\n",
      "* **profiling interface** encourages portable tools\n",
      "\n",
      "It Scales!"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "\n",
      "MPI - Quick Review\n",
      "-------------------------------------\n",
      "\n",
      "* processes can be collected into **groups**\n",
      "* each message is sent in a **context**, and must be received in the same context\n",
      "* a  **communicator** encapsulates a context for a specific group\n",
      "* a given program may have many communicators with any level of overlap\n",
      "* two initial communicators\n",
      " \n",
      "  - ``MPI_COMM_WORLD`` (all processes)\n",
      "  - ``MPI_COMM_SELF`` (current process)"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "First Example: Hello World\n",
      "-------------------------------------\n",
      "Ths user will have to install mpi4py\n",
      "```\n",
      "  pip install mpi4py\n",
      "```\n",
      "It is crucial here that the proper `mpiexec` for your MPI installation be in the path.\n",
      "\n",
      "For interactive convenience, we load the parallel magic extensions and make this view\n",
      "the active one for the automatic parallelism magics.\n",
      "\n",
      "This is not necessary and in production codes likely won't be used, as the engines will \n",
      "load their own MPI codes separately.  But it makes it easy to illustrate everything from\n",
      "within a single notebook here."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from IPython.parallel import Client\n",
      "c = Client()\n",
      "view = c[:]\n",
      "\n",
      "%load_ext parallelmagic\n",
      "view.activate()\n",
      "view.block = True"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "fragment"
      }
     },
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "Use the autopx magic to make the rest of these cell execute on the engines instead\n",
      "of locally"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%autopx"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "%autopx enabled\n"
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "fragment"
      }
     },
     "source": [
      "With autopx enabled, the next cell will actually execute *entirely on each engine*:"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_type": "slide"
      }
     },
     "source": [
      "## Hello World"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "\n",
      "size = MPI.COMM_WORLD.Get_size()\n",
      "rank = MPI.COMM_WORLD.Get_rank()\n",
      "name = MPI.Get_processor_name()\n",
      "\n",
      "print(\"Hello World! I am process %d of %d on %s.\\n\" % (rank, size, name))"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:0] \n",
        "Hello World! I am process 1 of 4 on wlan-floor3-1688.sc14.org.\n",
        "\n",
        "[stdout:1] \n",
        "Hello World! I am process 3 of 4 on wlan-floor3-1688.sc14.org.\n",
        "\n",
        "[stdout:2] \n",
        "Hello World! I am process 2 of 4 on wlan-floor3-1688.sc14.org.\n",
        "\n",
        "[stdout:3] \n",
        "Hello World! I am process 0 of 4 on wlan-floor3-1688.sc14.org.\n",
        "\n"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "\n",
      "Functionality\n",
      "-------------------------------------\n",
      "\n",
      "* There are hundreds of functions in the MPI standard, not all of them are necessarily available in MPI4Py, most commonly used are\n",
      "* No need to call MPI_Init() or MPI_Finalize()\n",
      " - MPI_Init() is called when you import the module\n",
      " - MPI_Finalize() is called before the Python process ends\n",
      "* To launch:\n",
      "    mpiexec -n < number of process > -machinefile < hostlist > python < my MPI4Py python script >\n",
      "* IPython automatically handles calling mpiexec for you with the `ipcluster` command"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_step": false,
       "slide_type": "slide"
      }
     },
     "source": [
      "## MPI Basic (Blocking) Send\n",
      "\n",
      "\n",
      "```\n",
      "   int MPI_Send(void* buf, int count, MPI_Datatype type, \n",
      "   int dest, int tag, MPI_Comm comm)\n",
      "```\n",
      "\n",
      "### mpi4py\n",
      "\n",
      "```\n",
      "  Comm.Send(self, buf, int dest=0, int tag=0)\n",
      "  Comm.send(self, obj=None, int dest=0, int tag=0)\n",
      "```"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## MPI Basic (Blocking) Recv\n",
      "\n",
      "```\n",
      "   int MPI_Recv(void* buf, int count, MPI_Datatype type, \n",
