Timing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Zw9q0J3wPec9"
   },
   "source": [
    "# Pylops - basic linear operators on GPU with cupy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "vB2RGSN2cCs3"
   },
   "source": [
    "In this notebook I will show how to use some of the basic linear operators implemented in *Pylops* and test both their numpy and cupy backends."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "BRQ0wMDU7jiS",
    "outputId": "cc6b0921-2df8-4af0-bc68-4abebcfdd12b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mon Feb 15 09:18:58 2021       \r\n",
      "+-----------------------------------------------------------------------------+\r\n",
      "| NVIDIA-SMI 460.39       Driver Version: 460.39       CUDA Version: 11.2     |\r\n",
      "|-------------------------------+----------------------+----------------------+\r\n",
      "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\r\n",
      "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\r\n",
      "|                               |                      |               MIG M. |\r\n",
      "|===============================+======================+======================|\r\n",
      "|   0  Quadro RTX 4000     Off  | 00000000:65:00.0  On |                  N/A |\r\n",
      "| 30%   47C    P0    51W / 125W |   2055MiB /  7979MiB |      2%      Default |\r\n",
      "|                               |                      |                  N/A |\r\n",
      "+-------------------------------+----------------------+----------------------+\r\n",
      "                                                                               \r\n",
      "+-----------------------------------------------------------------------------+\r\n",
      "| Processes:                                                                  |\r\n",
      "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\r\n",
      "|        ID   ID                                                   Usage      |\r\n",
      "|=============================================================================|\r\n",
      "|    0   N/A  N/A      1590      G   /usr/lib/xorg/Xorg                144MiB |\r\n",
      "|    0   N/A  N/A      1642      G   /usr/bin/gnome-shell               66MiB |\r\n",
      "|    0   N/A  N/A      2986      G   ..._14870.log --shared-files       45MiB |\r\n",
      "|    0   N/A  N/A     13176      G   /usr/lib/xorg/Xorg                880MiB |\r\n",
      "|    0   N/A  N/A     13340      G   /usr/bin/gnome-shell              237MiB |\r\n",
      "|    0   N/A  N/A     14424      G   /usr/lib/firefox/firefox           98MiB |\r\n",
      "|    0   N/A  N/A     14538      G   /usr/lib/firefox/firefox            2MiB |\r\n",
      "|    0   N/A  N/A     14596      G   /usr/lib/firefox/firefox           56MiB |\r\n",
      "|    0   N/A  N/A     15710      G   ...acnnk5/bin/glnxa64/MATLAB      120MiB |\r\n",
      "|    0   N/A  N/A     28881      G   /usr/lib/firefox/firefox           98MiB |\r\n",
      "|    0   N/A  N/A     29108      G   /usr/lib/firefox/firefox           98MiB |\r\n",
      "|    0   N/A  N/A     29967      G   /usr/lib/firefox/firefox           98MiB |\r\n",
      "|    0   N/A  N/A     31043      G   gnome-control-center               99MiB |\r\n",
      "+-----------------------------------------------------------------------------+\r\n"
     ]
    }
   ],
   "source": [
    "!nvidia-smi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "U3ZG-RRbomAk",
    "outputId": "5527cf60-c736-402b-abc7-19bb8e3e38fa"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "nvcc: NVIDIA (R) Cuda compiler driver\r\n",
      "Copyright (c) 2005-2019 NVIDIA Corporation\r\n",
      "Built on Wed_Oct_23_19:24:38_PDT_2019\r\n",
      "Cuda compilation tools, release 10.2, V10.2.89\r\n"
     ]
    }
   ],
   "source": [
    "!nvcc --version"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Y8yPDy420evR",
    "outputId": "b8f6e422-6aef-4533-b01b-332c5bb44210"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Architecture:        x86_64\r\n",
      "CPU op-mode(s):      32-bit, 64-bit\r\n",
      "Byte Order:          Little Endian\r\n",
      "CPU(s):              8\r\n",
      "On-line CPU(s) list: 0-7\r\n",
      "Thread(s) per core:  1\r\n",
      "Core(s) per socket:  8\r\n",
      "Socket(s):           1\r\n",
      "NUMA node(s):        1\r\n",
      "Vendor ID:           GenuineIntel\r\n",
      "CPU family:          6\r\n",
      "Model:               85\r\n",
      "Model name:          Intel(R) Xeon(R) W-2145 CPU @ 3.70GHz\r\n",
      "Stepping:            4\r\n",
      "CPU MHz:             1900.006\r\n",
      "CPU max MHz:         4500.0000\r\n",
      "CPU min MHz:         1200.0000\r\n",
      "BogoMIPS:            7399.70\r\n",
      "Virtualization:      VT-x\r\n",
      "L1d cache:           32K\r\n",
      "L1i cache:           32K\r\n",
      "L2 cache:            1024K\r\n",
      "L3 cache:            11264K\r\n",
      "NUMA node0 CPU(s):   0-7\r\n",
      "Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd mba ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts md_clear flush_l1d\r\n"
     ]
    }
   ],
   "source": [
    "!lscpu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "WJ3vTlJ7gHnQ"
   },
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "%matplotlib inline\n",
    "\n",
    "import timeit\n",
    "import os\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import scipy as sp\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "from scipy.sparse.linalg import inv, lsqr\n",
    "from scipy.sparse import rand\n",
    "\n",
    "from pylops.utils import dottest\n",
    "from pylops import LinearOperator\n",
    "from pylops.basicoperators import *\n",
    "from pylops.optimization.solver import cgls\n",
    "from pylops.optimization.leastsquares import *\n",
    "\n",
    "plt.style.use('seaborn-dark')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 37
    },
    "id": "NfTWAwO-7dqU",
    "outputId": "ae253f0a-7799-4a32-d2cb-37d270c14676"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'8.4.0'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cp.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "id": "D6SqvAFwGZs6"
   },
   "outputs": [],
   "source": [
    "def timing(cmd_setup, df, method, number=10, repeat=5):\n",
    "    # _matvec\n",
    "    mat_np = 'Op._matvec(x)'\n",
    "    mat_cp = 'cOp._matvec(xc)'\n",
    "\n",
    "    # _matvec\n",
    "    rmat_np = 'Op._rmatvec(y)'\n",
    "    rmat_cp = 'cOp._rmatvec(yc)'\n",
    "\n",
    "    # dry run\n",
    "    timeit.repeat(mat_cp, setup=cmd_setup, number=1, repeat=1)\n",
    "    timeit.repeat(rmat_cp, setup=cmd_setup, number=1, repeat=1)\n",
    "\n",
    "    # timing\n",
    "    tm_ = 1.e3 * np.array(timeit.repeat(mat_np, setup=cmd_setup,\n",
    "                                      number=number, repeat=repeat))\n",
    "    trm_ = 1.e3 * np.array(timeit.repeat(rmat_np, setup=cmd_setup,\n",
    "                                       number=number, repeat=repeat))\n",
    "    for i in range(repeat):\n",
    "        df.loc[len(df)] = [method, True, 'np', tm_[i]]\n",
    "    for i in range(repeat):\n",
    "        df.loc[len(df)] = [method, False, 'np', + trm_[i]]\n",
    "\n",
    "    tm_ = 1.e3 * np.array(timeit.repeat(mat_cp, setup=cmd_setup,\n",
    "                                      number=number, repeat=repeat))\n",
    "    trm_ = 1.e3 * np.array(timeit.repeat(rmat_cp, setup=cmd_setup,\n",
    "                                       number=number, repeat=repeat))\n",
    "    for i in range(repeat):\n",
    "        df.loc[len(df)] = [method, True, 'cp', tm_[i]]\n",
    "    for i in range(repeat):\n",
    "        df.loc[len(df)] = [method, False, 'cp', trm_[i]]\n",
    "    return df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "NoTH5zUBVbP4"
   },
   "source": [
    "# Elementary operators"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "fH1ohMVv8Cs9"
   },
   "outputs": [],
   "source": [
    "df = pd.DataFrame(columns=['Method', 'Forw', 'Backend', 'Time'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uZIXTxF-GNp2"
   },
   "source": [
    "## Diagonal"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "aXWgRcxHGxaE"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "d = np.arange(n) + 1.\n",
    "y = x = np.ones(n)\n",
    "yc = xc = cp.ones(n)\n",
    "Op = pylops.Diagonal(d)\n",
    "cOp = pylops.Diagonal(cp.array(d))\n",
    "\"\"\"\n",
    "\n",
    "df = timing(cmd_setup, df, 'Diagonal')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ibe5oiJjQbYx"
   },
   "source": [
    "## Matrix Multiplication"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "zLcsPQuSKREi"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = 1000\n",
    "D = np.ones((n, n))\n",
    "y = x = np.ones(n)\n",
    "yc = xc = cp.ones(n)\n",
    "Op = pylops.MatrixMult(D)\n",
    "cOp = pylops.MatrixMult(cp.array(D))\n",
    "\"\"\"\n",
    "\n",
    "df = timing(cmd_setup, df, 'MatrixMult')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ArXNiAVEGn0W"
   },
   "source": [
    "## Identity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "5AKKVy6vGlw8"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "x = np.ones(n//2)\n",
    "xc = cp.ones(n//2)\n",
    "y = np.ones(n)\n",
    "yc = cp.ones(n)\n",
    "Op = pylops.Identity(n, n//2)\n",
    "cOp = pylops.Identity(n, n//2)\n",
    "\"\"\"\n",
    "\n",
    "df = timing(cmd_setup, df, 'Identity')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "duzvLYbhKutX"
   },
   "source": [
    "## Zero"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "id": "NvS0uZJGKuCU"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = int(1e6), int(1e6)\n",
    "y = x = cp.arange(nx)\n",
    "yc = xc = cp.ones(nx)\n",
    "Op = pylops.Zero(ny, nx)\n",
    "cOp = pylops.Zero(ny, nx)\n",
    "\"\"\"\n",
    "\n",
    "df = timing(cmd_setup, df, 'Zero')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HFjgyviFPo0p"
   },
   "source": [
    "## Sum"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "id": "zIzjLLvZKuAl"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx)\n",
    "yc = cp.ones(nx)\n",
    "Op = pylops.Sum(dims=(ny, nx), dir=0)\n",
    "cOp = pylops.Sum(dims=(ny, nx), dir=0)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Sum')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "nljZK0xESzTt"
   },
   "source": [
