Benchmarking Julia implementations of generating Vandermonde matrices vs Numpy

vander_bench.ipynb

{
 "metadata": {
  "language": "Julia",
  "name": "",
  "signature": "sha256:e406752cd90928027c92f3c8f28f3e189d600ac0de461e4c9b1c519bc6659c8f"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# Benchmarking Julia implementations of generating Vandermonde matrices vs Numpy"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "versioninfo()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Julia Version 0.3.4\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Commit 3392026* (2014-12-26 10:42 UTC)\n",
        "Platform Info:\n",
        "  System: Darwin (x86_64-apple-darwin13.4.0)\n",
        "  CPU: Intel(R) Core(TM) i7-4870HQ CPU @ 2.50GHz\n",
        "  WORD_SIZE: 64\n",
        "  BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Haswell)\n",
        "  LAPACK: libopenblas\n",
        "  LIBM: libopenlibm\n",
        "  LLVM: libLLVM-3.3\n"
       ]
      }
     ],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Vandermonde matrix from MatrixDepot.jl (https://github.com/weijianzhang/MatrixDepot.jl)\n",
      "function vand{T}(p::Vector{T}, n::Int)\n",
      "    # n: number of rows\n",
      "    # p: a vector\n",
      "    m = length(p)\n",
      "    V = Array(T, (m, n))\n",
      "    for i = 1:n\n",
      "        V[:,i] = p.^(i-1)\n",
      "    end\n",
      "    return V\n",
      "end\n",
      "vand{T}(::Type{T}, n::Int) = vand(T[1:n], n)\n",
      "vand{T}(p::Vector{T}) = vand(p, length(p))\n",
      "\n",
      "function vand_devec{T}(p::Vector{T}, n::Int)\n",
      "    # n: number of rows\n",
      "    # p: a vector\n",
      "    m = length(p)\n",
      "    V = Array(T, (m, n))\n",
      "    for i = 1:n\n",
      "        for j = 1:m\n",
      "            V[j,i] = p[j]^(i-1)\n",
      "        end\n",
      "    end\n",
      "    return V\n",
      "end\n",
      "vand_devec{T}(::Type{T}, n::Int) = vand_devec(T[1:n], n)\n",
      "vand_devec{T}(p::Vector{T}) = vand_devec(p, length(p))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 2,
       "text": [
        "vand_devec (generic function with 3 methods)"
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "function vander{T}(p::Vector{T}, n::Int)\n",
      "    m = length(p)\n",
      "    v = Array(T, (m, n))\n",
      "    if m > 0\n",
      "        v[:,1] = 1\n",
      "    end\n",
      "    if m > 1\n",
      "        for k = 2:n\n",
      "            v[:, k] = p\n",
      "        end\n",
      "        w = sub(v, :, 2:n)\n",
      "        cumprod!(w, w, 2)\n",
      "    end\n",
      "    return v\n",
      "end\n",
      "\n",
      "vander{T}(::Type{T}, n::Int) = vander(T[1:n], n)\n",
      "vander{T}(p::Vector{T}) = vander(p, length(p))\n",
      "\n",
      "function vander_devec{T}(p::Vector{T}, n::Int)\n",
      "    m = length(p)\n",
      "    v = Array(T, (m, n))\n",
      "    if m > 0\n",
      "        v[:,1] = 1\n",
      "    end\n",
      "    if m > 1\n",
      "        for k = 2:n\n",
      "            for j = 1:m\n",
      "                v[j, k] = p[j]\n",
      "            end\n",
      "        end\n",
      "        w = sub(v, :, 2:n)\n",
      "        cumprod!(w, w, 2)\n",
      "    end\n",
      "    return v\n",
      "end\n",
      "\n",
      "vander_devec{T}(::Type{T}, n::Int) = vander_devec(T[1:n], n)\n",
      "vander_devec{T}(p::Vector{T}) = vander_devec(p, length(p))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 3,
       "text": [
        "vander_devec (generic function with 3 methods)"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "function vander2{T}(p::Vector{T}, n::Int)\n",
      "    m = length(p)\n",
      "    v1 = ones(T, (m,1))\n",
      "    m < 2 && return v1\n",
      "    \n",
      "    v2 = Array(T, (m, n - 1))\n",
      "    for k = 1:n-1\n",
      "        v2[:, k] = p\n",
      "    end\n",
      "    cumprod!(v2, v2, 2)\n",
      "    \n",
      "    return hcat(v1, v2)\n",
      "end\n",
      "\n",
      "vander2{T}(::Type{T}, n::Int) = vander2(T[1:n], n)\n",
      "vander2{T}(p::Vector{T}) = vander2(p, length(p))\n",
      "\n",
      "function vander2_devec{T}(p::Vector{T}, n::Int)\n",
      "    m = length(p)\n",
      "    v1 = ones(T, (m,1))\n",
      "    m < 2 && return v1\n",
      "    \n",