      "   int source, int tag, MPI_Comm comm, MPI_Status status)\n",
      "```\n",
      "\n",
      "### mpi4py\n",
      "\n",
      "```\n",
      "  comm.Recv(self, buf, source=0, tag=0, status=None)\n",
      "  comm.recv(self, obj=None, source=0, tag=0, status=None)\n",
      "```"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Send/Receive Example (lowercase convenience methods)"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "\n",
      "comm = MPI.COMM_WORLD\n",
      "rank = comm.Get_rank()\n",
      "\n",
      "if rank == 0:\n",
      "   data = {'a': 7, 'b': 3.14}\n",
      "   comm.send(data, dest=2, tag=11)\n",
      "elif rank == 2:\n",
      "   data = comm.recv(source=0, tag=11)\n",
      "   print data"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:2] {'a': 7, 'b': 3.14}\n"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "Send/Receive Example (MPI API on numpy)\n",
      "-------------------"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "import numpy\n",
      "\n",
      "comm = MPI.COMM_WORLD\n",
      "rank = comm.Get_rank()\n",
      "\n",
      "# pass explicit MPI datatypes\n",
      "if rank == 0:\n",
      "   data = numpy.arange(1000, dtype='i')\n",
      "   comm.Send([data, MPI.INT], dest=1, tag=77)\n",
      "elif rank == 1:\n",
      "   data = numpy.empty(1000, dtype='i')\n",
      "   comm.Recv([data, MPI.INT], source=0, tag=77)\n",
      "\n",
      "# or take advantage of automatic MPI datatype discovery\n",
      "if rank == 0:\n",
      "   data = numpy.arange(10, dtype=numpy.float64)\n",
      "   comm.Send(data, dest=1, tag=13)\n",
      "elif rank == 1:\n",
      "   data = numpy.empty(10, dtype=numpy.float64)\n",
      "   comm.Recv(data, source=0, tag=13)\n",
      "   print data"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:0] [ 0.  1.  2.  3.  4.  5.  6.  7.  8.  9.]\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Synchronization\n",
      "\n",
      "```\n",
      "   int MPI_Barrier(MPI_Comm comm)\n",
      "```\n",
      "\n",
      "### mpi4py\n",
      "\n",
      "```\n",
      "   comm.Barrier(self)\n",
      "   comm.barrier(self)\n",
      "```"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "\n",
      "comm = MPI.COMM_WORLD\n",
      "rank = comm.Get_rank()\n",
      "\n",
      "for r_id in range(comm.Get_size()):\n",
      "    if rank == r_id:\n",
      "        print \"Hello from proc:\", rank\n",
      "    comm.Barrier()\n"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:0] Hello from proc: 1\n",
        "[stdout:1] Hello from proc: 3\n",
        "[stdout:2] Hello from proc: 2\n",
        "[stdout:3] Hello from proc: 0\n"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Notice that MPI rank does not coincide with IPython rank"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Timing and Profiling\n",
      "\n",
      "the elapsed (wall-clock) time between two points in an MPI program can be computed using MPI_Wtime:"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "t1 = MPI.Wtime()\n",
      "t2 = MPI.Wtime()\n",
      "print(\"time elapsed is: %e\\n\" % (t2-t1))"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:0] \n",
        "time elapsed is: 8.388102e-05\n",
        "\n",
        "[stdout:1] \n",
        "time elapsed is: 1.196400e-04\n",
        "\n",
        "[stdout:2] \n",
        "time elapsed is: 6.836900e-05\n",
        "\n",
        "[stdout:3] \n",
        "time elapsed is: 7.039600e-05\n",
        "\n"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Broadcast Example\n",
      "\n",
      "```\n",
      "   int MPI_Bcast(void *buf, int count, MPI_Datatype type, \n",
      "   int root, MPI_Comm comm)\n",
      "```\n",
      "\n",
      "### mpi4py\n",
      "\n",
      "``` \n",
      "  comm.Bcast(self, buf, root=0)\n",
      "  comm.bcast(self, obj=None, root=0)\n",
      "```"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "\n",
      "comm = MPI.COMM_WORLD\n",
      "rank = comm.Get_rank()\n",
      "\n",
      "if rank == 3:\n",
      "   data = {'key1' : [7, 2.72, 2+3j],\n",
      "           'key2' : ( 'abc', 'xyz')}\n",
      "else:\n",
      "   data = None\n",
      "data = comm.bcast(data, root=3)\n",
      "print \"bcast finished and data \\\n",
      " on rank %d is: \\n\" % comm.rank, data\n"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:0] \n",
        "bcast finished and data  on rank 1 is: \n",
        "{'key2': ('abc', 'xyz'), 'key1': [7, 2.72, (2+3j)]}\n",
        "[stdout:1] \n",