    "## First Derivative"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "id": "M4-NfcOaSzwD"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx*ny)\n",
    "yc = cp.ones(nx*ny)\n",
    "Op = pylops.FirstDerivative(ny*nx, (ny, nx), dir=-1, edge=True)\n",
    "cOp = pylops.FirstDerivative(ny*nx, (ny, nx), dir=-1, edge=True)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'FirstDerivative')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "RGSUZAwQUlxH"
   },
   "source": [
    "## Second Derivative"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "id": "1YYLU1zaUlZj"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx*ny)\n",
    "yc = cp.ones(nx*ny)\n",
    "Op = pylops.SecondDerivative(ny*nx, (ny, nx), dir=-1, edge=True)\n",
    "cOp = pylops.SecondDerivative(ny*nx, (ny, nx), dir=-1, edge=True)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'SecondDerivative')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Jm0s1WeXUlB2"
   },
   "source": [
    "## Laplacian"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "wmfnD1vWUyKR"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx*ny)\n",
    "yc = cp.ones(nx*ny)\n",
    "Op = pylops.Laplacian((ny, nx), edge=True)\n",
    "cOp = pylops.Laplacian((ny, nx), edge=True)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Laplacian')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Ai_p16-fUk3Z"
   },
   "source": [
    "## Gradient"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "id": "Md2WUqaCVwch"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx*ny)\n",
    "yc = cp.ones(nx*ny)\n",
    "Op = pylops.Laplacian((ny, nx), edge=True)\n",
    "cOp = pylops.Laplacian((ny, nx), edge=True)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Gradient')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LhV7dzXWV_h5"
   },
   "source": [
    "## Causal Integration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "id": "1LN5uCZuWFQi"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx*ny)\n",
    "yc = cp.ones(nx*ny)\n",
    "Op = pylops.CausalIntegration(ny*nx, (ny, nx), dir=1)\n",
    "cOp = pylops.CausalIntegration(ny*nx, (ny, nx), dir=1)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'CausalIntegration')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "y2N7wZjSV_eb"
   },
   "source": [
    "## Flip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "id": "eS4SXDlZXXkz"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx*ny)\n",
    "yc = cp.ones(nx*ny)\n",
    "Op = pylops.Flip(ny*nx, (ny, nx), dir=1)\n",
    "cOp = pylops.Flip(ny*nx, (ny, nx), dir=1)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Flip')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sl0SCAfQV_Zc"
   },
   "source": [
    "## Pad"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "id": "YrzNIQTOY4tx"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "pad = (200, 200)\n",
    "x = np.ones(n)\n",
    "xc = cp.ones(n)\n",
    "y = np.ones(n + 400)\n",
    "yc = cp.ones(n + 400)\n",
    "Op = pylops.Pad(n, pad)\n",
    "cOp = pylops.Pad(n, pad)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Pad')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ONLcerOaaYsQ"
   },
   "source": [
    "## Roll"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "id": "uWDPIb05aZJq"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "x = np.ones(n)\n",
    "xc = cp.ones(n)\n",
    "y = np.ones(n)\n",
    "yc = cp.ones(n)\n",
    "Op = pylops.Roll(n, shift=20)\n",
    "cOp = pylops.Roll(n, shift=20)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Roll')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eAWELYq3asum"
   },
   "source": [
    "## Symmetrize\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "id": "BObcCFDVatoP"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "x = np.ones(n)\n",
    "xc = cp.ones(n)\n",
    "y = np.ones(n*2-1)\n",
    "yc = cp.ones(n*2-1)\n",
    "Op = pylops.Symmetrize(n)\n",
    "cOp = pylops.Symmetrize(n)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Symmetrize')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "zkVlMj9SasfC"
   },
   "source": [
    "## Restriction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "id": "s6-Ffu1Wat1S"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "iava = np.arange(0,n,10)\n",
    "x = np.ones(n)\n",
    "xc = cp.ones(n)\n",
    "y = np.ones(n//10)\n",
    "yc = cp.ones(n//10)\n",
    "Op = pylops.Restriction(n, iava)\n",
    "cOp = pylops.Restriction(n, iava)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Restriction')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3q-Rb7MmasO9"
   },
   "source": [
    "## Transpose"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "id": "-gLKFbUscj0r"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx*ny)\n",
    "yc = cp.ones(nx*ny)\n",
    "Op = pylops.Transpose((ny, nx), axes=(1,0))\n",
    "cOp = pylops.Transpose((ny, nx), axes=(1,0))\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Transpose')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Nj9Zru8SfT9B"
   },
   "source": [
    "## FFT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "id": "3rpDS-YEfYB4"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "x = np.ones(n)\n",
    "xc = cp.ones(n)\n",
    "y = np.ones(n*2-1)\n",
    "yc = cp.ones(n*2-1)\n",
    "Op = pylops.signalprocessing.FFT(dims=[n], nfft=n)\n",
    "cOp = pylops.signalprocessing.FFT(dims=[n], nfft=n)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'FFT')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "gn4-QSi784RL"
   },
   "source": [
    "## FFT2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "id": "ufrzvDrJ85Pk"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "ny, nx = 1000, 1000\n",
    "x = np.ones(nx*ny)\n",
    "xc = cp.ones(nx*ny)\n",
    "y = np.ones(nx*ny)\n",
    "yc = cp.ones(nx*ny)\n",
    "Op = pylops.signalprocessing.FFT2D(dims=[ny, nx])\n",
    "cOp = pylops.signalprocessing.FFT2D(dims=[ny, nx])\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'FFT2D')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "i4TWbT4kDLVe"
   },
   "source": [
    "## FFTND"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "id": "gG1KBIspB4fy"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "nz, ny, nx = 50, 100, 100\n",
    "x = np.ones(nz*nx*ny)\n",
    "xc = cp.ones(nz*nx*ny)\n",
    "y = np.ones(nz*nx*ny)\n",
    "yc = cp.ones(nz*nx*ny)\n",
    "Op = pylops.signalprocessing.FFTND(dims=[nz, ny, nx])\n",
    "cOp = pylops.signalprocessing.FFTND(dims=[nz, ny, nx])\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'FFTND')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "nUZPKkGWDKsh"
   },
   "source": [
    "## Convolve 1D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "id": "XL2kapI0B5gc"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "x = np.ones(n)\n",
    "xc = cp.ones(n)\n",
    "y = np.ones(n)\n",
    "yc = cp.ones(n)\n",
    "h = np.ones(25)\n",
    "hc = cp.ones(25)\n",
    "Op = pylops.signalprocessing.Convolve1D(n, h)\n",
    "cOp = pylops.signalprocessing.Convolve1D(n, hc)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Convolve1D')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Kt0tPxzvH40U"
   },
   "source": [
    "## Convolve2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "id": "S0-QvuEYH60V"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "nt, nx = 1000, 1000\n",
    "h = np.ones((5,5))\n",
    "hc = cp.ones((5,5))\n",
    "x = np.ones(nt*nx)\n",
    "xc = cp.ones(nt*nx)\n",
    "y = np.ones(nt*nx)\n",
    "yc = cp.ones(nt*nx)\n",
    "Op = pylops.signalprocessing.Convolve2D(nt*nx, h, (nt, nx))\n",
    "cOp = pylops.signalprocessing.Convolve2D(nt*nx, h, (nt, nx))\n",
    "\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Convolve2D')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "xqTlNhVWP6gm"
   },
   "source": [
    "## ConvolveND"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "id": "svCfiUZSP59q"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "nt, ny, nx = 100, 100, 100\n",
    "h = np.ones((10, 10, 10))\n",
    "hc = cp.ones((10,10))\n",
    "x = np.ones(nt*ny*nx)\n",
    "xc = cp.ones(nt*ny*nx)\n",
    "y = np.ones(nt*ny*nx)\n",
    "yc = cp.ones(nt*ny*nx)\n",
    "Op = pylops.signalprocessing.ConvolveND(nt*ny*nx, h, (nt, ny, nx))\n",
    "cOp = pylops.signalprocessing.ConvolveND(nt*ny*nx, h, (nt, ny, nx))\n",
    "\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'ConvolveND')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OEF5FBkbD8yV"
   },
   "source": [
    "## ChirpRadon2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "id": "k1Z8uhXvD7QU"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "nt, nh = 1000, 1000\n",
    "t = np.arange(nt)\n",
    "h = np.arange(nh)\n",
    "x = np.ones(nt*nh)\n",
    "xc = cp.ones(nt*nh)\n",
    "y = np.ones(nt*nh)\n",
    "yc = cp.ones(nt*nh)\n",
    "Op = pylops.signalprocessing.ChirpRadon2D(t, h, 0.001)\n",
    "cOp = pylops.signalprocessing.ChirpRadon2D(t, h, 0.001)\n",
    "\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'ChirpRadon2D')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "bdCvRV9OH063"
   },
   "source": [
    "## ChirpRadon3D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "id": "I9GSSGtJFr76"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "nt, nh1, nh2 = 100, 100, 100\n",
    "t = np.arange(nt)\n",
    "h1 = np.arange(nh1)\n",
    "h2 = np.arange(nh2)\n",
    "x = np.ones(nt*nh1*nh2)\n",
    "xc = cp.ones(nt*nh1*nh2)\n",
    "y = np.ones(nt*nh1*nh2)\n",
    "yc = cp.ones(nt*nh1*nh2)\n",
    "Op = pylops.signalprocessing.ChirpRadon3D(t, h1, h2, (0.001, 0.001), engine='fftw')\n",
    "cOp = pylops.signalprocessing.ChirpRadon3D(t, h1, h2, (0.001, 0.001))\n",
    "\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'ChirpRadon3D')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "xjqCp8z1QsnO"
   },
   "source": [
    "## Interp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "id": "sZHAHHM4QtCr"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "n = int(1e6)\n",
    "iava = np.arange(0,n,10) + 0.5\n",
    "x = np.ones(n)\n",
    "xc = cp.ones(n)\n",
    "y = np.ones(n//10)\n",
    "yc = cp.ones(n//10)\n",
    "Op = pylops.signalprocessing.Interp(n, iava, kind='nearest')[0]\n",
    "cOp = pylops.signalprocessing.Interp(n, iava, kind='nearest')[0]\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Interp')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ngmQilEJRy_4"
   },
   "source": [
    "## Bilinear"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "OPgPPYsFRzXW",