      "    v2 = Array(T, (m, n - 1))\n",
      "    for k = 1:n-1\n",
      "        for j = 1:m\n",
      "            v2[j, k] = p[j]\n",
      "        end\n",
      "    end\n",
      "    cumprod!(v2, v2, 2)\n",
      "\n",
      "    return hcat(v1, v2)\n",
      "end\n",
      "\n",
      "vander2_devec{T}(::Type{T}, n::Int) = vander2_devec(T[1:n], n)\n",
      "vander2_devec{T}(p::Vector{T}) = vander2_devec(p, length(p))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 4,
       "text": [
        "vander2_devec (generic function with 3 methods)"
       ]
      }
     ],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Suggested by Steven G. Johnson: https://github.com/stevengj on the Julia-user list\n",
      "function vander3{T}(p::Vector{T}, n::Int)\n",
      "    # n: number of rows\n",
      "    # p: a vector\n",
      "    m = length(p)\n",
      "    V = Array(T, m, n)\n",
      "    for j = 1:m\n",
      "        @inbounds V[j, 1] = 1\n",
      "    end\n",
      "    for i = 2:n\n",
      "        for j = 1:m\n",
      "            @inbounds V[j,i] = p[j] * V[j,i-1]\n",
      "        end\n",
      "    end\n",
      "    return V\n",
      "end\n",
      "\n",
      "vander3{T}(::Type{T}, n::Int) = vander3(T[1:n], n)\n",
      "vander3{T}(p::Vector{T}) = vander3(p, length(p))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 5,
       "text": [
        "vander3 (generic function with 3 methods)"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## Benchmarking\n",
      "\n",
      "Here we're just going to use `Int64` arrays of ones since I don't care about Vandermonde matrices per se and just want to benchmark implementations, not worrying about over/underflow (https://groups.google.com/d/msg/julia-users/Q96aPufg4S8/IBU9hW0xvWYJ) "
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Adapted code from TimeIt.jl package to actually return timing values for plotting\n",
      "module TimeIt\n",
      "\n",
      "export @timeit\n",
      "\n",
      "macro timen(ex, n)\n",
      "    quote\n",
      "        local t0 = time_ns()\n",
      "        for i = 1:$(esc(n))\n",
      "            local val = $(esc(ex))\n",
      "        end\n",
      "        local t1 = time_ns()\n",
      "        (t1 - t0) / 1.e9\n",
      "    end\n",
      "end\n",
      "\n",
      "macro timeit(ex)\n",
      "    quote\n",
      "        local val = $(esc(ex))  # Warm up\n",
      "        t = zeros(3)\n",
      "        \n",
      "        # Determine number of loops so that total time > 0.1s.\n",
      "        n = 1\n",
      "        for i = 0:9\n",
      "            n = 10^i\n",
      "            t[1] = @timen $(esc(ex)) n\n",
      "            if t[1] > 0.1\n",
      "                break\n",
      "            end\n",
      "        end\n",
      "\n",
      "        # Two more production runs.\n",
      "        for i = 2:3\n",
      "            t[i] = @timen $(esc(ex)) n\n",
      "        end\n",
      "        best = minimum(t) / n\n",
      "        best_ns = best\n",
      "        \n",
      "        # Format to nano-, micro- or milliseconds.\n",
      "        if best < 1e-6\n",
      "            best *= 1e9\n",
      "            pre = \"n\"\n",
      "        elseif best < 1e-3\n",
      "            best *= 1e6\n",
      "            pre = \"\\u00b5\"\n",
      "        else\n",
      "            best *= 1e3\n",
      "            pre = \"m\"\n",
      "        end\n",
      "        @printf \"%d loops, best of 3: %4.2f %ss per loop\\n\" n best pre\n",
      "\n",
      "        best_ns\n",
      "    end\n",
      "end\n",
      "\n",
      "end  # module"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "using TimeIt\n",
      "\n",
      "msizes = [2, 5, 10, 50, 100, 1000, 5000]\n",
      "vmethods = [vand, vand_devec, vander, vander_devec, vander2, vander2_devec, vander3]\n",
      "results = Array(Float64, (length(msizes), length(vmethods)))\n",
      "\n",
      "for (i, N) in enumerate(msizes)\n",
      "    println(\"---------\")\n",
      "    println(\"N: \", N)\n",
      "    println(\"---------\")\n",
      "    for (j, f) in enumerate(vmethods)\n",
      "\n",
      "        p = ones(Int64, (N,))\n",
      "        println(string(f), \":\")\n",
      "        results[i, j] = @timeit f(p)\n",
      "    end\n",