        "bcast finished and data  on rank 3 is: \n",
        "{'key2': ('abc', 'xyz'), 'key1': [7, 2.72, (2+3j)]}\n",
        "[stdout:2] \n",
        "bcast finished and data  on rank 2 is: \n",
        "{'key2': ('abc', 'xyz'), 'key1': [7, 2.72, (2+3j)]}\n",
        "[stdout:3] \n",
        "bcast finished and data  on rank 0 is: \n",
        "{'key2': ('abc', 'xyz'), 'key1': [7, 2.72, (2+3j)]}\n"
       ]
      }
     ],
     "prompt_number": 10
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Scatter Example"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "\n",
      "comm = MPI.COMM_WORLD\n",
      "size = comm.Get_size()\n",
      "rank = comm.Get_rank()\n",
      "\n",
      "if rank == 0:\n",
      "   data = [(i+1)**2 for i in range(size)]\n",
      "else:\n",
      "   data = None\n",
      "data = comm.scatter(data, root=0)\n",
      "assert data == (rank+1)**2\n",
      "print \"data on rank %d is: \"%comm.rank, data"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:0] data on rank 1 is:  4\n",
        "[stdout:1] data on rank 3 is:  16\n",
        "[stdout:2] data on rank 2 is:  9\n",
        "[stdout:3] data on rank 0 is:  1\n"
       ]
      }
     ],
     "prompt_number": 11
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Gather (and Barrier) Example"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "\n",
      "comm = MPI.COMM_WORLD\n",
      "size = comm.Get_size()\n",
      "rank = comm.Get_rank()\n",
      "\n",
      "data = (rank+1)**2\n",
      "print \"before gather, data on \\\n",
      "  rank %d is: \"%rank, data\n",
      "\n",
      "comm.Barrier()\n",
      "data = comm.gather(data, root=0)\n",
      "if rank == 0:\n",
      "   for i in range(size):\n",
      "       assert data[i] == (i+1)**2\n",
      "else:\n",
      "   assert data is None\n",
      "print \"data on rank: %d is: \"%rank, data"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:0] \n",
        "before gather, data on   rank 1 is:  4\n",
        "data on rank: 1 is:  None\n",
        "[stdout:1] \n",
        "before gather, data on   rank 3 is:  16\n",
        "data on rank: 3 is:  None\n",
        "[stdout:2] \n",
        "before gather, data on   rank 2 is:  9\n",
        "data on rank: 2 is:  None\n",
        "[stdout:3] \n",
        "before gather, data on   rank 0 is:  1\n",
        "data on rank: 0 is:  [1, 4, 9, 16]\n"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Collective Examples"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "## Reduce Example"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "comm = MPI.COMM_WORLD\n",
      "\n",
      "sendmsg = comm.rank\n",
      "recvmsg1 = comm.reduce(sendmsg, op=MPI.SUM, root=0)\n",
      "recvmsg2 = comm.allreduce(sendmsg)\n",
      "\n",
      "print recvmsg2"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:0] 6\n",
        "[stdout:1] 6\n",
        "[stdout:2] 6\n",
        "[stdout:3] 6\n"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "Compute Pi Example\n",
      "-------------------\n",
      "<!--====-->\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mpi4py import MPI\n",
      "import math\n",
      "\n",
      "def compute_pi(n, start=0, step=1):\n",
      "    h = 1.0 / n\n",
      "    s = 0.0\n",
      "    for i in range(start, n, step):\n",
      "        x = h * (i + 0.5)\n",
      "        s += 4.0 / (1.0 + x**2)\n",
      "    return s * h\n",
      "\n",
      "comm = MPI.COMM_WORLD\n",
      "nprocs = comm.Get_size()\n",
      "myrank = comm.Get_rank()\n",
      "if myrank == 0:\n",
      "    n = 40\n",
      "else:\n",
      "    n = None\n",
      "\n",
      "n = comm.bcast(n, root=0)\n",
      "\n",
      "mypi = compute_pi(n, myrank, nprocs)\n",
      "\n",
      "pi = comm.reduce(mypi, op=MPI.SUM, root=0)\n",
      "\n",
      "if myrank == 0:\n",
      "    error = abs(pi - math.pi)\n",
      "    print (\"pi is approximately %.16f, error is %.16f\" % (pi, error))"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "subslide"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[stdout:3] pi is approximately 3.1416447369226570, error is 0.0000520833328639\n"
       ]
      }
     ],
     "prompt_number": 14
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "Mandelbrot Set Example\n",
      "-------------------\n",
      "<!--====-->\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def mandelbrot (x, y, maxit):\n",
      "    c = x + y*1j\n",
      "    z = 0 + 0j\n",
      "    it = 0\n",
      "    while abs(z) < 2 and it < maxit:\n",