    "outputId": "17378301-3106-4981-afac-a3d5b8074ba5"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ravasim/miniconda3/envs/pylops_cupy_cusignal/lib/python3.8/site-packages/cupyx/_scatter.py:56: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[cupy.array(seq)]`, which will result either in an error or a different result.\n",
      "  a.scatter_add(slices, value)\n",
      "/home/ravasim/Documents/OpenSource/pylops/pylops/signalprocessing/Bilinear.py:122: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n",
      "  ncp_add_at(y, [self.iava_t, self.iava_l],\n",
      "/home/ravasim/Documents/OpenSource/pylops/pylops/signalprocessing/Bilinear.py:124: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n",
      "  ncp_add_at(y, [self.iava_t, self.iava_r],\n",
      "/home/ravasim/Documents/OpenSource/pylops/pylops/signalprocessing/Bilinear.py:126: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n",
      "  ncp_add_at(y, [self.iava_b, self.iava_l],\n",
      "/home/ravasim/Documents/OpenSource/pylops/pylops/signalprocessing/Bilinear.py:128: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n",
      "  ncp_add_at(y, [self.iava_b, self.iava_r],\n"
     ]
    }
   ],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "nt, nx = 1000, 1000\n",
    "#iava = np.vstack((np.arange(0,nt,2),\n",
    "#                  np.arange(0,nx,2)))\n",
    "iava = np.vstack((np.hstack([i * np.ones(100) for i in range(nt-1)]),\n",
    "                  np.hstack([np.random.permutation(np.arange(100)) for i in range(nt-1)])))\n",
    "iavac = cp.asarray(iava)\n",
    "x = np.ones(nt*nx)\n",
    "xc = cp.ones(nt*nx)\n",
    "y = np.ones(iava.shape[1])\n",
    "yc = cp.ones(iava.shape[1])\n",
    "Op = pylops.signalprocessing.Bilinear(iava, (nt, nx))\n",
    "cOp = pylops.signalprocessing.Bilinear(iavac, (nt, nx))\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Bilinear')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "nYNEKpT3RymX"
   },
   "source": [
    "## Fredholm1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "id": "_1drzUZhUPB0"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "nt, nh1, nh2, nz = 100, 50, 50, 20\n",
    "h = np.ones((nt, nh1, nh2))\n",
    "hc = cp.ones((nt, nh1, nh2))\n",
    "x = np.ones(nt*nh2*nz)\n",
    "xc = cp.ones(nt*nh2*nz)\n",
    "y = np.ones(nt*nh1*nz)\n",
    "yc = cp.ones(nt*nh1*nz)\n",
    "Op = pylops.signalprocessing.Fredholm1(h, nz)\n",
    "cOp = pylops.signalprocessing.Fredholm1(hc, nz)\n",
    "\"\"\"\n",
    "df = timing(cmd_setup, df, 'Fredholm1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "id": "zUiGE2U0WiFI"
   },
   "outputs": [],
   "source": [
    "#df=df[df['Method']!='Bilinear']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 629
    },
    "id": "9PHJ10Yx8MUw",
    "outputId": "7287731f-938a-4f42-d6e0-b24b217cae7d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtkAAAEyCAYAAAAvJsxCAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAACAEElEQVR4nO3dd1gU19fA8e/SkWKj2A2KYtcY7BqjgD2KFXuMPXbF3nvXEHsXe1QswYpdUbEAKmLsWEGlKr3v+0fe3R8mMSosi+V8nidPdNydc3d3dvbMnXvPVSiVSiVCCCGEEEIIjdHJ6QYIIYQQQgjxpZEkWwghhBBCCA2TJFsIIYQQQggNkyRbCCGEEEIIDZMkWwghhBBCCA2TJFsIIYQQQggNkyRbCCE+Qx07dqRly5bv/HdfX18aNmwIwKJFi9ixY8d/7u/48eOMGzfuvXG9vb0JCQn5uMYKIcRXSJJsIYT4zNy/fx8zMzMKFSrEtWvX3vt4V1dXOnXq9J+PcXJyYs6cOe/dl7u7uyTZQgjxAfRyugFCCCE+zt69e2nSpAmGhobs37+fb7/9FoAVK1awc+dO8uXLR4MGDdSPHzt2LMWKFWPAgAHcuXOHqVOn8vr1awwNDRk5ciT16tVj7969eHp64u7uztixY9UJ/OPHj/nmm29YsWIFq1ev5tKlSwQFBTFq1CiaNWuWU2+BEEJ88qQnWwghPiNpaWkcP36cxo0b4+DgwLlz50hOTubBgwe4u7uzZ88ePDw8uHv37j+em56ezogRI+jatStHjx5l5syZuLq6Ehsb+4/HHj16lF9//ZXjx48TGRnJ8ePHGTZsGNbW1ixYsEASbCGEeA/pyRZCiM/I+fPnqVixIqampgBUr16d06dPExkZSbVq1bCwsACgZcuW3L59+63nPn/+nPDwcJo3bw5AxYoVKVSoEDdv3vxHnPr165MnTx4ASpcuzYsXL7LxVQkhxJdHkmwhhPiM7N27l3PnzmFvbw/81bP95s0bqlSpgpmZmfpx5ubm/3huZGQkZmZmKBSKtx4XGRn5j8dm3Jeuri5paWmafBlCCPHFkyRbCCE+E9HR0Vy5coXLly9jYGAAQGpqKvXr16dixYrExMSoHxsVFfWP5+fPn583b96gVCrVifbr16/Jnz+/TGYUQggNkzHZQgjxmTh48CA1a9ZUJ9gAenp61K1bF1NTU/z8/IiMjCQtLQ1PT89/PL9IkSIUKFCAw4cPA+Dv7094eDiVKlX64Dbo6em9lcwLIYT4d5JkCyHEZ2L//v04Ojr+Y7uTkxN//PEHHTt2pHXr1rRp04aqVav+43EKhYLFixezdetWmjZtysyZM/ntt9/IlSvXB7ehcePGDB8+nI0bN2bptQghxJdOoVQqlTndCCGEENknYwk/IYQQ2iE92UII8YWLjo7G2Ng4p5shhBBfFUmyhRDiCzZ+/Hhu376Ng4NDTjdFCCG+KjJcRAghhBBCCA2TnmwhhBBCCCE07LOukx0WJmWkhBBCCCFE9rO0NHv/gzKQnmwhhBBCCCE0TJJsIYQQQgghNEySbCGEEEIIITRMkmwhhBBCCCE0TJJsIYQQQgghNEySbCGEEEIIITRMkmwhhBBCCCE0LFvrZM+fPx8/Pz9SU1Pp168fFStWZPTo0aSlpWFpacmCBQswMDDA09OTTZs2oaOjg4uLC+3atSMlJYWxY8cSEhKCrq4uc+bMoWjRotnZXCGEEEIIITQi25LsS5cucf/+fXbu3ElUVBStW7emVq1adO7cmaZNmzJ//nw8PDxwdnZm+fLleHh4oK+vj7OzM46Ojpw+fRpzc3MWLVrE2bNnWbRoEW5ubtnVXCGEEEIIITQm24aLVKtWjd9++w2A3Llzk5CQwOXLl3FwcADAwcEBHx8fbty4QcWKFTEzM8PIyAh7e3v8/f3x8fHByckJgLp16+Ln55ddTRVCCCGEEFrw4kUITk7fM2hQXwYP7ke/fj9z48b1j3p+r17dNNaeiRNH4+/vq7H9ZZRtPdm6urrkypULgN27d/P9999z/vx5DAwMALC0tCQsLIzw8HDy5cunfp6FhcU/tuvq6qKjo0NycrL6+UIIIYQQX4uafSI+6HGX1ubP5pZkXbFixVm2bA0A16/7s2nTOhYvXpbDrdK8bB2TDXDixAk8PDzYsGEDjRs3Vm9XKpVv/T/jdoVC8c7tQgghhBDiyxAZGYmFhSX3799j8eJ56OnpoaOjw4wZczE3z822bZs4c+YkCoUO/fsPomDBQurn+vhcYM+encyb9yt//LGX48ePoFDoUK/eD3Tq1JX161cTFxfL06dPCA5+zpAhrtSqVYdt2zZx8uQxihYtRnR0dLa9tmxNsr29vVm1ahXr1q3DzMwMY2NjEhMTMTIy4tWrV1hZWWFtbc2ZM2fUzwkNDaVKlSpYW1sTFhZGmTJlSElJQalUoq+vn53NFUIIIYQQ2ezp0ycMGtSX5ORkwsPDWLRoKeHhoQwfPorSpcuwbt0qjh07Qo0atTlz5iSrV7sTEhLM1q3u/PRTLwCeP3/Gpk3rWbhwCa9eveT06ROsWLEegF9+6UWDBo4AhIa+YuHCJVy6dJE//thDhQqV2LfPg23bPEhLS6VDB+dse53ZlmTHxMQwf/583N3dyZMnDwC1a9fGy8uLVq1acezYMerVq0flypWZOHEi0dHR6Orq4u/vz/jx44mNjeXo0aPUq1eP06dPU6NGjexqqhBCCCGE0JKMw0WePHnMpEljmDx5JitXLiUpKZHw8DCcnJpw795dypWrgI6ODkWKFGXs2Em8eBFCYmIC48a5MnHidExNTbl8+SLPnz9j8OB+AMTHx/HyZQgAlSpVAcDKyorY2FiCg59hY1MCQ0NDwBA7u7LZ9jqzLck+fPgwUVFRDBs2TL1t7ty5TJw4kZ07d1KoUCGcnZ3R19fH1dWVXr16oVAoGDhwIGZmZjRr1oyLFy/SqVMnDAwMmDt3bnY1VQghhBBC5IDixb/B0NCQ335bSJcuP1GzZm22b99CQkI8uro6pKcr//Gc0NBQGjduyr59uxk7dhJ6evrUqlWH0aMnvPU4P7+r6Orqqv+uVCr/f/ixToZt6dn22rItyXZxccHFxeUf2zdu3PiPbU2aNKFJkyZvbVPVxhZCCCGEEF+m6Og3REREYGxsTOHCRUhOTubSpQuUL18RO7uyuLuvJzU1lejoNyxYMIchQ0ZQrFhxXF3HMmRIf65cuYSdXVlWrlxKYmLi/yfsi/jll0H/Gq9w4SI8efKI1NRUkpISuXv3dra9tmyf+CiEEEII8TXx9/flwIF9/Phja6pWtc/p5nxyVGOyAZKTkxk+fBSRkZGMGzeSwoUL07atC25uC2jY0InGjZsxaFBflEol/foNVO9DoVAwduwkxowZzpo17nTo0ImBA/ugo6PD99//gKGh0b/GNjfPTdOmLejbtweFChWmTJny2fY6Fcq/l/H4jISFxeR0E4QQQggh3jJu3AgePQrCxqYEc+Ys1sg+v6QSfp8rS0uzj3q89GQLIYQQQmTB3xPgwrFx6AP3Hse99W+SAH9dsm3FRyGEEEIIIb5WkmQLIYQQQmhQusLorf+Lr5Mk2UIIIYQQGvTasBUJuna8NmyV000ROUjGZAshhBBCaFCCXmUS9CrndDNEDpOebCGEEEIIITRMerKFEEIIIUS2+9AyhB/qU6/WIj3ZQgghhBBCaJj0ZAshhBBCiC/O4cMHCAi4zuvXUTx9+oTOnbvh7r6epk1b4Od3FX19A2bOnIeZ2cctMvOhpCdbCCGEEEJ8kR4+fMCsWQuYM2cRHh67AChe/BtWrFiHrW0pjhw5mG2xJckWQgghhBBfpAoVKqGrq4ulpRVxcbEA2NvX+P9/q8jTp0+yLbYk2UIIIYQQ4oukq6ur/rNSqfz//6f///9BoVBkW2xJsoUQQgghxFcjIOA6ALduBfDNNzbZFkcmPgohhBBCiGz3qZTcu3PnNnv37gYU9OrVL9viSJIthBBCCCG+OM2a/aj+c65cufDwOEC7dj/SrdvP5MqVK9vjZ2uSfe/ePQYMGECPHj3o2rUrQ4YMISoqCoDXr19TpUoVZsyYQd26dbGx+V93vbu7O+np6YwdO5aQkBB0dXWZM2cORYsWzc7mCiGEEEIIoRHZlmTHx8czY8YMatWqpd62ZMkS9Z/HjRtH+/btUSqVWFlZsWXLlree7+npibm5OYsWLeLs2bMsWrQINze37GquEEIIIYT4wnl4HNBarGyb+GhgYMDatWuxsrL6x78FBQURExNDpUqViI+PJy0t7R+P8fHxwcnJCYC6devi5+eXXU0VQgghhBBCo7ItydbT08PIyOhf/23z5s107doV+KvHOyIigiFDhtCxY0c2b94MQHh4OPny5QP+Kr+io6NDcnJydjVXCCGEEEIIjdH6xMfk5GT8/PyYOnUqAMbGxgwdOpRWrVqRkpJC165dqVq1qrqWoYpSqczWWoZCCCGEEEJoitbrZF+9epVKlSqp/25qakr79u0xMDDAxMSEWrVqcffuXaytrQkLCwMgJSUFpVKJvr6+tpsrhBBCCCHER9N6T/bNmzcpU6aM+u93795lw4YNzJ07l7S0NPz9/WnSpAkGBgYcPXqUevXqcfr0aWrUqKHtpgohhBBCCA2JW1rp/Q/6CCaDAzS6P03LtiQ7MDCQefPmERwcjJ6eHl5eXixdupSwsDCKFSumfpydnR158uShffv26Ojo0KBBAypVqkT58uW5ePEinTp1wsDAgLlz52ZXU4UQQgghhNAohfLvg58/I2FhMTndBCGEEEJ85Wr2ifigx2VlxUNtxMhu2u7JTk1NZebMKbx69QIDA0OqVrUnKOgBcXFxhIWF0qFDZ5o3b/nB8SwtzT6qfbLioxBCCCGE+OIcOXKQ/PnzM3XqLE6c8CImJoZHj4LYsGEbsbGx9OjRiaZNW6Cjkz1TFLU+8VEIIYQQQojsdvfuHSpWrAyAo2NjDA0NqVKlKnp6euTJkwczMzPevHmdbfElyRZCCCGEEF8cXV0d0tPfHhWd8e9/DZjOvvLQkmQLIYQQQogvTpky5fD3vwrAhQvehIeHcetWAGlpabx+/Zr4+Dhy586dbfFlTLYQQgghhMh22i655+jYGF/fKwwa1BddXV2qVrWnQIFCTJo0luDgZ/TtOyDbxmODJNlCCCGEEOILpK+vz6RJ09V/P3z4ADExMQwaNEwr8WW4iBBCCCGEEBomPdlCCCGEEOKL16zZj1qNJz3ZQgghhBBCaJgk2UIIIYQQQmiYJNlCCCGEEEJomCTZQgghhBBCaJgk2UIIIYQQQmiYJNlCCCGEEEJomCTZQgghhBBCaJgk2UIIIYQQQmiYJNlCCCGEEEJoWLYm2ffu3cPR0ZGtW7cCMGPGDNq0aUO3bt3o1q0bZ86cAcDT05O2bdvSvn17PDw8AEhJScHV1ZVOnTrRtWtXnj17lp1NFUIIIYQQQmOybVn1+Ph4ZsyYQa1atd7aNmvWLMqWLfvWtuXLl+Ph4YG+vj7Ozs44Ojpy+vRpzM3NWbRoEWfPnmXRokW4ubllV3OFEEIIIYTQmGzryTYwMGDt2rVYWVmpt8XFxf3jcTdu3KBixYqYmZlhZGSEvb09/v7++Pj44OTkBEDdunXx8/PLrqYKIYQQQgihUdnWk62np4ee3tu7j4uLY9myZURHR2Ntbc3EiRMJDw8nX7586sdYWFgQFhb21nZdXV10dHRITk7GwMAgu5oshBBCCCGERmRbkv1vOnbsiK2tLTY2NqxcuZKlS5dSuXLltx6jVCpRKBQolcp/3S6EEEIIIcSnTqvVRZycnLCxsVH/+e7du1hbWxMeHq5+TGhoKJaWllhbWxMWFgb8NQlSqVSir6+vzeYKIYQQQgiRKVpNsvv3709ISAgAly9fplSpUlSuXJmbN28SHR1NXFwc/v7+2NvbU6dOHY4ePQrA6dOnqVGjhjabKoQQQgghRKZl23CRwMBA5s2bR3BwMHp6enh5edGpUycGDx5Mrly5MDY2Zs6cORgZGeHq6kqvXr1QKBQMHDgQMzMzmjVrxsWLF+nUqRMGBgbMnTs3u5oqhBBCCCGERimUfx/8/BkJC4vJ6SYIIYQQ4itXs0/EBz3u0tr8n3QM8d8sLc0+6vGy4qMQQgghhBAaJkm2EEIIIYQQGiZJthBCCCGEEBomSbYQQgghhBAaJkm2EEIIIYQQGiZJthBCCCGEEBomSbYQQgghhBAaJkm2EEIIIYQQGiZJthBCCCGEEBomSbYQQgghhBAaJkm2EEIIIYQQGiZJthBCCCGEEBomSbYQQgghhBAaJkm2EEIIIYQQGiZJthBCCCGEEBomSbYQQgghhBAappedO7937x4DBgygR48edO3alRcvXjBu3DhSU1PR09NjwYIFWFpaUrduXWxsbNTPc3d3Jz09nbFjxxISEoKuri5z5syhaNGi2dlcIYQQQgghNCLberLj4+OZMWMGtWrVUm9zc3OjQ4cObN26FScnJzZu3IhSqcTKyootW7ao/9PV1eXgwYOYm5uzY8cO+vTpw6JFi7KrqUIIIYQQQmhUtiXZBgYGrF27FisrK/W2KVOm0LhxYwDy5s3L69eviY+PJy0t7R/P9/HxwcnJCYC6devi5+eXXU0VQgghhBBCoz44yVYqlSiVyg/esZ6eHkZGRm9ty5UrF7q6uqSlpbF9+3Z+/PFH4uPjiYiIYMiQIXTs2JHNmzcDEB4eTr58+QDQ1dVFR0eH5OTkD44vhBBCCCFETvnPMdmBgYFs2rQJX19fUlJSANDX18fe3p7u3btTsWLFjw6YlpbG6NGjqVmzJrVq1SI2NpahQ4fSqlUrUlJS6Nq1K1WrVv1HQq9UKlEoFB8dTwghhBBCCG17Z5I9Z84cgoOD6dKlCzNnzsTQ0BCA5ORk/Pz8WLNmDQULFmT8+PEfFXDcuHEUL16cQYMGAWBqakr79u2Bv4aY1KpVi7t372JtbU1YWBhlypQhJSUFpVKJvr5+Zl+nEEIIIYQQWvPOJLtmzZo0aNDgH9tViXCtWrU4e/bsRwXz9PREX1+fIUOGqLfdvXuXDRs2MHfuXNLS0vD396dJkyYYGBhw9OhR6tWrx+nTp6lRo8ZHxRJCCCGEECKnvDPJViXY58+f582bNzRr1owJEyYQFBRE7969cXR0pH79+u/ccWBgIPPmzSM4OBg9PT28vLyIiIjA0NCQbt26AVCyZEmmTp1Knjx5aN++PTo6OjRo0IBKlSpRvnx5Ll68SKdOnTAwMGDu3LkafulCCCGEEEJkD4XyPbMZXVxcWL9+PT4+Ppw7d44pU6bQs2dP9QTFnBQWFpPTTRBCCCHEV65mn4gPetyltfk/6Rjiv1lamn3U499bXcTAwABTU1NOnDhB69at0dPT+9eSe0IIIYQQQoi/vHfFRwsLC3r06EF8fDxVq1bF09MTY2NjbbRNCCGEEEKIz9J7k+wFCxZw7949SpYsCUCpUqX49ddfs71hQgghhBBCfK7em2T/+eefHDp0iJiYmLdqV8+ZMydbGyaEEEIIIcTn6r1J9qhRo+jTpw8WFhbaaI8QQgghhBCfvfcm2SVKlKBt27ay2qIQQgghhBAf6L1JdosWLXB2dsbOzg5dXV31dhkuIoQQQgghxL97b5Lt5uZG3759sbS01EZ7hBBCCCGE+Oy9N8kuWbIk7du310ZbhBBCCCGE+CK8N8nOmzcvXbp0oUKFCm8NFxk9enS2NkwIIYQQQojP1XuT7OrVq1O9enVttEUIIYQQQogvwjuT7JkzZzJy5Ehat279r/+emJjIwoULmThxYrY1TgghhBBCiM/RO5Ps77//ng4dOlCrVi2qVauGtbU1CoWCly9f4uvri4+PD66urtpsqxBCCCGEEJ+F/0yya9asydGjRzl16hShoaEAWFlZUaNGDUaMGIGBgYHWGiqEEEIIIcTn4j/HZBsYGNCyZUtatmyprfYIIYQQQgjx2dPJ6QYIIYQQQgjxpZEkWwghhBBCCA37oCRbNdkRIDk5+YN3fu/ePRwdHdm6dSsAL168oFu3bnTu3JmhQ4eq9+Xp6Unbtm1p3749Hh4eAKSkpODq6kqnTp3o2rUrz549+6gXJoQQQgghRE55b5Lt7u7OsGHDmD59OgALFixgzZo1791xfHw8M2bMoFatWuptS5YsoXPnzmzfvp3ChQvj4eFBfHw8y5cvx93dnS1btrBu3Tpev37NwYMHMTc3Z8eOHfTp04dFixZl4WUKIYQQQgihPe9Nsk+cOMHvv/9O7ty5ARg/fjwnT558744NDAxYu3YtVlZW6m2XL1/GwcEBAAcHB3x8fLhx4wYVK1bEzMwMIyMj7O3t8ff3x8fHBycnJwDq1q2Ln59fpl6gEEIIIYQQ2vbeJDstLQ0AhUIBQFJSEqmpqe/dsZ6eHkZGRm9tS0hIUJf9s7S0JCwsjPDwcPLly6d+jIWFxT+26+rqoqOj81FDVYQQQgghhMgp711W/ccff6R79+48efKEKVOmcPnyZbp3756pYKpEHUCpVL71/4zbFQrFO7cLIYQQQgjxqXtvkt25c2e+//57AgICMDAwoH///hQsWDBTwYyNjUlMTMTIyIhXr15hZWWFtbU1Z86cUT8mNDSUKlWqYG1tTVhYGGXKlCElJQWlUom+vn6m4gohhBBCCKFN7x0uEhAQwJYtWzh//jwnT55kyZIljBs3LlPBateujZeXFwDHjh2jXr16VK5cmZs3bxIdHU1cXBz+/v7Y29tTp04djh49CsDp06epUaNGpmIKIYQQQgihbe/tyR41ahR9+vTBwsLio3YcGBjIvHnzCA4ORk9PDy8vLxYuXMjYsWPZuXMnhQoVwtnZGX19fVxdXenVqxcKhYKBAwdiZmZGs2bNuHjxIp06dcLAwIC5c+dm+kUKIYQQQgihTQrl3wc//80vv/zCihUrPsnx0GFhMTndBCGEEEJ85Wr2ifigx11am/+TjiH+m6Wl2Uc9/r092S1atMDZ2Rk7Ozt0dXXV2+fMmfPxrRNCCCGEEOIr8N4k283Njb59+2JpaamN9gghhBBCCPHZe2+SXbJkSdq3b6+NtgghhBBCCPFFeG+SnTdvXrp06UKFChXeGi4yevTobG2YEEIIIYQQn6v3JtnVq1enevXq2miLEEIIIYQQX4R3Jtk3btygcuXKby15LoQQQgghhHi/dybZV65coXLlyuoFYf6ufv362dYoIYQQQgghPmfvTLJv3rwJSKk+IYQQQgghPtY7l1V//fq1FpshhBBCCCHEl+OdPdlPnz5l/vz573yiVBcRQgghhBDi370zyTY2NqZUqVLabIsQQgghhBBfhHcm2RYWFrRu3VqbbRFCCCGEEOKL8M4x2RUqVNBmO4QQQgghhPhivDPJHjNmjDbbIYQQQgghxBfjnUm2EEIIIYQQInMkyRZCCCGEEELDJMkWQgghhBBCw95ZXSQ77N69G09PT/XfAwMDadOmDdeuXcPExASAXr168cMPP+Dp6cmmTZvQ0dHBxcWFdu3aabOpQgghhBBCZJpWk+z27dvTvn17AK5cucKRI0eIj49n1qxZlC1bVv24+Ph4li9fjoeHB/r6+jg7O+Po6EiePHm02VwhhBBCCCEyJceGiyxfvpwBAwYQFxf3j3+7ceMGFStWxMzMDCMjI+zt7fH398+BVgohhBBCCPHxtNqTrRIQEEDBggWxtLQkLi6OZcuWER0djbW1NRMnTiQ8PJx8+fKpH29hYUFYWFhONFUIIYQQQoiPliNJtoeHh3o1yY4dO2Jra4uNjQ0rV65k6dKlVK5c+a3HK5VKFApFTjRVCCGEEEKIj5Yjw0UuX77Mt99+C4CTkxM2NjbqP9+9exdra2vCw8PVjw8NDcXS0jInmiqEEEIIIcRH03qS/erVK0xMTDAwMACgf//+hISEAH8l36VKlaJy5crcvHmT6Oho4uLi8Pf3x97eXttNFUIIIYQQIlO0PlwkLCzsrfHWXbt2ZfDgweTKlQtjY2PmzJmDkZERrq6u9OrVC4VCwcCBAzEzM9N2U4UQQgghhMgUhVKpVOZ0IzIrLCwmp5sghBBCiK9czT4RH/S4S2vzf9IxxH+ztPy4Dl9Z8VEIIYQQQggNkyRbCCGEEEIIDZMkWwghhBBCCA2TJFsIIYQQQggNkyRbCCGEEEIIDZMkWwghhBBCCA2TJFsIIYQQXw1/f1+mTZuAv79vTjdFfOG0vhiNEEIIIYS2/L2+dMG4LRimP+H63VhemNiot0t9aaFp0pMthBBCCCGEhkmSLYQQQoivxmvDViTo2vHasFVON0V84WS4iBBCCCG+Ggl6lUnQq5zTzRBfAenJFkIIIYQQQsMkyRZCCCGEEELDJMkWQgghhBBCwyTJFkIIIYQQQsMkyRZCCCGEEELDJMkWQgghhBBCw7Rawi8wMJABAwZQvHhxAEqXLk3v3r0ZPXo0aWlpWFpasmDBAgwMDPD09GTTpk3o6Ojg4uJCu3bttNlUIYQQQgghMk2rSXZ8fDyNGzdmwoQJ6m3jxo2jc+fONG3alPnz5+Ph4YGzszPLly/Hw8MDfX19nJ2dcXR0JE+ePNpsrhBCCCGEEJmi1eEicXFx/9h2+fJlHBwcAHBwcMDHx4cbN25QsWJFzMzMMDIywt7eHn9/f202VQghhBBCiEzTek+2n58fvXv3JiEhgcGDB5OQkICBgQEAlpaWhIWFER4eTr58+dTPs7CwICwsTJtNFUIIIYQQItO0mmSXKVOGgQMH4uDgwKNHj/j5559JTU1V/7tSqXzr/xm3KxQKbTZVCCGEEEKITNPqcJGSJUuqh4bY2NhgYWFBdHQ0iYmJALx69QorKyusra0JDw9XPy80NBRLS0ttNlUIIYQQQohM02qS7eHhwebNmwEICwsjIiKCNm3a4OXlBcCxY8eoV68elStX5ubNm0RHRxMXF4e/vz/29vbabKoQQgghhBCZptXhIk5OTowcORIvLy+Sk5OZOnUqZcuWZcyYMezcuZNChQrh7OyMvr4+rq6u9OrVC4VCwcCBAzEzM9NmU4UQQgghhMg0rSbZuXPnZu3atf/YvnHjxn9sa9KkCU2aNNFGs4QQQgghhNAoWfFRCCGEEEIIDZMkWwghhBBCCA2TJFsIIYQQQggNkyRbCCGEEEIIDZMkWwghhBBCCA2TJFsIIYQQQggNkyRbCCGEEEIIDZMkWwghhBBCCA2TJFsIIYQQQggNkyRbCCHEV83f35dp0ybg7++b000RQnxBtLqsuhBCCJHTavaJeOvvBeO2YJj+hOt3Y3lhYqPefmltfm03TQjxBZGebCGEECKbSW+5EF8f6ckWQgjxVXtt2ArzZC+iDRprbJ/SWy6EkCRbCCHEVy1BrzIJepVzuhlCiC+MDBcRQgghstlrw1Yk6Nrx2rBVTjdFCKEl0pMthBBCZDPpLf9w/v6+HDiwjx9/bE3VqvY53RwhMk2SbCGEEEJ8Mnbv3s6jR0EkJiZIki0+a1pPsufPn4+fnx+pqan069ePy5cvc+3aNUxMTADo1asXP/zwA56enmzatAkdHR1cXFxo166dtpsqhBBCfNX+PoHzXbIygfOfk0TTMATuPEl7699kkqj43Gg1yb506RL3799n586dREVF0bp1a2rVqsWsWbMoW7as+nHx8fEsX74cDw8P9PX1cXZ2xtHRkTx58mizuUIIIYTQsuyo9iJETtBqkl2tWjUqVaoEQO7cuUlISCA6Ovofj7tx4wYVK1bEzMwMAHt7e/z9/WnYsKE2myuEEEIILZPx6+JLodUkW1dXl1y5cgGwe/duvv/+eyIjI1m2bBnR0dFYW1szceJEwsPDyZcvn/p5FhYWhIWFabOpQnyRZEKREEIIoR05MvHxxIkTeHh4sGHDBi5duoStrS02NjasXLmSpUuXUrny21ewSqUShUKRE00V4osiE4qEEEII7dB6ku3t7c2qVatYt24dZmZmODk5qf/NycmJqVOn0qhRI86cOaPeHhoaSpUqVbTdVCGylUwoEkIIIb5cWl2MJiYmhvnz57N69Wr1JMb+/fsTEhICwOXLlylVqhSVK1fm5s2bREdHExcXh7+/P/b20usmRFbJghhCCCGEdmi1J/vw4cNERUUxbNgw9ba2bdsyePBgcuXKhbGxMXPmzMHIyAhXV1d69eqFQqFg4MCB6kmQQojMkwlFQgghhHZoNcl2cXHBxcXlH9udnZ3/sa1JkyY0adJEC60SQgghhBDaoI2hkp8KrQ4XEUIIIUT28Pf3Zdq0Cfj7++Z0U4QQyLLqQgghxGfpnxObt2CY/oTrd2N5YWKj3v4l9AgK8TmSnmwhhBBCCCE0TJJsIYQQ4gsg1YOE+LTIcBEhhBDiCyDVg4T4tEiSLYQQQgjxhYhbWumDHmcyOCCbWyJkuIgQQgghhBAaJj3ZQvyLr6mOpxBCCCE0T3qyhRBCCCGE0DBJsoUQQgghhNAwGS4iPksynEMIIcTnzt/flwMH9vHjj62pWtU+p5sjNEySbCGEEEIILfh75Y+d1wvzJM6Q+Kf+2FUJVm+Xyh9fBhkuIoQQQgghhIZJT7YQIktk6I4QQmROq2KReAXnoXHh1zndFJENJMkWQgghxGcnu8cza2O8dOV8CVTOl5At+xY5T5JsIYQQQnx2du/ezqNHQSQmJmRLEpzd+xdfPkmyhRBCCPHJ+/ukwfTQwoAh6aG33/q3zE4azO79i6/PJz3xcfbs2bi4uNCxY0cCAuSgFkIIIcRfWhWLxM48gVbFIj/L/Ysv3yfbk33lyhWePHnCzp07efDgAePGjWP37t3vfLxMvhKfguwYw/f33pV3yUrvijZiiOz3pZwHv5TXIbJXdo9nlvHSH07qff+7TzbJ9vHxwdHREQBbW1uio6OJjY3F1NQ0R9ojJ/0P9zW9V1LzVDuy+5jSxjH7pXwvPvR1nKzU4IMe97V/N+QiWnxKMvv9lt++f6dQKpXKnG7Ev5k0aRL169dXJ9qdO3dm1qxZ2NjY5HDLhBBCCCGE+G+f7Jjsv+f+SqUShUKRQ60RQgghhBDiw32ySba1tTXh4eHqv4eGhmJhYZGDLRJCCCGEEOLDfLJJdp06dfDy8gLgzz//xMrKKsfGYwshhBBCCPExPtmJj1WrVqV8+fJ07NgRhULBlClTcrpJQgghhBBCfJBPduLj10D11mtqrHlaWhq6urrqfcsYdqEpcjz9z549e8iVKxdNmzbNlv0rlUouXbpErVq1smX/n5rP5dhKT09HR+eTvfmbaZ/L+y8EfH7fw8+npVqwa9cuJk+e/I9Jl9klLS1Noyc3XV1dlEoljx8/RqFQZOl1pKenA/+cgKpJqn0nJydnW4ysUL0Hn3rMvz8nOz6zT+0z+rf3SRvvQ2xsLPr6+jx48ABvb2+N7x8gKiqKsWPHcvjw4WzZ/8fKzu+BKsFLTEzMtv2rpKWl/ev2D3Hu3Dnu3LmjsXZ9SlS/QTlxvvsSfY39ltp8zTo6OoSFhXHw4EFSU1O1FjcsLCxTz5MkO4M2bdqQkJDAwoULs+2gUSqV3L59G4BffvmF9evXk5aWlqUTXMa2Xrp0iVGjRhEeHp7pBF6pVKKjo0NkZCS7du36z0WAMkv14/r69WumTJnC8uXLNR4jK1RXy5GRkXh4eBAQEEBkZPau+qWK+fr1a3bs2MGdO3fem3yoPiulUsnBgwdJSUnReK/UmzdvmDhxItu2bSM9PT3Hf0QyvuaQkBDS0tLe2nbkyBFev36t8fchLS0NU1NTvv32W0JCQjh16hRnz57VaIwpU6Zw7Ngxdu3axbp16zh06JBG9/+xVMekUqlk//79Gm1Peno6CoWCN2/eMHr0aJYtW4afn5/Gkj3VOSY9PZ358+ezcuVK3N3dgY+7e7h//342bdpE4cKFNdKuD3HixAm2bdumPvdkhyNHjrBu3TpmzJjB3bt30dHR0WqifeXKFY4dO4afnx/BwcHvf8InGuPkyZMcOnSItWvXqo+57DxH3rp1C19f32y9CP+YGBlf8+3btwkJCcm2dgEkJSUxZcoUHj9+jJ6enlaOWS8vL3788Uf1d/FjPl9JsvnfFbyenh7z5s0jJiaGuXPnZssXxc/Pj2nTphEYGEihQoU4fvw4ERERWTrBKRQK0tLSSEtLo1atWjg7O6sP9MzsU6FQEB0dzaRJkzA0NMTLyws3N7dMte3fqL6UsbGxnDp1ikaNGhEUFMSvv/6qsRhZpaOjQ0REBEuWLCEpKYljx45x6NChbO3RVcWcPn06CoWCZcuWcfz48XfGVCUpAJs2beLo0aMMHTpUYxcDqmPHyMiIN2/ecOHCBV68eJHtPyLva5MqvpubG1OnTmXv3r08f/4cgA0bNuDt7c2MGTO4cuWKRk/Aurq6REZG8ttvv1GvXj2KFSvGtWvXNNqjPWjQIPbv34+Pjw8rV67M8URblWCPGjUK+OvHZtGiRRrpedbR0eHNmzds3bqVtm3bYm1tjY+PD9euXdPI56b6bixZsoQKFSrQrFkz9uzZ81E90h4eHixcuJCOHTuSO3fuLLfpQxw+fBgPDw8KFizI06dPOXXqFMePH9d4jKNHj1KxYkVKlCjB+PHjuXPnjtZuwx88eJB169bx9OlTDh8+zPbt27l169ZnF+Pw4cPs3bsXAwMDHj9+zKBBg946R2mKal9+fn4sXryYx48fs3PnTpYtW0ZSUlKOxciYYC9fvpwtW7awZcsWPDw8NNImFdX5ID09HT09PerVq8eFCxd4/PixVi4Ozc3NadGiBStWrODatWvqc8uHfMZffZKt6gFLT09n/fr17Nixg+nTp5OUlMS8efM0+uEplUrs7e3p06cPK1asoHz58vzyyy9Mnz4904nRs2fPCA8PZ9u2bQwePJiQkBDi4+O5dOkSwEedNFWvNSUlhRs3blC6dGmcnZ1ZtmwZt27dYufOnZlq499jqHqvJk+ezNWrVylRogRz587l1atXzJo1K8sxsiLjl+bu3bsUKFAABwcHgoODqVy5crbenkpJSSE4OJi6devSrl070tLSKFiwICkpKf/6eNVnO2/ePKysrFi2bBnlypVj1qxZb5W/zAzV9+L169cEBQXRrFkz2rdvz8aNG3n16lWW9p0Vqte8du1aSpQoQeXKlQkODubSpUuMGTMGU1NTZs+ejYODAx4eHoSGhmY5ZsZzQHx8PObm5ur3I0+ePBw6dAgfHx+NxLC0tGT16tXs2bOHCxcusGrVKtzd3dm3b1+W9v+xMh7nL168oFixYjg7O5M3b15Kly5NRMSHrQr3bzK+n56enjx79ozixYvTunVr8ufPz+nTp7PUe5vxO5yamkpERAS5cuXi4MGDjB07FoBHjx69dz/nz5/H19eX3r17c+XKFa5evfrO76Km7N+/n2PHjjFr1iwaNmxI//79qVChAg8fPtRYjAMHDrB69WoWLFhAjRo16NKlC7169WLOnDm8efNGY3HeZf/+/Zw9e5ZZs2bRu3dvfvnlF7755psP+kw+pRi7d+9m9+7d/Pbbbzg5OTFr1iwKFSrEhAkTAM3NtVLt69y5c6xdu5YpU6bQrl071q5dy4sXL1i5cmWOxMg4ln/p0qUAzJ49myZNmnDnzh1u3LihkXap7qaFhYUxe/Zs9uzZQ9WqVenWrRu//vorDx8+zJaLw4znKTs7OyZOnEirVq347bffePLkCdevX/+gO5lffZKtugobO3Ys5ubm+Pv7M2XKFKZOnUpYWBhLlizRaKyYmBhy585Nr169OHXqFEZGRvz444/89NNPLFy48IN7Sn18fEhKSuLatWv88ssvFC1alPr167N3714ePXqEp6cnFy5c+Kj2qYZHHDx4EGNjY06dOsXp06cxMjKiQ4cOWR7Woorx5s0bli5dSsuWLbG1teXIkSOEhYWpL26ye1jGf1EqlSQkJODq6kpcXBzR0dHMnDmTQYMGERcXx5o1a7Il0U5OTmbVqlX4+/uzf/9+Jk6cyNSpU0lNTWXhwoVv/bhn/AyUSiXh4eHqi6pBgwZRsGBBpk2blumEQHXyjI6Oxs3NDU9PTwICAkhISODbb79l1KhR/Pbbb1odp53xNT969IigoCBiYmKoWbMmzZo1IygoiBcvXnDp0iVSUlJo1qwZBQoUyHJymnHY0J9//klaWhqGhoacPHkSU1NTihQpgq6uLqVKlcpyjIiICNzd3Tl9+jRLly7Fw8MDHx8flixZwt69e4mNjdXKHYS0tDT1bVh3d3cMDAw4ceIEnTp1omvXrtjb27NkyZJMXWylpaWho6NDVFQUgYGBVKhQgWLFiqnvkrRr147ixYtTrFixTLU9Y8/apk2bCAsLo379+syfPx8TExPs7e1ZtWrVe89jPj4+PHnyhPHjx9OjRw/Kli3LH3/8wc2bN7P1uA8ICMDIyIj8+fOjVCrJlSsX5cuX586dO8THx2d5/zExMejq6lK2bNm3enUbNWpEhQoVMDMzy3KMd0lPTyc2NpZdu3ZRvnx5LC0tSU1NxcLCgvz583P58mWNHN/x8fF4eHhQpkyZbIsRExNDWFgY1tbWb30uffr0wdLSMsv7/zdnzpwhLi5O/d0wMDBgyJAhPH/+/K35BpmlVCo5d+7cB8XI+D3bv38/BgYGhIeH8/DhQypXrky+fPk09juuOl+4ublRvnx54uLi2LVrF6VLl6ZRo0asXLmSlJQUjd85UJ2T9+/fz7NnzwBo2rQpvXv35ueff+bXX3+ldOnS72+/xlr1mcl4kn327Bl2dna0b98eGxsb7ty5w5AhQ6hRowYNGjTIcqyMH76/vz+///47ycnJ9OzZky1btpA/f34mTJhAhw4dMDAweO/+nj17Rnp6OoaGhsTGxlKiRAlu3bqFs7MzgwYNonv37tjZ2fH48eOPamdycjJjxozh8ePH2NvbM3r0aJYsWcLKlSvZu3cv9erVy/IVY1JSEvfu3ePhw4fkypWLLl26oKury86dO9VDJfLly5elGFmho6ODsbExdnZ2KBQKTE1NKVCgAFevXuXAgQO0atUKPT3NV75UnSQcHR1p2rQpwcHB/Pnnn/z+++/8/PPP6OvrA3/1zKnuvGzYsIHt27ezYMECdHV1mT17NgAjR47E1dVV/ZyPpVAoiIuLY+bMmVSrVo0xY8bw3Xffcf36dYyMjOjTpw+tWrX6oGNVEzKOC968eTPx8fGYmZmxfPlyHj9+zP379/H19WXatGkUKlSIiRMncvHiRWJjY7NUoUN1og0LC2POnDmsX7+eq1evYmNjg4+PD/PmzWPnzp0MHDgwSwtl6ejokJyczMSJE7GysuLkyZPs2LGDlStXsnHjRvz8/HB3d8fU1DTbq0AolUr1BGo3NzfWrl2rHkYXERHBrVu3mDZtGu3bt8fa2vqj96+rq0t4eDhLlixh27ZtXL9+XX1n69ChQ4SFhdG+fftMnwNUP/xTpkwhd+7cJCYmYmdnx5gxY/jzzz+ZNm0aLi4ulCxZ8p372L9/P9u3b6dp06aYm5sDf83XqVatGlu3blXPqdGkK1eucPbsWSZPnoyZmRkzZ85UJzX37t1DT09PXTkqswICAjhx4gS2trY0bdoUT09PdU+ct7c3jx49IjY2Nsuv5V1SUlIwNTVlypQpeHt7s2fPnrfOpaamplnuxLl9+7b6+Lp8+TK7d+/WeIyAgAD8/f1xdHRUX3CqkjBfX1+Cg4NJSkrSWNIXFBTEq1ev6Nu3Lz/88APjx48nISEB+F8uoJqXklkPHz7k4cOH9OrVi7p16zJhwoT/jKE6Dy1evJiXL1/SrVs3vvnmG3bv3o2XlxePHz/O8gVbxtfz5MkTChYsSLVq1Xjw4AFGRkbMnDkTMzMzpk2bhr6+vsbOjaq77fHx8WzevJmAgADOnDnDwYMHAShSpAi5c+dm8uTJFCpU6L37+ypL+GW8CvP29ub777/n6tWr+Pv7U79+ffUP9ZAhQ7C1tdVIzJiYGGJjYylYsCDHjx/nwoULNG3alOTkZPbt28e8efM+Oiny8fHB19eXOnXqcO/ePcLDw6lfvz5ly5YlOjqaOXPmMHr06P+8slYlMCkpKejr63PixAkWLVrEsmXLKFmyJA8fPuTZs2eULVs2Uz+qGSUmJrJ7927y5MlDnjx5OHv2rLoHZdeuXbRq1Yq8efNmKUZmLFq0iGHDhpGYmMjJkydp2bIlJ06cICgoiD59+nD9+nX1hUfRokU1EnPnzp00a9YMhUJBUFAQxYsX588//8Td3Z1Vq1Zx5swZFAoFJUqU+EevnlKpZMyYMVSrVo1Lly5hamrKtGnTcHV1pVChQri6umaqTRlLI0VGRvL7778TEhLC0KFDyZ07NwcOHODBgwcMHz5cawm2qiylUqlk3rx5KBQK8ubNS1RUFPv27cPIyIjhw4fz/Plz7t+/j5ubG25ubly7do1x48ZRpkyZTJcoe/DgAQUKFGDGjBn8+OOPmJubM2fOHFxcXHjz5g1OTk4oFAoKFiyY6den6sHNlSsXQUFBDBo0CIAhQ4bQsGFDGjZsSHR0NEWKFMl0jA919uxZrK2tKVOmDOvXrwf+Sopr1KhB2bJluXPnDlFRUeTNm5cyZcp88H5VPWQ1atTAyMgINzc3ChYsiIuLCydPnuTp06dUqFCBgIAA2rRpk6lzQMbP+OrVq3h5edG3b19mz55Neno6DRs2xNHRkZiYmP/8vDw9PTl9+jSjRo2iUKFC6sSzYsWKwF9jcO3t7bGysvroNr7L7t27OXnyJGZmZhgZGTFjxgzGjh1LgQIFKF++PIcOHWLw4MH/eWHwIe7du8fhw4cpVaoUpUqVIjg4GF9fX9LT0wkNDWXIkCEUL15cQ6/qbQcOHCAwMJCff/6ZAgUKcPfuXebMmUPXrl0xNDRk586dDB8+PEuvcefOnZw9e5bIyEg6duxItWrVmDBhgkZjwF+TUgMDA6lRowb58+fn+vXrhISEYGlpib+/P4MHD+abb77JUgwVPz8/FixYQM2aNbG0tKRly5asXbuW5ORkmjRpwpYtW2jXrl2WOhOuXr3Kr7/+SpUqVWjQoAEVK1Zk8+bNhIaG0qJFi7diZPyeBQcHs3HjRnLlysUvv/xCdHQ0f/zxB4GBgbRo0YJGjRpluk2q36Lo6GiSkpLQ1dUlJSWFNWvW0KFDB9LT0zly5Aht2rTR2HsNf3Vi6enpERkZyc6dO0lMTGT48OGcOXOGwMBAChQoQOnSpbG2tv7gfOir68nOOFls/fr1bNy4kWnTplGhQgXi4uLYvHkzixcvpl+/fhpJsO/cuUNYWBgXLlxQX/E6OTlRqVIlfv/9d8zMzJg1a9YHJdjp6emEh4czevRoVq5cSVxcHEZGRgQEBFCsWDHMzc0ZPnw4jx49IjExkZiYmP9MhlQHckhICIsXL2by5MkUKVKEwYMHM3bsWO7evUvJkiX54YcfMp1gq67hUlNTUSgU6OvrExISQmxsLA0bNmTfvn3cvn2bHj165EiCDdC8eXOSk5O5ePEijx8/ZtSoUdjY2HD58mXOnDnDt99+S+vWrTWWYMfGxmJvb4+xsTHLly/H39+fSZMmYWVlRc2aNbl+/ToNGjTghx9+UCfYHh4eHDhwAIDnz59TunTpf73z4ujomKk2qW7jh4eHs3LlSi5dukTp0qX59ttv2bRpE1FRUTg7OzNw4ECtJdienp7qCY23bt0iMTGRMWPGULhwYby9vXF1daVly5Zs27aNvn37YmtrS9++fRk2bBh2dnZs376d5OTkTCXYBw8eZNq0aWzfvh0zMzNSUlLw8PBgwIAB3LhxA29vb4yMjLKUYKuqVjx//hx/f3/Onj3L3r17AWjbti0RERGYm5trJcFOTk6mQIEClClThrNnz1K7dm169eqFiYkJsbGx3L9/Hz8/PypVqvRRCTZAREQEhoaG6h/NwoULkydPHgAcHBx4/PgxxsbG9OrVK0sJtqraTN68eYmOjmbv3r107tyZXr16ERgYiIGBwXsT7MWLF9O+fXsKFSrE9evX/3FubtasmUYT7JcvXxIYGIibmxsLFixQv/65c+fy4sUL1q1bl+XE8N69e8TGxlK6dGl69+7No0ePePLkCYUKFVJf3AwYMCDbEuyEhAT8/f0xNTVl8+bNvHjxAjs7O8aNG8eaNWtYvXo1o0ePztJrfPLkCVeuXGHFihUsXbqUx48fU7hwYcaNG8fatWs1EuPFixcAODo64ujoiJ+fHxEREVSuXBkTExN27NjBoEGDNJb0vXjxglWrVjF+/Hh++OEHQkNDMTMzY8CAAQD07NmTDh06ZCnBfvr0KStWrGDRokUYGxtz4cIFtm7dSoMGDTA3N6dXr17qGBm/Z7t37yYqKopKlSphYWHBvn37MDY2Vt/xuXfvHg8ePMh0u1RD9CZMmMCRI0cYNWoUSUlJlClThlWrVrFhwwZ69Oih0QQbUCfYHh4e6txq27Zt/PDDD5QuXZqoqChsbGw+Kh/6ZFd8zC6q286//fYbFSpUoHfv3kyZMoVVq1bRtWtX9QFWvnz5LMVJT0/n7t279OjRg4EDB1KuXDlq1KjBpk2bGDJkCIUKFaJAgQIUKlQIY2PjD9qfjo4OFhYWmJubs3v3bo4cOULZsmU5fPgwgYGBNG/eHEdHR/UtjFmzZv3njPirV69So0YNVqxYQdeuXblz5w5Tpkxh586dxMfHM2PGDDZs2JClWzEPHz6kcOHC7N27l5o1a+Ls7MzRo0d5+PAhJUqUoGXLltl2cv8QkZGRWFlZcezYMZYvX87hw4c5fvw4/v7+JCUl4efnx/fff5/lW7UZGRkZUbx4cW7evMnr169p1KgRVatWZd26dTx58oSbN2/y7bffvvWcdu3aAbB582a6d+/Ot99+y8qVK3F0dKR79+5MnDiRqlWrZurCUNU7GRUVxZIlS6hbty4hISE8f/5cnezt2LGDwYMHY2pqqpH34ENUqFCBokWLcu/ePfXxf+LECdatW4ezszMPHjzA1dWV48ePM2jQIPXwhocPHzJ+/HhmzZrFmzdvPnqM5J49e/D392fu3Lk8ePCAuLg4vvnmGypWrKgeLztkyJAsXRTu3r2bP//8kyVLluDp6Ul8fDw//fQTI0eOJDw8HH9/f/UEquyWlpaGgYEBpUuXxtvbm2vXrlG6dGnKli2LiYkJv//+O4aGhrRp0wYTE5OP3n96ejo1a9bk119/xcLCgoIFC7Jz5050dXVJTk5GqVRmaYhYxmozDx48wMHBgXbt2lGsWDHCwsJYu3YtP//8839eHKqGTnTu3Jlt27bx+vVr9u3bR8+ePT/6ouJDnT9/HnNzc6ytrYmOjubVq1fqWry3b99mxowZxMbGZum98fHxYcaMGdSpUwddXV3KlSuHgYEBN27cICQkhBYtWlCvXr1s/V4bGhry/fffU6xYMc6cOcPmzZsZPnw4JUqUYPHixejq6mbpYhXAxMQEKysr0tPTefHiBTExMcBfE4mXLVtGampqlmKcOHGCtWvX0qRJE6Kjo9W9tLdu3aJEiRK0atWKtm3bamSooyqZLViwIA4ODtja2pKQkKAe5vL69Wt69epF+/btM33RoIphbW3NTz/9hL6+Prdu3cLFxYVHjx5x6NAh+vTpQ5MmTShduvRbCfb8+fPJmzcvly5dIjExkQoVKnDz5k327dvHTz/9RI0aNbh27VqW3ovU1FTWr19Pu3btsLe35/Lly4SEhFC+fHn09fWpUqWKRoeVzpw5E11dXcaNG8esWbPUd4cdHBzU1VV69uxJXFzcR58Dv5rhIhlvhSclJdG/f3++/fZbhgwZAvw1jtXCwoKRI0dqdMzt7NmzMTQ0xMbGhiJFivD8+XO8vLwoWLAgAwcO/KAEQHWAp6WlcejQIb755huePn3K9u3b2b59O9evX+f27dvUrVuXokWLftDt8RMnTjB48GBWrVqFp6cnXbt25dixYzRu3JikpCSqVq1KUlJSlk6+hw4dYu/evbRo0YJXr16ho6ODg4MDNjY2TJ8+HQsLC37++edM/XBrQnR0NFu2bOH27dvMmzePvXv3cuzYMZYuXUqePHm4fPkyVlZW2NjYaCReeno6r1+/ZurUqSQnJzNgwACePXvGlStX6NGjBzY2Njx+/Jj09HRKlCihfk7GcfC9e/emWLFiTJ48mSVLlhASEoKxsTHt2rXL1IWht7c3+fPnp1y5cty9e5fdu3fTqVMn1q9fT5cuXbCwsCApKYncuXNrrYQZ/DUOUaFQ8ODBAw4ePEivXr04cuQIN2/eJCYmhm+++QYXFxe2b99Oy5YtefHiBV5eXmzdujVL8wYePnxI3759GTVqFE2aNOHGjRts376dVq1acfToUfWdpKz0oDx8+JBffvmFqVOnEhUVxd69e9HX16datWo8evQIe3t7vv/+e63MTVANx0lPT2fOnDmUL1+evHnzcv36dfUEpgULFjBixIh/XPi9j+rYHTJkCCYmJri4uHDgwAGqVq2KoaEhr1694tWrV+qEOCuWLFmCubk5rVq1Yu/evSgUCurXr8/JkyepWrUq9vb273zu7t27uXz5MqNGjcLa2pply5Zx4cIFOnToQOvWrbNlhbn9+/fzxx9/sGbNGnVPeUhICN7e3tjY2ODm5sZvv/2WpUl0Pj4+GBoasnv3bkxMTBg2bBiXLl3i/v37BAYGcuPGDTw9PbPlOFO9Z8nJyaSlpak7kp4+fcrly5c5dOgQKSkprFq1KtNjd1UxkpKS1BdQCoWCgIAAnj17RqFChXBzc2P58uVZvoj4888/mTVrFk5OTlSuXJl79+5x48YN7t+/r75roonfsYy13SMjI0lISKBo0aIEBQVx69Ytypcvz8qVKxk0aJC6c+pjh8NlTJgjIyNJTU3F2tpafS7w8fHh3LlzjBkz5h/PuXv3Lrt27WL48OHqYXRVq1bl9u3bWFhYqNukGn76MVSfp2rIxpEjRzhz5gypqamMHTuWp0+fkpCQQN26dT9qvx8iOjqawYMH4+TkRJMmTejVqxc9evSgdevWBAUFsW7dOsaMGZOp38Cvoif77xOnLC0tmTRpEvPmzWPjxo38/PPPLFy4UD3BJKtUPYMAlStX5tq1axgYGBAcHEzFihWxs7Mjd+7cH5VgK5VK3N3defnyJevXr2fNmjUkJCTg7OyMjY0NQ4cOVQ9neN8XzsPDg4sXL/Ljjz9iZ2dHxYoVGTBggPqHasWKFZQoUSJLJ/jDhw9z5coV9S3Qq1evEhwczMGDB7G1tcXCwoL27dvnSIKtOh7Mzc0xMDAgNjaWmzdv0q1bN3R0dOjcuTObNm2iRo0aGo+ZL18+atasydatWzEwMMDJyQmAVatW0aFDB7777ru3nqc6bmfNmsU333zDkiVLGD9+PPPmzWPkyJGsWLGCOnXqZCrBPnToEKamppQrV469e/dSsmRJ8uTJw9SpUxk9ejQpKSmsWLGCsWPHftDdFk169uwZM2fOZOzYsXTt2pU9e/bQtm1bypYty/3797l79y6LFy+mV69e6l6lBg0aZDkZUg2XunPnDq9fv+bKlSt07NiR6tWrU6ZMGQwMDLL8g12kSBEGDBjAhg0bUCqV6gWpjh8/jkKh0HgvzX9JTEzExMSE8+fPExwcTMOGDalVqxbp6enqW/zjxo2jXLlyH7XfjIlppUqVOHv2LOHh4QwfPpxff/2VkiVL0qxZs0zfDfh7YmFiYkKVKlXImzcvHTt2ZOrUqdSvX58+ffr85/nw4cOHrFy5kjFjxmBtbU1AQAABAQHUrFmTnTt3UqlSpSyP4f07Dw8P3NzcmDJlyluJiGp+jqqkaVbOv/v372ffvn1s3LiRKlWqMGTIEPbs2cNPP/2kHlKWmV65D6Wjo8OrV69YtWoVqamp2Nra8tNPP6krySQkJDBnzpwsTY7LGCMlJQVbW1t69OgB/HU3qnDhwkydOjXT31fVMZyYmEi5cuWYO3cuEyZMoFKlSri4uNChQwcAwsPDNfI+ZsxTZs2ahZ6eHiYmJvz888/o6Oiwfft28ufPT7du3d66+/sxCXbGGKreW1NTU/r378+9e/fw8fHh5s2b6temVCp59eoVa9asoUSJEtSqVQuFQsH48eNp3749RYoUYeHChYwYMQJTU1P19/JjE+yMY6FXr16NkZERZcqUoUCBAoSHh/PkyRPWrVvHxIkTP2q/H/p+qPKegQMHoqury4YNGxg4cCCJiYl06tRJPbkyM76anmyAcePGUbZsWR48eEDx4sWpWbMmixcv5rvvvlOPc8qqU6dO4e7uTv/+/SldujQWFhasWbOG/Pnz8+bNG/LkyUObNm0+er8TJ06kYsWKuLi4cOjQITZt2sSvv/6qXtmxUqVKH7SfP//8k3Xr1jF37lzOnj2rLne2detWrly5Qq5cuejXr1+Wem+Tk5OZMmUKLVu2JHfu3AQGBnL69Gl0dXXR1dXFyMiIQYMGaWyM88dQfanCw8M5evQoTk5O3Lhxg1u3blGvXj1sbGzw9PTEyclJY2NhM5ZoO3XqFLa2thgbGzN79mwmTJhATEwM8fHxlCtX7q0qFaoT1tSpUylUqBBNmzalaNGiJCYmMnToUAoWLMjUqVMz1aZnz55x4sQJUlJSKFq0KMHBwejp6akriyQlJREWFkafPn001pP/PqqeFPhrLoOHhwdhYWFMnjyZx48fs2vXLtq0aUOlSpU4dOgQXl5ejBkzBltb23/Mes+KhIQEjh07xurVq+ncuTNdu3bN8j7/LikpiYMHDxIQEECnTp2ybUjCu6Snp5Oeno6rqytKpZIuXbrw7Nkznj59iqOjI5UqVeL48eOULVv2o78HGevaHjx4EFNTU+rUqcPcuXNp3bo1efPmxdPTk2HDhqmrd3yMjB0PS5Ys4bvvvuPOnTscOXKEuXPnYmRkxIoVK+jZs+d7SysmJSVx5MgRHj9+TIECBfD29qZbt27UrFmT/fv3U7169Q+qIPChzp8/z8GDBylTpgzBwcE0atSIKlWqoK+vT3x8PP369WPq1KlZSuwzJvGqi/jExETGjx9PkSJFGDFiBPDxPaAfQrXPlJQU5s+fT5s2bYiLi2PWrFksWrSI4sWLs2jRIjp06JDpu0H/FWP58uUkJibSs2dPNm3alOWhiKokPikpSf37MH/+fLp166aRymP/ZvHixZQtWxZ7e3t2796NQqGgZcuWLFiwABcXlyyNwX5XDNXvj2poXLVq1d46PkaMGMGtW7fw8vLi4MGDnDt3jlq1anHp0iWcnZ2z1KY3b96QO3du9WTDUqVKkZCQwKFDh2jZsiXGxsacO3eOnj17ajRnUL0+VZk+c3NzmjRpwoABA2jRogUODg64urqybNmyLFV2+qKT7Iw/2i9evGDDhg3qcY5z5syhaNGitGrViidPnlChQoUsxVL9sBw6dIht27ZRoUIFLCwsMDIywtbWFhMTE/UXXjXp50P2p7J69WqOHTvG1q1bMTY2Zu/evWzcuJFdu3Z9VC+jqtaumZkZu3fv5tWrVwwaNIgLFy5w9uxZBgwY8EHte5+DBw9y6NAhLC0t1eV99uzZQ758+bI061gTwsLC1Au4hIaG0r17d54/f87Nmze5efMmq1at0vgYxcjISEaNGoWDgwPu7u4sWrQIIyMj5syZoy6/p+q5+vtn7+7uTqVKlahatSqxsbEcPHiQjh07cufOnUwlZ6r9BwQEsHPnTooVK4alpaW6jm2FChUwNjYmT548WltKWtWm9PR0Vq5cSfHixSldujSPHj3Cy8uL7t27c+7cOZo2bUqpUqXUydG1a9cYPny4Ro7ZjJKTkzl69Ci+vr60a9fugy9iP0ZCQgKnT5/m9OnTdOvWLVti/JuMx5enpye//fYb48aNw9HRUV1nv379+v85xOJdMq7mOnjwYFq1akXJkiWpWLEiDx48YNasWeTJk4dJkyZlqbdeqVSyZcsWkpOTiY6OpnTp0ty/f5+oqChSU1Np1arVB9+JUl1UrVixgs6dO/PTTz9lul3/xcfHh6CgIHWlmr179+Lv70+bNm0oV64cRkZGb/1mZca/JfGquQQJCQlMnz6d0aNHZ+sk8/j4eB48eMCWLVv4/vvv+fPPP2nVqhWRkZEUK1aMggULZnmOy7tihIeHU6pUKczNzTN99+1dSfyMGTPYvn07QUFBrFixgsWLF2NkZJTlC5WM38eHDx+qFxXr06cPwcHB7Nq1i4SEBLp16/bBw0E/JsazZ8/w9PQkJSWFYcOGvfUepKamcvr0aYoVK8bp06fV80hUVWl0dXX/cff1Y1y6dEm9jz179rBlyxY8PT2Bv+7G7N27l4ULF5IvX75sKZ0bExPDhAkTaNq0KcePH6d48eL06tWLn376iS5duuDs7Jzlu6NfbJKd8bbIkSNHqF27Nn369KFdu3a4uLhw584ddu/ezZgxYzRWLSEmJoYXL17w7Nkzzp49S58+ffD29iYkJISbN2+yfv36DzpQVLdO0tPT2bx5MwkJCXTq1Ek9zGPJkiWYmpry6tWrLJXV+/PPP7l37x42NjasXr2akSNHqscCZ1VaWhqvX78mX758KBQKrly5wo4dOxg6dKjGZwR/jOTkZK5evcrTp0/58ccf+eWXX9Q9bfny5aNWrVoa+wFSnagSExO5ePEiiYmJNGvWjGHDhnHt2jWGDh0KQP369cmfPz/wv89eqVTy+++/U7duXby8vEhMTMTBwYG0tDSWLl3K5MmTs5QAR0ZGsnTpUr755ht1FQvV3RZ9fX26dOmikffgY6jK9OXOnRtzc3Pu379P165duX37NsuWLWPJkiXY2dmpH5+QkEB0dHSWS0u+S3x8POfOneO7777LtgUm4uPj8fb2pmrVqtkWI6OMFzM+Pj6kpaVRsWJFevfurf5u+vr6Uq1atY/uNdq8eTNlypShevXqREZGsmHDBkaOHAmgLnNWt25dYmNjM9U7/PcV5lQLWKkm5pYpUwZHR0cSExMxMjL6qH1n90XV/v37OXnyJNOmTXvr4uKPP/7g7Nmz/PTTT1SuXDlLMf4riS9btmy2D/lKSUlBV1eXLVu2kDt3booVK8aQIUMYPXo0lSpVYunSpYwePTpL31dtxIB/T+JbtmxJREQEZcqUUQ81zKqM5/szZ85gaGiInp4eV69eJV++fHTq1EldgzuzvbgfEiM4OJiUlJS3fpuVSiWrV6/m2bNnPHnyhA0bNrBixQquXr1K0aJFmTZtGoaGhpl+7Zs2bcLOzo6aNWuydu1aHBwcOHz4MMHBwUyePBljY2P27dtHrVq1KFCgQKbj/J3q/YC/zksHDhxg0qRJALi6ulKjRg1atGhBVFSURjqZvtgkG/53kKjqPLZp04bhw4fTqFEjrl+/Tt++falatarG4l25coV9+/YxZ84ctm3bRkhICG3btqVEiRJER0d/0K3RmJgYzMzMSE1NZebMmVSvXp3Q0FBOnjzJli1b+PXXX7ly5Qpbt24FyFKPwPPnz+nRowe2traMHTs2W5Lf69evc+7cOQIDA5k4cWKWJzhlRXJyMjNnzsTBwQFzc3M8PDxwdXXFw8OD169f06VLF4333CYlJbFnzx5KlChBbGws58+fZ+LEiezfv5/Lly8zduxYdYIdGxuLqakpaWlpzJ8/n4SEBPLmzUuFChUICgrC0NCQc+fO0b9/f6pXr/7RbcmYpJw9exZvb28mTpzIw4cPOXDgAEqlEjs7O+rUqaO1SY5/HyKyf/9+9fwA+KvEWYsWLahatapGS6d9qOy4pZ4TMf4eb+TIkdSoUUNd5q5MmTLqiT1jx47N9PCVxMREfv/9d5ydnRk2bBg//PADPXr04OTJkwQEBDB48OBM9UhlHCLy/PlzdQUEJycnHB0dOXfuHDt27GDBggXkypUrU71P2XVRpY3a29pI4t/l2rVrb02Kff78OePGjWPkyJFERERw6NAhTExM6NGjR6Y7cbQRA7SXxGeUmpqKm5sbKSkpmJmZ8eDBAzp27Mi5c+coXLiwRjo8PibGrVu3CA0N5datW9jY2NC8eXPWrl3LpUuXWLp0KRcvXiRPnjyZutOlcvPmTW7dusWDBw+oW7cuz549U0+CPnv2rLo+eFaS+P8SFRXF48ePqVixIj///DPNmzenY8eOeHt7c+fOHfr06aOxWF/0xEc3NzeKFi3KyZMnGTx4MIaGhqxcuZLnz5/ToEGDt3rFMsPLy4u6deuiVCo5ePAgVlZW6oSpS5cu7Nixg3379jFkyJAPmuDh7e3NtGnT2LZtG5aWlqSkpPD999+zYcMGOnTooF4I5NmzZxopKWdqakqhQoWYMGFCto2PLlWqFCYmJrRv3z7LZZqyysDAgG+++QYLCwtKlSrF8ePH8fX15enTpwwbNixLq/a9S1xcHDdv3sTOzo4GDRrg4eGhLmE4cuRI9fFy9uxZ7ty5Q7t27bh37x76+vqMHj2agQMHEhERgY2NDVWrVqVdu3aZGsqi6sGMiYlBR0cHS0tLTE1N1UNO8ufPz4MHD6hVq5ZWq4ioFpqZPXs2/fr1o1mzZqxevZqGDRtSpEgRtm3bRu7cuXMkwQbNjPH+FGJkvF0cGBhIsWLF6NChAwEBASQmJmJra8uGDRuIiIj46DKQqgul2NhYfHx8OHPmDMnJyeo5B+Hh4dy4cYMZM2Zk+pavKsGeNWsW1tbW9O7dG2NjY/bs2aNecrx///5ZGuaVK1cuGjdurNHPQ1V7e/bs2era28uWLVP38MNftbezGuPs2bOMGzeOfPnyvZXEt2rVCn19/Ww796qGGYaGhnLo0CGaN29OyZIlGTZsGM+fP6d58+ZUrlwZY2NjcuXK9cnGUCXxqsltDg4O6iR++vTpHDp0CF9fXwYOHKiRBNvLy4s///xTvdBJVFQUs2bNQqlUsmLFCoKCgmjUqFGWJoZmNsabN2/YsWMHN27coGHDhjRv3pw+ffqoy9ht3bo1SxfKZ8+excrKivr165M7d27Onz9PzZo1MTExwcPDg9atW6Onp8ebN280et6fMmUKdnZ2dO7cmRkzZvDy5UuqVq3K/PnzGTNmDI8ePeL58+dvVVXRhC+qJ/vvY1lnzZpFlSpVaN68OQCNGzemTZs29OvX718f/zFevnxJZGQkdnZ2HDlyBAMDA65du4aXlxfNmzcnJCSETp06YWtr+0HjRVUHoLu7O6dOnWLFihW4u7uzbds2Fi9eTMWKFZk+fToDBgzQaI9zZm6tfi5Wr15Nt27dUCqV+Pr68t133+Hr68uBAwdYtGgRK1asICYmBhcXF429p4cPH6ZOnTrqWdk2NjY8evSI3377jQULFnDo0CH8/Pz+MfHz0qVLBAQEkCtXLurUqUNSUhI7duygTZs26oVAWrdunakebJWIiAhmzJhB7ty5qVevHnfu3CExMZG0tDRCQ0MZOXKk1i6EMn73vLy8mDt3LqdPnwZg165dBAYGEhcXx6BBg7CxsdF6b++XJON79+jRIywtLWnfvj0WFhZMnToVQ0NDNm/ezODBgzP9ox4TE8P06dNp37491atXZ9SoUeoFk9LS0lAoFFmumLJs2TLS0tLo168fK1asoEaNGjx//pxLly5Rp04ddS35T4Uq+bWzs+PGjRs0b96cffv28fPPP1O7dm2NxVAl8bVr134ridfGZNrIyEguXLhAdHQ0QUFBfPfdd1y+fJno6Gh0dHSYPn16litvZHcMVRL/3XffvZXEv3nzhpcvX9K8eXMiIiKylMRnlJqaysOHDzl8+DCFChWiZs2aeHl5Ubt2bSpUqICXlxevX7/GxcUlx2IEBATg7u7OpUuXaNasmbqqx7Nnz7LcIXf9+nWqVKnCvn37MDExIT09HV9fX2rWrElYWBihoaEMGDAg09U8/k6pVHL8+HGqVavGwIEDsbKyokePHlSpUoW+fftSuXJlBg4cyPXr1ylcuLDGh+19MUl2xuWX9+/fT4ECBbCxsWHAgAH06dOH2rVr89tvv/Hnn3/SpEkTdamfzFD9aMXHx3Px4kVu3rxJ9+7dyZ8/P7Nnz+aHH35AX1+fYsWKffRV76ZNm/Dx8eHp06esXr2a/fv38/TpU1JTU7O8utPXRKlUEhQURNGiRdm0aRO5cuXC29ubGTNmsHz5cho3bkytWrVITk7W2Jj8qKgowsLCKFWqFAsWLECpVJKSkkK1atVISUnBxsaG8uXLk5SUpL4NlnG4xM2bN/Hz8wOgUaNGnDlzhvz583P+/Hl69eqVqQsB1bGanJzM8OHD6dq1K2FhYRw/fpzOnTsTHBzMN998g6WlpdYWBco4X2LdunU4Ojqybds2kpOTmTp1Ki9evOD48eOUKVOGmjVraqVNX6qMFzPLli3j0qVLtGvXjtKlSzN8+HAGDRrE0aNH6dGjB9WqVfvo/d++fVu94uzSpUupX78+/fr1Q6lUMmjQIGxtbRk+fHim2v73C6uTJ09y+fJlXrx4QbVq1Xj69Cnt2rUjOjqarVu30qFDB+rUqfNJXIxpo/a2NpL4d8nY/tjYWE6cOEFsbCxFixalRo0abN68mevXr6srmnyqMUA7Fwp/f02q8qhPnz6lUKFCmJmZYWhoiFKp5MGDB7Rq1SpT30dNxVAqldy4cYOtW7dy/PhxWrVqxfTp07N83O7fv5/ff/+dkiVLUqlSJeLj4ylRogRJSUmcPn2axo0b891332WpB//vEhIScHFxwcXFhVatWvHzzz/TqFEj+vTpQ3p6Op06daJWrVrqSZ+a9sUk2fC/sYYVK1bE09OTXr16UadOHVxdXbG0tMTV1ZVHjx5x/fp1evbsmalbHqqDLCoqisWLF6uXBVZ9kJcvX0apVH7ULcBbt24RFhZGTEwMERER9OjRg9WrV3P48GG2bt1KXFycutSaeD9V4pqQkICfnx/Lli1j8eLF6gUEfH191QviaIrquIiNjeXy5cvcv3+fsmXLYmtry+LFiwkODsbMzIy1a9e+teCAjo4OaWlpuLu7Y2try7Nnz9DX1ychIYGQkBCMjIyoWbNmpn40Mx6rcXFxPHnyBF1dXU6fPo21tTX37t3j1atXrFixQmt1sDOOrd21axfHjx+nQIEC9O3bl6NHj/Ls2TNmzJihlbZ86TJezHh4eGBubk6dOnXo2bMnnTt35vvvv+fChQvq1Swz4/bt2wwbNow+ffrQtGlTRowYQb169dRlDzM7OTvjcTJ79mysra0xNjZWV6kwNDRkzpw59OzZk0qVKuHj40OJEiWybRLsx3j48CF9+vRhzJgxNG7cmICAAJYtW0b58uXx8fFh1qxZWa69nRML6KioPpvIyEhmzZpFpUqViIqKolixYkRGRlKuXDlq166dpQ4MbcTQVhL/b3FnzJhB+fLlsbCw4NatW8TGxtKwYUPu3LlDxYoVqVKlSo7HUCqV+Pv7s3nzZlxcXLJ84bZ7925u377N8OHDmTJlCiVKlMDQ0BBdXV2KFi2qXiNAk8M2VYvivHr1iu7du9OxY0fatm1Lnz59aNWqFZ07dwb+qj6XXXdxP/sx2WvWrFGv6HXt2jXKlStHjx49ePHiBQsXLiQ4OBhnZ2cKFizI9evX2bJlCzNnzsz02EDVl/LIkSNERERQr149ypYti5eXF5s3b8bCwuKjx9i9efOGLVu2kJaWRtmyZQHo168ft2/fZuDAgWzYsCFbytd8qXR1dYmIiGD69OmMGzcOFxcXZs6cybhx4/jhhx+4d++eRpdJh7+Oi9evXzN//nwaN26MhYUF169fp0iRIsybN48nT568tSqZ6jlKpZJNmzaRlJREVFQU9+/fx8bGBgMDAwoWLEi7du0yfVWvWnFtxowZNGvWDAcHB44dO0a1atUoUqQISqUSV1dXrS40o3rtM2bMoEyZMkyZMoU//viDFStWMGDAAP74449MlyYU/6NUKt+6W3D69GkaNmyIqakpS5cupXfv3sTGxmaqBrgqOYmLiyMyMlI9XyQwMBA3Nzd+/vlncuXKRZs2bTKd9KoS7HXr1lGmTBkKFizIsWPHOHfuHHZ2dixevBgXFxd1FZBP6Q5fkSJFGDJkCLdv3yYqKgpvb2969uxJzZo1KV68eJa/bzmxgI5Kxs9+xowZNG7cmCJFinD+/HmePn1K4cKFCQgI4Ntvv83069RGDNX349+S+Pv376Ovr0/fvn01dqcz4x3L9PR00tLSqFChAmXKlMHExITff/+d+/fv061bt08mhkKhoGrVquqSiFkZsvfw4UPWrFnDpEmTuHz5snoV4bi4OPT19Xn58iVt27bVyHCcjPT19QkPD2fdunUMHjyYFStWoFAoWL9+Pe3bt0dfXz/b54t99plbt27dMDY25vr161StWpWoqCgWLlxI586dad68OStWrGD06NEUK1aMc+fOMXv27ExdlSqVSh4/fszixYuxs7PD3t6eFy9esG3bNtq0aYOjoyMXLlzAycnpo2v21q5dGyMjIzZt2sSpU6cwMDCgRYsWWFlZ0b17d0mwP1DGk8Dp06cJDAykQIECtG7dWr1K1YQJEzSawGWMeejQISIjI9VF/o8cOYKHhwdNmzZ9qyRYxuds27aNy5cvs2TJEvT09NDX1ycmJoZSpUphaWmZ6QT79OnThIWFoVAo0NHRITo6GoVCgYGBAceOHePly5dMnz5dK2Xj4J/zH8zMzNDR0aFo0aL89NNPTJ8+nbVr1zJs2DDy588vY7CzKGOpOzMzM9zc3Pj111/Zv38/zs7ObNiwQV0a7GOokpPw8HCmT59O9+7dadGiBQ8ePMDb25tbt27RuXPnTNfOzfi5nzt3jpMnTzJp0iTKly9Peno6x44dIzk5mWnTpml1gu7HMDQ0pHHjxm/V3lYNe3J2ds7y/rM7if8vqnOJkZERNjY2VK9enXz58qGjo4Ovry8tWrR4axn1TzGGNpL4v8dTDWW9dOkSRYsWpXz58uoVVXPnzo2JiUmWqr9kVwyFQqGuipaV83GRIkUYOHAg27ZtQ6FQsGrVKgAWLlxIQkICPXv21GiCnfE8cvHiRfT09GjRogUNGzbExcWFxMREduzYQWxsrMZivstnnb35+/vz+PFj7ty5w+bNm1mzZg0ODg6cOHGC8+fP8/z5c4YMGaIu6dOgQYNMHSiqD0zVw3j06FEGDRqElZUVR44cYfv27XTu3Fk9KzYzqlatql46/cCBA6SkpODo6KjREoNfsozDNdLS0qhTpw6RkZGMHDmSKVOm4OzsjL6+vkZ/mFUxo6OjSU1NVfce7d+/nx9//JEmTZpw4sSJty7qMiab6enp2Nraqle66tq1K6ampnh7e9O+fftM97YfOHCAffv20apVK0qXLo2pqal6hUdHR0dKly6NsbGx1pbuzjhs4dixY+TKlYvvvvuOBw8ecPz4cUqUKEGhQoVISUnh9OnTtGvXThLsTMp4fIWGhqKvr6++2OrXrx+//fYbKSkptG/fPlMXWKq676tWraJHjx5YWFjg4+PDy5cvcXJyYsWKFUyZMiVLdbCVSiUvXrygWLFidO/eXb3qXd26dUlPT6dgwYKfbIKtYmxsTNOmTVEoFPj6+hIQEKCx2tvZncS/z/r163nz5g2mpqacPn0aJycnIiMj8fPzo3379hpZyCs7Y2jjQuHv8dLT05k9ezbR0dGUKlUKIyMjkpOT2bFjB9evX2fMmDGUL1/+k46RFYaGhjRt2pS0tDQCAgJ48eIFwcHBPHnyhNGjR2fL73JSUhKpqanUrl2bbdu2cfLkSRwcHPjll19Yv349Xbt21egwoHf57MZkq07Efn5+rF27lsaNG5OQkMCqVatIS0tj8eLF5MuXjxUrVtClS5cs1XKE/31gkZGR6pn558+f5/jx42zcuJGzZ88SHh6Oo6OjRg6Uq1evsnHjRkaNGqW15ay/FJGRkUyePBl7e3suXbrE6NGjOXv2LL6+vsyfP19jk1cyioiIYNGiReqkonDhwsTHxxMeHk7Hjh2xsrJSJ4wZk4gpU6ZgYGCgrkQSGBjIq1evMDY25scff8x0r8b58+fZsmUL06dPx9ramrt377JgwQIcHBwICwujfPnyODg4aOz1f6j09HTmzJmjHlP5/PlznJycuHbtGoGBgcyZMwd/f38iIiL4+eeftd6+L0HG42vWrFlUqFCBtLQ00tLSiImJoXnz5iQlJZGQkPDRd3MyJu8pKSkMGDCAsmXLcuPGDXr16sXJkycZMGAA+fPnz9KdN9V3IyUlhRIlSlC+fHni4+M5deoUXbp0ybEkIbOyc0EjbaxK+vd4BgYGPH36lMuXL1OpUiX++OMPFAoFISEhjBw5MsvrDGgjBsCvv/6qTuJtbGxwcnIiICCA3bt3M3fu3Cwn2H+/E+fu7k5SUhIdO3ZkwoQJlC1bFn19fTp06EBERESmhvdoI4amqVa53bFjB6mpqcybNy9b1s549eoVy5cvJy0tjUqVKlGhQgWmTp1K7dq1CQ0NZciQIVqrpPXZ9WQrFAouX77MggULcHNzU1+JWFhYsGDBAlxdXRk7diyLFy/WSG+Y6tboyJEj+fbbb7l8+TLr1q0jNTWVdu3aUapUKUaNGqWxK7Fq1apRqlQpjS8T/aXKWD1jzZo1dOnShbJly3L58mUiIyPVvaKxsbEaT7JTU1P59ddfcXR0xMrKivXr19O0aVOSkpK4fv06BgYGbx2Dqj9PmDCB6tWrY2BggKenJy4uLtja2pIrVy6SkpKy9GMZFRVFw4YNsba2JjIyks2bN9OwYUNCQkIoXLjwWws6aJOfnx/p6el069aNWbNm0axZM8qUKUOjRo24evUqfn5+nDt3jkGDBuVI+z53GX9w//jjD4oUKUJKSgpRUVHkzZsXIyMj9u3bR69evT56jKkqwY6IiOD06dPkzp2bFStW8OjRI1q3bk1ycjJRUVEYGRllqX4uwIIFCyhTpgy1a9dm0aJF6Orqqldi1VQVIG3KjtrbKgYGBjg6Oqrnb2QH1WevWmQI/loD4sKFCxQrVozRo0cTHx9PQkKCuub/pxhDRZXEt23b9q0kPigoiJCQEEaPHq2RHmyFQoG/vz+vXr3im2++oUWLFgQGBrJs2TI6dOjAo0ePOHr0KPXr18/0eh3aiKFpxsbGNGjQgJSUFCpVqpQtCXZKSgru7u507NiR9PR0JkyYwIYNG1iwYAEXLlzQ+podn11Pdnp6Onv27OHUqVOMHj1aXUM3NTVVvQRqdHS0RoZZqJazTklJwdramrp169K7d2/S09MZOXIk169fp2HDhjm+yMrXLCoqijx58qBQKNi3bx/Xrl0jOTmZESNG8PTpU+CvoTianGkfGxuLjo4OuXLl4tChQ4SFhXHr1i0GDBjA8+fPyZcvH6VKlXpnUnD48GGKFSvG8ePH1YnQ8+fP6datG7du3aJIkSK0adMmU8NFfHx8OHLkCIMHD1b3nAUGBrJx40bGjRuXLQvuvEtYWJh6YRInJyeWLFnC48ePGT58OAULFmTHjh3069cPExMTduzYQZ06dbRWRvBLkrF8qZubm3rFMnt7ezw8PHj58iUWFhbUr18/0+eq5ORkBg4cSM+ePdU1qvv06cORI0c4fvw4o0ePztJnl5qayoMHD9DR0SEkJIQrV65QunRpHj9+zPnz55k9ezalS5fO9P6/ZNk1f0G135iYGE6dOkXZsmVZt24dlStXJj4+nj///JPJkyeTN2/eTzrGu5L4UaNG0alTJ2rUqKGxJF71eq5evcrKlSspUaIENjY2dOnShcjISH7//XcqVqyIn58fbdu2zVTFMG3EyG6aPmZV+4uNjeXJkyds2rSJatWqERQUhLOzM6GhoRQqVChHevM/u55sHR0dWrRogYmJCZs2beLHH3/ku+++486dO1y4cIG2bdt+9Ipl/+bNmzccP36coKAg9RXX1KlTmT17Nhs2bGD58uXMmTPnkx8b+CXbv38/ly5dwtTUlJSUFOrVq4ehoSF6enq8evWKtWvXMmXKFI0m2B4eHly/fp2IiAjq1KnD8+fPiYqKolKlSrx8+ZJdu3YxevTotxLshw8fEh0dre5Ftre358aNG9jY2PD999+zc+dO+vTpQ4kSJciXLx8lS5bM9HjsChUqcO3aNY4dO0bJkiVJSEhgz549jBgxQmsJtupHzcjIiGPHjpGYmIiTkxP16tUjNTVVfVu2ZcuW6rGVqlJK4uOpEuwtW7ZgZ2dHxYoVWblyJQMHDqRly5bs27eP2rVrZyrBvnDhAi9evMDc3JzatWtTq1YtDhw4QKVKlQgLC6NZs2bUqVMn00Mhrly5wsWLF3n+/DmFChVixIgRpKWlkTt3bsqWLUtoaCgLFy7U6AJcX5rsSLAzrhC7YsUKnj17xvXr15k/fz5XrlwhICCABw8eZGlokDZiqCbqqpL42rVrs27dOnbt2kX58uX5/fffKV26NHnz5tXIxDuFQsGlS5fYvHkzS5Ys4cyZM1y7do2IiAjy58+Pvr4+O3fupHPnzplOfrURI7tp+pi9ffs2uXPnZvny5QwdOpQWLVowadIkBg4ciKGhIfv372fUqFEajfmhPruebJWEhAROnDjB7du3KV68OFeuXKFDhw7UqFEjU/vLOJ5RdQD4+Phw9OhRgoODKVKkCLlz58bJyYmLFy/SqlWrT6Im69fqjz/+4Ny5c4wcOZKUlBT8/Pw4dOgQPXr04MmTJ9y/f59evXpp9CRz6NAhTpw4wdSpU3n8+DHh4eE8fPhQ3cvi5+fHwIED34qZnp7OoUOHuHjxIi4uLlSpUkU9A/zw4cPqnl1NTnB99eoVFy9exMfHh9y5c9OxY0etXcGrvj9paWn4+PhgaWnJ7t27sbW1xcnJiaioKK5fv65OBoVm7Nmzh8WLF7Nhwwbs7Ow4c+YMGzZsYMiQIXz33XeZ+lHbv38/hw8fpmzZsiQkJODj44NCoWD79u0EBQXxxx9/4OrqmunkxMfHh+XLl9OpUyeePn3Ks2fPmDhxInFxcWzcuJFr164xbNiwTJ/TRda8efOGOXPm0K1bN8qXL8+wYcMwNzdXry+gGor0qcb4tyTe0tKSKVOmqJP4P/74g+3bt2ts8ZPg4GDGjh3LgAEDSE9PZ/HixZQuXZrU1FSqV69O7ty5qVOnTpaGLmojxufE09OTP/74gxYtWjB//nz27t1LwYIFOXnyJIcPH8bc3Jxu3bqpC2BonfIzFh8fr/Tw8FA2atRIeeHCBY3s8+7du2/9/cSJE8qQkBDlwIEDlXZ2dsoRI0Yonz59qpFYInN8fHyU7dq1U964cUOpVCqVKSkpSqVSqZw7d67684uLi9NozPPnzyudnZ2VkZGR6m1Pnz5VDh06VPn48WOlUqlUJiUl/etz4+LilIcOHVJOnTpV6evrq1QqlcobN24oFy9e/I/jTZNSU1OVqamp2bb/v0tLS1P/ef369cqff/5ZuWbNGuW1a9eUEyZMUM6bN085ceJEZXJystba9CVTvd+enp7KkydPKk+dOqVs1aqVMigoSKlUKpUnT55UBgQEZGrfu3btUk6dOlWZkJCg3Llzp9Ld3V35+PFjpbOzs3Lnzp3Kvn37qo/7zLh69aqySZMmyoEDBypfvnypfPDggXL16tXKpUuXKuPj45VpaWnKly9fZnr/InMyfofj4uKUbdq0Ubq5uam39e3bVzl+/Ph/PPZTi6Hy+vVr5ZgxY5SBgYFKpVKpHDp0qHLSpEnqf894Ps+KjO1cunSpcvz48cpevXopo6KilKmpqUp/f3/lypUrs3S+10aMz82uXbuUo0ePVr548UKpVCqV06dPV/bo0UP56tUrpVKpVEZFRSnj4+NzsonK7FkSSkuMjY1p0aIF7u7uGllG9s6dO7Rs2ZJJkyaxc+dOlEolt2/f5uHDh8ycOZNWrVoxYsSIT/YWzNcgPT0dS0tLatasSWhoKCEhIerbiUlJSbx58wZAY+WXZs6cya+//krlypXJkycPFy5cUP9b0aJFKViwoDqmvr7+v+4jV65cNGjQgO+++46TJ0+yY8cO1qxZQ506dbJ1nKmurq7GF935N8r/vxmWcfluc3NzNmzYQFxcHNevX1cv5a0qpSiyTvV+29rasmHDBipUqEDfvn3p378/QUFBNGzYMFN3Cx4+fMjatWtxcnLi5MmTeHl5cfHiRU6cOEGFChUwMDBg9uzZmR6DHR8fz4EDB3Bzc6N+/fqsW7eO5ORkddWbFStWkJ6eLncKtSzjBFd3d3eOHz/Ohg0b8PX1ZeXKlQCsXr2aIUOGAGRqGJ62Yqjo6+tz//59Tpw4AYCbmxuvXr1iwoQJABoZ7qn8/yEpqampXLp0id69e2Nra6vuOU1NTeXbb7+ld+/emT7fayPG50a1KFPDhg0pUKCAujRgiRIl6NatG6GhoeTJk0eri639m892uEh2SElJYf/+/QQFBfHs2TOaNm1KTEwMf/75J9OnT39rRSWhXenp6bx584YxY8aok9Pz589ja2tL8+bN8ff3Z/369cycOVOjP85paWl07doVFxcXHBwccHV1xcHBARcXF06dOoW7uzsLFiz4oJgJCQkcPnyYNWvWMHnyZOrUqaOxdua0+/fvc+jQIYYNG8aYMWMoUqQI/fr1Q0dHh0WLFmFiYqLxBQe+Vr6+vjx48ICLFy/Stm1b7OzsePHiBU+ePMHZ2Znt27dTunTpTJcvTUpK4siRIxw8eBClUsn69etJS0vj+PHjhISE0LBhwyyPkU5ISFD/+G3cuJHQ0FBatGiBoaEhhoaG0pGRQxITE3F1daV58+YcOXKEcuXK0aVLF3r06EGzZs3o3bv3Jx0jYxJ/4MAB8ubNyw8//MCgQYOoXbs2v/zyC/DXkDpN/E5kLG25Zs0abt++Tb58+WjZsiVnz54lOjqaLl26ZKkcrzZifI5U56nHjx9ToEABvL296datGzVr1mTbtm3Ur19fK3Ww30eS7L+Jj4/nypUr3L59m1KlSvHNN9+wfv16JkyYoJEi++LjKTOMkx87diyPHz9mxYoVREVF4enpSUxMDE+fPmXSpEkaq06hzDC2eMmSJezbt49u3brRsWNHRo0aRf78+YmLi2P48OEfFTMpKYnIyMgvpiLNvn37uH//PsOHD6d79+40a9YMFxcXdf1TFxcXDAwMePLkyVf3I5AdfH192bx5M02aNCE0NBQDAwNOnjxJ4cKFSUpKYu7cuRqZVJSUlMTBgwcJCAigU6dO2bLM/d+Th9DQUIYNGybn2Rxw8eJFEhISSE9PJzAwkOHDhwMwZMgQGjZsiIODA2/evMlS0qKNGKCdCwV4ex6Xm5sbxYoVo23btuzYsYOgoCAGDhzIzp07cXR0zPScGG3E+JwlJCS8tSjTTz/9lNNN+ofPerhIdsiVKxfVqlXD1tYWX19fzM3NmTZtmpz4c0h6ejoKhULdM9GuXTt69OjBpEmTyJ8/PzVr1sTMzEyjCTb8b2W7yZMn06BBAw