      "end"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "---------\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "N: 2\n",
        "---------\n",
        "vand:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 377.41 ns per loop\n",
        "vand_devec:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 130.64 ns per loop\n",
        "vander:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 1.82 \u00b5s per loop\n",
        "vander_devec:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 1.79 \u00b5s per loop\n",
        "vander2:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 326.47 ns per loop\n",
        "vander2_devec:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 309.00 ns per loop\n",
        "vander3:\n",
        "10000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 84.82 ns per loop\n",
        "---------\n",
        "N: 5\n",
        "---------\n",
        "vand:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 1.06 \u00b5s per loop\n",
        "vand_devec:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 304.89 ns per loop\n",
        "vander:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 2.13 \u00b5s per loop\n",
        "vander_devec:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 2.06 \u00b5s per loop\n",
        "vander2:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 555.71 ns per loop\n",
        "vander2_devec:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 500.62 ns per loop\n",
        "vander3:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 160.93 ns per loop\n",
        "---------\n",
        "N: 10\n",
        "---------\n",
        "vand:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 3.02 \u00b5s per loop\n",
        "vand_devec:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 1.09 \u00b5s per loop\n",
        "vander:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 2.66 \u00b5s per loop\n",
        "vander_devec:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 2.62 \u00b5s per loop\n",
        "vander2:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 1.50 \u00b5s per loop\n",
        "vander2_devec:\n",
        "100000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 1.22 \u00b5s per loop\n",
        "vander3:\n",
        "1000000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 536.44 ns per loop\n",
        "---------\n",
        "N: 50\n",
        "---------\n",
        "vand:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 47.47 \u00b5s per loop\n",
        "vand_devec:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 32.32 \u00b5s per loop\n",
        "vander:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 23.42 \u00b5s per loop\n",
        "vander_devec:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 24.10 \u00b5s per loop\n",
        "vander2:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 19.29 \u00b5s per loop\n",
        "vander2_devec:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 22.25 \u00b5s per loop\n",
        "vander3:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 9.03 \u00b5s per loop\n",
        "---------\n",
        "N: 100\n",
        "---------\n",
        "vand:\n",
        "1000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 178.02 \u00b5s per loop\n",
        "vand_devec:\n",
        "1000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 158.29 \u00b5s per loop\n",
        "vander:\n",
        "1000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 84.43 \u00b5s per loop\n",
        "vander_devec:\n",
        "1000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 103.96 \u00b5s per loop\n",
        "vander2:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 87.98 \u00b5s per loop\n",
        "vander2_devec:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 88.99 \u00b5s per loop\n",
        "vander3:\n",
        "10000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 43.34 \u00b5s per loop\n",
        "---------\n",
        "N: 1000\n",
        "---------\n",
        "vand:\n",
        "10"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 21.44 ms per loop\n",
        "vand_devec:\n",
        "10"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 19.44 ms per loop\n",
        "vander:\n",
        "100"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 7.61 ms per loop\n",