      "        z = z**2 + c\n",
      "        it += 1\n",
      "    return it\n",
      "\n",
      "x1, x2 = -2.0, 1.0\n",
      "y1, y2 = -1.0, 1.0\n",
      "w, h = 150, 100\n",
      "maxit = 127\n",
      "\n",
      "from mpi4py import MPI\n",
      "import numpy\n",
      "  \n",
      "comm = MPI.COMM_WORLD\n",
      "size = comm.Get_size()\n",
      "rank = comm.Get_rank()\n",
      "\n",
      "# number of rows to compute here\n",
      "N = h // size + (h % size > rank)\n",
      "\n",
      "# first row to compute here\n",
      "start = comm.scan(N)-N\n",
      "\n",
      "# array to store local result\n",
      "Cl = numpy.zeros([N, w], dtype='i')\n",
      "\n",
      "# compute owned rows\n",
      "\n",
      "dx = (x2 - x1) / w\n",
      "dy = (y2 - y1) / h\n",
      "\n",
      "for i in range(N):\n",
      "    y = y1 + (i + start) * dy\n",
      "    for j in range(w):\n",
      "        x = x1 + j * dx\n",
      "        Cl[i, j] = mandelbrot(x, y, maxit)\n",
      "\n",
      "# gather results at root (process 0)\n",
      "counts = comm.gather(N, root=0)\n",
      "C = None\n",
      "if rank == 0:\n",
      "    C = numpy.zeros([h, w], dtype='i')\n",
      "\n",
      "rowtype = MPI.INT.Create_contiguous(w)\n",
      "rowtype.Commit()\n",
      "\n",
      "comm.Gatherv(sendbuf=[Cl, MPI.INT], recvbuf=[C, (counts, None), rowtype],root=0)\n",
      "\n",
      "rowtype.Free()"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_step": true,
       "slide_type": "fragment"
      }
     },
     "outputs": [],
     "prompt_number": 15
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "source": [
      "We now need to \"pull\" the C array for plotting so we disable autopx. Make sure to re-enable it later on"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%autopx"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "%autopx disabled\n"
       ]
      }
     ],
     "prompt_number": 16
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "view['rank']"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 17,
       "text": [
        "[1, 3, 2, 0]"
       ]
      }
     ],
     "prompt_number": 17
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# CC is an array of C from all ranks, so we use CC[0]\n",
      "CC = view['C']\n",
      "ranks = view['rank']\n",
      "\n",
      "# Do the plotting\n",
      "from matplotlib import pyplot\n",
      "# Some magic to get to MPI4PY rank 0, not necessarily engine rank 0\n",
      "pyplot.imshow(CC[ranks.index(0)], aspect='equal')\n",
      "pyplot.spectral()\n",
      "pyplot.show()"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAAD8CAYAAABTjp5OAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCRsE/x/poJUtBGjQUQAAAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x1040bab50>"
       ]
      }
     ],
     "prompt_number": 18
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "source": [
      "Toggle autopx back"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%autopx"
     ],
     "language": "python",
     "metadata": {
      "slideshow": {
       "slide_start": false,
       "slide_type": "fragment"
      }
     },
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "%autopx enabled\n"
       ]
      }
     ],
     "prompt_number": 19
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "Advanced Capabilities\n",
      "----------------------------\n",
      "\n",
      "Other features supported by mpi4py\n",
      "\n",
      "* dynamic process spawning: Spawn(), Connect() and Disconnect()\n",
      "* one sided communication: Put(), Get(), Accumulate()\n",
      "* MPI-IO: Open(), Close(), Get_view() and Set_view()"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {
      "slideshow": {
       "slide_start": true,
       "slide_type": "slide"
      }
     },
     "source": [
      "---\n",
      "More Comprehensive mpi4py Tutorials\n",
      "----------------------------\n",
      "<!--====-->\n",
      "\n",
      "* basics - http://mpi4py.scipy.org/docs/usrman/tutorial.html\n",
      "* advanced - http://www.bu.edu/pasi/files/2011/01/Lisandro-Dalcin-mpi4py.pdf\n",
      "\n",
      "Interesting Scalable Applications and Tools\n",
      "---------------------------------------------\n",
      "* PyTrilinos - http://trilinos.sandia.gov/packages/pytrilinos/\n",
      "* petsc4py - http://code.google.com/p/petsc4py/\n",
      "* PyClaw - http://numerics.kaust.edu.sa/pyclaw/\n",
      "* GPAW - https://wiki.fysik.dtu.dk/gpaw/"
     ]
    }
   ],
   "metadata": {}
  }
 ]
}