4cOMDRo0fZvXs3CxcuJDk5mWHDhn10TENDwy8mwQaoVasWUVFR7Ny5k82bN3PmzBm2b9/O6NGjuXv3Llu3blWvliqy5sGDB6xbt44ZM2bQoEEDateuTWRkJBUrViR37tyEhoayfPlyjcQyNDSkWbNm1KhRg/Xr1xMQEKCR/WakWgUUoG/fvvz0009yns0Bnp6erF27lpiYGJKTk/Hx8WH//v0AtG3blsjISMzMzLKU/GojxsWLFzl58iTe3t7Y2trSrFkzli5dyu3btzl16hSbNm2iSZMmmd5/Rhk7fs6ePcvNmze5efMmAJ06dSIlJYXjx4/Tr1+/LCfY2Rnjc6daWXXgwIE8fPgwW85TWfXZlfDTBhMTE/UYbz09vc9yEYQvhY6ODlFRUUyZMoXOnTtjYWFBmTJl0NfXp3v37tSvX5/evXtrrJSiqnctLS1NHTs+Pp7cuXPj7u7O999/T8GCBZk/f/5XvfS3auhUgQIFGD58OMuWLWPbtm2sXLmSQYMGkZyczJQpU3j69OlX/T5pkoGBAVZWVhw8eFBd91VVQaBt27aEhYVx6dIljcUzNjbmhx9+QFdXN9suDDNWdJIhItp3/vx5Dhw4wNy5c9XDJ0qWLMnQoUMJDQ3l2rVrjB8//pOP4enpyb59+2jVqhX6+vr4+PhgY2ODs7Mzbdu25eHDh5iZmWmkisiFCxfUK/XOnj2bunXrsmHDBlxdXZk1axZNmjQhPT09S2OjtRHjS6GNRZmyQnqy38HExARHR0fy5cuX0035Kql6uJKSkjh+/Dg1atTAxsYGNzc3xo4di5mZGWPGjKFt27YaT7CjoqJYvnw5Pj4+uLq64u7ujo+PDy9fvqRt27aUL1/+q04c09PT0dXVJT09nSNHjhASEkL//v158uQJv//+O8uWLePChQuEhYVRqlSpnG7uF0O1INb9+/dp3bo1Li4ufP/995w7d460tDRCQ0Px9vYmLi4OTY0CzJUrF40aNdL4kuAZfc3fpZwWERFBgwYNsLa2JiUlhfT0dPT09KhcuTLlypVj+vTpWb74ye4YGZN4Z2dnmjdvzvTp01m5ciVr1qzh999/x8nJKUuvQSU+Pp4HDx6o/3z9+nX132fOnMmbN29Yu3Yt3bt3p3Llypn6HmojxpdGtbJqdp6nMkvGZItP1ps3b/D29iYmJgYfHx+srKzo3r07fn5+pKWl0a5du2yJuXv3bvLmzcv27dtp06YN9vb2zJs3D3Nzc1xdXaXHjb8ugmbNmoWOjg6GhoZYW1vj6OjIsmXLsLGxoWfPnpI8ZaOEhAQuX76Mp6cnHTt2pHr16jx48AAzMzOpyiE+2KVLlzhw4ABDhgxRHzdeXl7cu3ePAQMGaGSif3bH+OOPP4iLi6Nz586kpKSgq6tLUFAQa9asoWXLltjZ2Wkk+VItyQ6wYMECLCwscHZ2ZuTIkTRv3pw2bdqQlJTE5MmTsbCwyNTiJ9qIIbRLerLFJ0F1rbdo0SL1ErcjR44kKiqKTp06MXfuXLp27crz5885deoUtWrV0njs1NRUxo4dS0xMDG3btmX58uUcOHCAGzdusGHDBmbMmPFVJ9gZr8cPHz6MmZkZo0aN4vXr19y7d48tW7bQt29ffvjhB0mws1FsbCxnzpzh4MGDdOjQgerVqwN/lfKTBFt8jAoVKlCkSBFOnDiBj48PJ06cYN++fTRr1kxjlbSyO4a1tTW3bt3i1atX6Ovro6Ojw8OHDylatCi1atXSWO+mgYEBaWlpPHr0iF9++QVvb29OnDjBggUL8PDwYNeuXRgaGjJt2jSioqKIjIz8JGMI7ZKebPFJ2bZtGy1atMDAwIC9e/dy4cIFZsyYQf78+blz5w579uyhW7duFCtWTCPxVENEkpKS0NPT49atW4wePZqJEydSt25dgoODGT9+PEuXLsXc3FwjMT9HqvdJqVTy6NEjihYtysWLFzl//jyNGzcmMTGR9evXM3z4cCpVqpTTzf0qZCyDJ0RmvXr1igsXLnD+/Hly585N165dNT6RLjtjxMbGsmXLFszNzSlRogRxcXF4eHgwatQojcRYvnw5ZmZmdO/enalTp3Lv3j2qVKnCsGHD6N+/Pw4ODjRt2pTnz59n+tynjRgiZ0iSLT4ZBw4cYP/+/dSqVYt9+/bh4eGBh4cHfn5+TJw4EQsLC5KSkjA0NNRIPNWEK1VVhlevXtGuXTtMTExYsGABAwcOxMHBQeqj/z+lUsmUKVOoVKkS9erVw9LSEnd3d3R0dAgPD8fFxeWr7ukX4nOWmpoKoF7c63OKkZ1J/O3btxk7diz29vbUqFGDRo0a0bdvXypXrkyfPn3o2bMnCxYsyNKkO23EEDlDkmzxSdi9eze+vr6MGDECa2trZs6cyfPnz3Fzc2P79u3cvn2befPmoVAoNDoUITk5mTlz5tCoUSNKlCjBxIkTcXFxwcrKitmzZ7NhwwaMjIwytfLYlyBjLeP169djbm5Oy5Yt+f3334mJiaFcuXIcP36cH3/8USOrrgohRGZpKolXdcCcPXuWGzduEBwcTEBAAC1atGDgwIEAdOvWjapVqzJs2LBM/SZpI4bIeVLCT+Q41fKoY8aMwdramoCAAEJCQrC2tqZDhw7s3LmT5ORkjSW6Gcv06erqEh4ejpWVlTq5nzJlCqtWrWLjxo1f/e141RARhULBmzdvePXqFaNHj6Zp06Y8fvwYfX19pk+fLkulCyFynKZ6yBUKBT4+Pmzfvp3+/ftja2tLamoqffv2pWDBgrRp04YtW7YQEBCQ6eRXGzFEzpMkW+S4IkWKMGTIEG7fvk1UVBTe3t50796dmjVrsmXLFiIjIylcuLBGYmVcdnfVqlV06tSJhg0bMmfOHMaPH09SUpJ6NUdVDeKvkeoGl0KhYPny5QQFBbF48WKePXvG69evKV68OF5eXhQuXFgSbCHEFyUhIQFPT09GjBiBnZ0dSUlJmJmZUbduXebNm0dCQgJdunTJ0vhobcQQOe/rvAcuPimGhoY0btwYGxsbNm7cSPXq1alZsybw1+0yTSXY8FfP7Js3b5g8eTKlSpUiISGB1q1b06RJExYsWMCOHTvo3bv3V51gA28Nyxk0aBAA48ePp2jRosTFxamH1ZQoUSInmymEEBqnunOXmJgI/K+HvHDhwkyYMIEyZcp8FjFEzpMx2eKTkZyczNGjR/H19aVdu3YavYLPOLY4PDycdevWMWLECAwMDLh58yb6+vrY2tqSnJxMrly5NBb3c5NxKd8ZM2ZgYGDAmDFjABgyZAhpaWksX76c4OBgjV78CCHEp2T37t3o6OhQu3ZtChYsSGBgIDNnzmTevHkUL178s4khcpYk2eKTEh8fz7lz5/juu+80Vt8040qOoaGhpKSkcOTIEezt7albty5eXl5cuXKFiRMnqhcC+BplvBBJTk4mOTmZoUOH8u233zJo0CBCQkL45ZdfcHNzw8bGJodbK4QQ2efly5fs27ePwMBAKlSogLe3N4MHD9boGg3aiCFyliTZ4pOTsTdVU8LDw1myZAmWlpYYGBiQkJBAYmKiuu7zpEmTpPwckJaWxsSJE9HR0aFw4cL07NmT3r17U6FCBVJTU+nQoQOlS5fO6WYKIUS2i4+P5+7du8TExGBhYUG5cuU+yxgi50iSLb5od+7coVSpUsybN49vv/2W0qVL89tvvzFy5Mi36m1/rSvlqS5oEhIS0NPTY9u2bZiYmNC0aVM6d+5M+/btady4MUeOHKFGjRoyTlAIIYT4QJJkiy/WoUOH2LNnD/379yckJISXL1/y+PFj+vTpw7Nnz0hPT6dhw4Y53cwco0qwL168yNGjRzExMcHS0pKqVaty6tQpKlWqxIYNGyhevDgjRozQ2PAdIYQQ4msgJfzEF+nw4cNcuXKFhQsXYm5uzvPnzzlw4ABOTk68evUKT09PhgwZktPNzFGqOq2bNm1iwIABFChQAGtra06cOIG1tTWOjo4EBQVRo0YNSbCFEEKIjyRJtvjiJCcn4+3tTcuWLXn58iUnTpzg8uXL6Orq8uzZMxISEhg6dOhXP3v773Vak5OTAQgMDGT37t2cOHGC4cOHS51WIYQQIhMkyRZfHAMDA+rUqcPmzZuxtLSka9eutGnThr1795I7d24aN26c0038JPy9Tquuri4AlpaWTJw4kRIlSmBnZ5eTTRRCCCE+W5Jkiy9S06ZNqVWrFvny5UOhUHDlyhV8fHwYOnRoTjftk2FkZESDBg148OABVlZWFCxYkFu3bnHw4EFmz54tZfqEEEKILJCJj+KLdv36dc6dO0dgYCATJ06kWLFiOd2kT4rUaRVCCCGyhyTZ4osWFxdHSEgIpqamFCxYMKeb80mSOq1CCCGE5kmSLYQQQgghhIbp5HQDhBBCCCGE+NJIki2EEEIIIYSGSZIthBBCCCGEhkmSLYQQQgghhIZJki2EEEIIIYSGSZIthBBCCCGEhkmSLYQQQgghhIb9HwlmaGYrVt7FAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(12, 4))\n",
    "sns.barplot(x='Method', y='Time', hue='Backend', data=df[df['Forw']==True], palette=\"bright\")\n",
    "ax.set_xlabel('');\n",
    "ax.set_ylabel('Time (ms)')\n",
    "ax.set_title('Forward')\n",
    "ax.set_xticklabels(df['Method'].unique(),rotation=45, fontsize=9)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12, 4))\n",
    "sns.barplot(x='Method', y='Time', hue='Backend', data=df[df['Forw']==False], palette=\"bright\")\n",
    "ax.set_xlabel('');\n",
    "ax.set_ylabel('Time (ms)')\n",
    "ax.set_title('Adjoint')\n",
    "ax.set_xticklabels(df['Method'].unique(),rotation=45, fontsize=9);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_np = df[df['Backend'] == 'np']\n",
    "df_cp = df[df['Backend'] == 'cp']\n",
    "\n",
    "df_speedup = df_np.groupby(['Method', 'Forw']).mean() / df_cp.groupby(['Method', 'Forw']).mean()\n",
    "df_speedup = pd.DataFrame(df_speedup.to_records())\n",
    "df_speedup['Forw'] = df_speedup['Forw'].map({True: 'Forw', False: 'Adj'})\n",
    "df_speedup = df_speedup.rename(columns={'Time': 'Speedup', 'Forw':'Kind'})\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(18, 4))\n",
    "g = sns.barplot(x='Method', y='Speedup', hue='Kind', data=df_speedup, palette=\"bright\")\n",
    "g.set_yscale(\"log\")\n",
    "ax.set_xlabel('')\n",
    "ax.set_title('GPU Speed-up')\n",
    "ax.set_xticklabels(df_speedup['Method'].unique(),rotation=45, fontsize=9);\n",
    "\n",
    "plt.savefig('Speedup.eps', dpi=300, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "32LnopoUifer",
    "outputId": "9d6c5ac5-a87a-4a26-fa18-1b170fd91674"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "               Method  Kind     Speedup\n",
      "0            Bilinear   Adj   87.320490\n",
      "1            Bilinear  Forw   10.265064\n",
      "2   CausalIntegration   Adj   56.626825\n",
      "3   CausalIntegration  Forw   52.327019\n",
      "4        ChirpRadon2D   Adj   23.105806\n",
      "5        ChirpRadon2D  Forw   40.702982\n",
      "6        ChirpRadon3D   Adj   14.075977\n",
      "7        ChirpRadon3D  Forw   19.558677\n",
      "8          Convolve1D   Adj  300.081329\n",
      "9          Convolve1D  Forw  255.197904\n",
      "10         Convolve2D   Adj    6.790194\n",
      "11         Convolve2D  Forw   59.052445\n",
      "12         ConvolveND   Adj   16.766678\n",
      "13         ConvolveND  Forw   82.706551\n",
      "14           Diagonal   Adj  178.529788\n",
      "15           Diagonal  Forw  196.659363\n",
      "16                FFT   Adj  333.504175\n",
      "17                FFT  Forw  434.478695\n",
      "18              FFT2D   Adj   73.129410\n",
      "19              FFT2D  Forw  215.286123\n",
      "20              FFTND   Adj   54.184452\n",
      "21              FFTND  Forw  151.840721\n",
      "22    FirstDerivative   Adj   61.289490\n",
      "23    FirstDerivative  Forw   62.893917\n",
      "24               Flip   Adj   36.948576\n",
      "25               Flip  Forw   37.205458\n",
      "26          Fredholm1   Adj   10.324390\n",
      "27          Fredholm1  Forw   17.223799\n",
      "28           Gradient   Adj   16.123690\n",
      "29           Gradient  Forw   31.698659\n",
      "30           Identity   Adj    0.621425\n",
      "31           Identity  Forw   65.482733\n",
      "32             Interp   Adj    1.681032\n",
      "33             Interp  Forw    3.787271\n",
      "34          Laplacian   Adj   15.967672\n",
      "35          Laplacian  Forw   31.619879\n",
      "36         MatrixMult   Adj   34.752847\n",
      "37         MatrixMult  Forw    5.445526\n",
      "38                Pad   Adj   51.381740\n",
      "39                Pad  Forw   28.262812\n",
      "40        Restriction   Adj    1.395795\n",
      "41        Restriction  Forw    1.368300\n",
      "42               Roll   Adj   21.189911\n",
      "43               Roll  Forw   20.472498\n",
      "44   SecondDerivative   Adj   30.510026\n",
      "45   SecondDerivative  Forw   53.998740\n",
      "46                Sum   Adj   27.204947\n",
      "47                Sum  Forw   14.580669\n",
      "48         Symmetrize   Adj   78.697778\n",
      "49         Symmetrize  Forw   92.585928\n",
      "50          Transpose   Adj  112.991058\n",
      "51          Transpose  Forw   87.142365\n",
      "52               Zero   Adj    0.947535\n",
      "53               Zero  Forw    0.962258\n"
     ]
    }
   ],
   "source": [
    "print(df_speedup)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "OXC5MSx6_8Qj",
    "outputId": "414b5d04-919f-4508-ecbc-367620ab528b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average speed-up operators 68.8694332140358\n"
     ]
    }
   ],
   "source": [
    "print('Average speed-up operators', np.mean(df_speedup['Speedup']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QOjsRcOZVWdv"
   },
   "source": [
    "# Composite operators"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ypKYKn2xVf-j"
   },
   "source": [
    "## PoststackLinearModelling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "id": "TCB_3cKoVe6z"
   },
   "outputs": [],
   "source": [
    "df_comp = pd.DataFrame(columns=['Method', 'Forw', 'Backend', 'Time'])\n",
    "\n",
    "number, repeat = 10, 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "id": "De3htGLkVeXu"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "wav = (pylops.utils.wavelets.ricker(np.arange(51)*0.004, 20)[0])\n",
    "wavc = cp.asarray(wav)\n",
    "nt, nx = 1000, 1000\n",
    "x = np.ones(nt*nx)\n",
    "xc = cp.ones(nt*nx)\n",
    "y = np.ones(nt*nx)\n",
    "yc = cp.ones(nt*nx)\n",
    "Op = pylops.avo.poststack.PoststackLinearModelling(wav, nt, nx)\n",
    "cOp = pylops.avo.poststack.PoststackLinearModelling(wavc, nt, nx)\n",
    "\"\"\"\n",
    "df_comp = pd.DataFrame(columns=['Method', 'Forw', 'Backend', 'Time'])\n",