        "vander_devec:\n",
        "100"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 7.49 ms per loop\n",
        "vander2:\n",
        "10"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 6.68 ms per loop\n",
        "vander2_devec:\n",
        "10"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 7.02 ms per loop\n",
        "vander3:\n",
        "100"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 3.32 ms per loop\n",
        "---------\n",
        "N: 5000\n",
        "---------\n",
        "vand:\n",
        "1"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 603.39 ms per loop\n",
        "vand_devec:\n",
        "1"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 574.85 ms per loop\n",
        "vander:\n",
        "1"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 194.65 ms per loop\n",
        "vander_devec:\n",
        "1"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 207.63 ms per loop\n",
        "vander2:\n",
        "1"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 205.05 ms per loop\n",
        "vander2_devec:\n",
        "1"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 220.20 ms per loop\n",
        "vander3:\n",
        "10"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " loops, best of 3: 108.48 ms per loop\n"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "results"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 8,
       "text": [
        "7x7 Array{Float64,2}:\n",
        " 3.77411e-7   1.30637e-7   1.81882e-6  \u2026  3.26468e-7  3.09e-7     8.482e-8  \n",
        " 1.05665e-6   3.04895e-7   2.13367e-6     5.55715e-7  5.00623e-7  1.60928e-7\n",
        " 3.0199e-6    1.09436e-6   2.65604e-6     1.50177e-6  1.22363e-6  5.36436e-7\n",
        " 4.74703e-5   3.23248e-5   2.3416e-5      1.92936e-5  2.22491e-5  9.0292e-6 \n",
        " 0.000178016  0.000158291  8.4435e-5      8.79787e-5  8.89919e-5  4.33375e-5\n",
        " 0.0214381    0.0194448    0.00760928  \u2026  0.00667609  0.0070193   0.00332397\n",
        " 0.603395     0.574849     0.194648       0.205047    0.220196    0.108478  "
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## Comparison to numpy's np.vander()\n",
      "\n",
      "For the code:\n",
      "\n",
      "```python\n",
      "import numpy as np\n",
      "for N in [2, 5, 10, 50, 100, 1000, 5000]:\n",
      "    p = np.ones((N + 1,), dtype=np.int64)\n",
      "    print \"N: \", N\n",
      "    %timeit np.vander(p)\n",
      "```"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "npresults = [4.54e-6, 4.65e-6, 4.84e-6, 11.7e-6, 32.2e-6, 2.89e-3, 152e-3]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 9,
       "text": [
        "7-element Array{Float64,1}:\n",
        " 4.54e-6\n",
        " 4.65e-6\n",
        " 4.84e-6\n",
        " 1.17e-5\n",
        " 3.22e-5\n",
        " 0.00289\n",
        " 0.152  "
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "using PyPlot"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "INFO: Loading help data...\n"
       ]
      }
     ],
     "prompt_number": 10
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Normalize the timings against numpy results\n",
      "for k in 1:length(npresults)\n",
      "    results[k,:] ./= npresults[k]\n",
      "end"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 11
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for k in 1:length(vmethods)\n",
      "    mname = string(vmethods[k])\n",
      "    marker = endswith(mname, \"_devec\") ? \"o\" : \"s\"\n",
      "    \n",
      "    loglog(msizes, results[:,k], linestyle=\"-\", marker=marker, label=mname)\n",
      "end\n",
      "\n",
      "legend(loc=\"lower right\")\n",
      "\n",
      "plot(msizes, ones(length(msizes)), color=\"black\", linestyle=\"--\")\n",
      "xlabel(\"Array size: N\")\n",
      "ylabel(\"Ratio Julia Method/numpy\")"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "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",
       "text": [
        "Figure(PyObject <matplotlib.figure.Figure object at 0x127a71b50>)"
       ]
      },
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 12,
       "text": [
        "PyObject <matplotlib.text.Text object at 0x127addb50>"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 13
    }
   ],
   "metadata": {}
  }
 ]
}