    "df_comp = timing(cmd_setup, df_comp, 'PoststackLinearModelling_2D', number=number, repeat=repeat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "id": "1IQDRKt7W5Jm"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "wav = (pylops.utils.wavelets.ricker(np.arange(51)*0.004, 20)[0])\n",
    "wavc = cp.asarray(wav)\n",
    "nt, nx, ny = 1000, 10, 100\n",
    "x = np.ones(nt*nx*ny)\n",
    "xc = cp.ones(nt*nx*ny)\n",
    "y = np.ones(nt*nx*ny)\n",
    "yc = cp.ones(nt*nx*ny)\n",
    "Op = pylops.avo.poststack.PoststackLinearModelling(wav, nt, (nx, ny))\n",
    "cOp = pylops.avo.poststack.PoststackLinearModelling(wavc, nt, (nx, ny))\n",
    "\"\"\"\n",
    "df_comp = timing(cmd_setup, df_comp, 'PoststackLinearModelling_3D', number=number, repeat=repeat)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "-CriBZEOrjgu"
   },
   "source": [
    "## PrestackLinearModelling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "id": "kdZOonVurkUG"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "wav = (pylops.utils.wavelets.ricker(np.arange(51)*0.004, 20)[0])\n",
    "wavc = cp.asarray(wav)\n",
    "theta = np.arange(0, 30, 5.)\n",
    "thetac = cp.arange(0, 30, 5.)\n",
    "ntheta = len(theta)\n",
    "nt, nx = 1000, 1000\n",
    "x = np.ones(nt*nx*3)\n",
    "xc = cp.ones(nt*nx*3)\n",
    "y = np.ones(nt*nx*ntheta)\n",
    "yc = cp.ones(nt*nx*ntheta)\n",
    "Op = pylops.avo.prestack.PrestackLinearModelling(wav, theta, 0.5, nt, nx)\n",
    "cOp = pylops.avo.prestack.PrestackLinearModelling(wavc, thetac, 0.5, nt, nx)\n",
    "\"\"\"\n",
    "df_comp = timing(cmd_setup, df_comp, 'PrestackLinearModelling_2D', number=number, repeat=repeat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "id": "7s-ZX1HFqUIh"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "wav = (pylops.utils.wavelets.ricker(np.arange(51)*0.004, 20)[0])\n",
    "wavc = cp.asarray(wav)\n",
    "theta = np.arange(0, 30, 5.)\n",
    "thetac = cp.arange(0, 30, 5.)\n",
    "ntheta = len(theta)\n",
    "nt, nx, ny = 1000, 10, 100\n",
    "x = np.ones(nt*nx*ny*3)\n",
    "xc = cp.ones(nt*nx*ny*3)\n",
    "y = np.ones(nt*nx*ny*ntheta)\n",
    "yc = cp.ones(nt*nx*ny*ntheta)\n",
    "Op = pylops.avo.prestack.PrestackLinearModelling(wav, theta, 0.5, nt, (nx, ny))\n",
    "cOp = pylops.avo.prestack.PrestackLinearModelling(wavc, thetac, 0.5, nt, (nx, ny))\n",
    "\"\"\"\n",
    "df_comp = timing(cmd_setup, df_comp, 'PrestackLinearModelling_3D', number=number, repeat=repeat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 699
    },
    "id": "bsLfyVD3hSHL",
    "outputId": "c1e399c1-d702-44bc-ffd7-61900129ad3e"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(12, 4))\n",
    "sns.barplot(x='Method', y='Time', hue='Backend', data=df_comp[df_comp['Forw']==True], palette=\"dark\")\n",
    "ax.set_xlabel('');\n",
    "ax.set_ylabel('Time (ms)')\n",
    "ax.set_title('Forward')\n",
    "ax.set_xticklabels(df_comp['Method'].unique(),rotation=45, fontsize=9)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12, 4))\n",
    "sns.barplot(x='Method', y='Time', hue='Backend', data=df_comp[df_comp['Forw']==False], palette=\"dark\")\n",
    "ax.set_xlabel('');\n",
    "ax.set_ylabel('Time (ms)')\n",
    "ax.set_title('Adjoint')\n",
    "ax.set_xticklabels(df_comp['Method'].unique(),rotation=45, fontsize=9);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 358
    },
    "id": "Y4p74WSshXBq",
    "outputId": "5e4a2600-fe5d-4822-e2b4-4436170c3854"
   },
   "outputs": [],
   "source": [
    "df_comp_np = df_comp[df_comp['Backend'] == 'np']\n",
    "df_comp_cp = df_comp[df_comp['Backend'] == 'cp']\n",
    "\n",
    "df_comp_speedup = df_comp_np.groupby(['Method', 'Forw']).mean() / df_comp_cp.groupby(['Method', 'Forw']).mean()\n",
    "df_comp_speedup = pd.DataFrame(df_comp_speedup.to_records())\n",
    "df_comp_speedup['Forw'] = df_comp_speedup['Forw'].map({True: 'Forw', False: 'Adj'})\n",
    "df_comp_speedup = df_comp_speedup.rename(columns={'Time': 'Speedup', 'Forw':'Kind'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 297
    },
    "id": "qyMJgyErhcx0",
    "outputId": "10a6c54b-4cb8-44c1-d7f4-fb07ec867c9d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                        Method  Kind    Speedup\n",
      "0  PoststackLinearModelling_2D   Adj   4.213610\n",
      "1  PoststackLinearModelling_2D  Forw  12.606533\n",
      "2  PoststackLinearModelling_3D   Adj   5.551755\n",
      "3  PoststackLinearModelling_3D  Forw   9.629610\n",
      "4   PrestackLinearModelling_2D   Adj  10.137013\n",
      "5   PrestackLinearModelling_2D  Forw  19.663094\n",
      "6   PrestackLinearModelling_3D   Adj   7.766390\n",
      "7   PrestackLinearModelling_3D  Forw  12.713874\n"
     ]
    }
   ],
   "source": [
    "print(df_comp_speedup)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Ji5cwmcOj55V",
    "outputId": "bd14065b-bc0b-4964-a298-7b8332afa016"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average speed-up Convmodelling 10.28523492566096\n"
     ]
    }
   ],
   "source": [
    "print('Average speed-up Convmodelling', np.mean(df_comp_speedup['Speedup']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HJR0G6lfpp_n"
   },
   "source": [
    "## PhaseShift"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "id": "jax4v2yuh1Ko"
   },
   "outputs": [],
   "source": [
    "df_comp = pd.DataFrame(columns=['Method', 'Forw', 'Backend', 'Time'])\n",
    "\n",
    "number, repeat = 10, 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "id": "HSM3rELdtOoQ"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "vel = 1500.\n",
    "zprop = 100\n",
    "pad = 11\n",
    "nt, dt = 1000, 0.004\n",
    "nx, dx = 1000, 10\n",
    "freq = np.fft.rfftfreq(nt, dt)\n",
    "kx = np.fft.fftshift(np.fft.fftfreq(nx + 2*pad, dx))\n",
    "x = np.ones(nt*(nx+2*pad))\n",
    "xc = cp.ones(nt*(nx+2*pad))\n",
    "y = np.ones(nt*(nx+2*pad))\n",
    "yc = cp.ones(nt*(nx+2*pad))\n",
    "Op = pylops.waveeqprocessing.PhaseShift(vel, zprop, nt, freq, kx)\n",
    "cOp = pylops.waveeqprocessing.PhaseShift(vel, zprop, nt, cp.asarray(freq), cp.asarray(kx))\n",
    "\"\"\"\n",
    "df_comp = timing(cmd_setup, df_comp, 'PhaseShift_2D', number=number, repeat=repeat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "id": "2kgyreMlpwC3"
   },
   "outputs": [],
   "source": [
    "# setup command\n",
    "cmd_setup =\"\"\"\\\n",
    "import numpy as np\n",
    "import cupy as cp\n",
    "import pylops\n",
    "vel = 1500.\n",
    "zprop = 100\n",
    "pad = 11\n",
    "nt, dt = 100, 0.004\n",
    "nx, dx = 100, 10\n",
    "ny, dy = 100, 10\n",
    "freq = np.fft.rfftfreq(nt, dt)\n",
    "kx = np.fft.fftshift(np.fft.fftfreq(nx + 2*pad, dx))\n",
    "ky = np.fft.fftshift(np.fft.fftfreq(ny + 2*pad, dy))\n",
    "x = np.ones(nt*(nx+2*pad)*(ny+2*pad))\n",
    "xc = cp.ones(nt*(nx+2*pad)*(ny+2*pad))\n",
    "y = np.ones(nt*(nx+2*pad)*(ny+2*pad))\n",
    "yc = cp.ones(nt*(nx+2*pad)*(ny+2*pad))\n",
    "Op = pylops.waveeqprocessing.PhaseShift(vel, zprop, nt, freq, kx, ky)\n",
    "cOp = pylops.waveeqprocessing.PhaseShift(vel, zprop, nt, cp.asarray(freq), cp.asarray(kx), cp.asarray(ky))\n",
    "\"\"\"\n",
    "df_comp = timing(cmd_setup, df_comp, 'PhaseShift_3D', number=number, repeat=repeat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "id": "_20nIvi50hiJ"
   },
   "outputs": [],
   "source": [
    "#df_comp['Time'] /= 1000 # ms -> s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 615
    },
    "id": "ToMVDR1IVj5y",
    "outputId": "b49dff0f-a230-40fa-f1de-3e69cb98b87b"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(12, 4))\n",
    "sns.barplot(x='Method', y='Time', hue='Backend', data=df_comp[df_comp['Forw']==True], palette=\"bright\")\n",
    "ax.set_xlabel('');\n",
    "ax.set_ylabel('Time (ms)')\n",
    "ax.set_title('Forward')\n",
    "ax.set_xticklabels(df_comp['Method'].unique(),rotation=45, fontsize=9)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12, 4))\n",
    "sns.barplot(x='Method', y='Time', hue='Backend', data=df_comp[df_comp['Forw']==False], palette=\"bright\")\n",
    "ax.set_xlabel('');\n",
    "ax.set_ylabel('Time (ms)')\n",
    "ax.set_title('Adjoint')\n",
    "ax.set_xticklabels(df_comp['Method'].unique(),rotation=45, fontsize=9);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 316
    },
    "id": "NnE1FGJlUAGU",
    "outputId": "fe59b034-150d-4425-a980-f3904d9b8633"
   },
   "outputs": [],
   "source": [
    "df_comp_np = df_comp[df_comp['Backend'] == 'np']\n",
    "df_comp_cp = df_comp[df_comp['Backend'] == 'cp']\n",
    "\n",
    "df_comp_speedup1 = df_comp_np.groupby(['Method', 'Forw']).mean() / df_comp_cp.groupby(['Method', 'Forw']).mean()\n",
    "df_comp_speedup1 = pd.DataFrame(df_comp_speedup1.to_records())\n",
    "df_comp_speedup1['Forw'] = df_comp_speedup1['Forw'].map({True: 'Forw', False: 'Adj'})\n",
    "df_comp_speedup1 = df_comp_speedup1.rename(columns={'Time': 'Speedup', 'Forw':'Kind'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 173
    },
    "id": "uNXV8VlWYvWY",
    "outputId": "d184d22c-f4b7-454d-938c-ba6eb425597c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          Method  Kind     Speedup\n",
      "0  PhaseShift_2D   Adj   77.415956\n",
      "1  PhaseShift_2D  Forw  103.284773\n",
      "2  PhaseShift_3D   Adj  169.239121\n",
      "3  PhaseShift_3D  Forw  203.486204\n"
     ]
    }
   ],
   "source": [
    "print(df_comp_speedup1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "lqyLABHlknAS",
    "outputId": "75beb40c-fb21-4099-b110-c4baff4f048f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average speed-up Phase shift 138.35651343699814\n"
     ]
    }
   ],
   "source": [
    "print('Average speed-up Phase shift', np.mean(df_comp_speedup1['Speedup']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(1, 2, figsize=(15, 4))\n",
    "sns.barplot(x='Method', y='Speedup', hue='Kind', data=df_comp_speedup, palette=\"bright\", ax=axs[0])\n",
    "axs[0].set_xlabel('')\n",
    "axs[0].set_title('Convolutional Modelling')\n",
    "axs[0].set_xticklabels(df_comp_speedup['Method'].unique(),rotation=45, fontsize=9);\n",
    "sns.barplot(x='Method', y='Speedup', hue='Kind', data=df_comp_speedup1, palette=\"bright\", ax=axs[1])\n",
    "axs[1].set_xlabel('')\n",
    "axs[1].set_title('Phase-Shift Modelling')\n",
    "axs[1].set_xticklabels(df_comp_speedup1['Method'].unique(),rotation=45, fontsize=9);\n",
    "\n",
    "plt.savefig('Speedup_composite.eps', dpi=300, bbox_inches='tight')"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Cupy_PYLOPS_timing.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}