https://www.reddit.com/r/learnpython/comments/i1cctb/i_need_help_in_optimisating_appending_to_numpy/

On-Off Keying.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import numba as nb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def setup():\n",
    "    bin_data = np.random.randint(0, 2, 10**3, dtype=np.uint8)\n",
    "    sign_length = 1000\n",
    "    return bin_data, sign_length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# https://www.reddit.com/r/learnpython/comments/i1cctb/i_need_help_in_optimisating_appending_to_numpy/fzwghgf/\n",
    "def func_b(bin_data, sign_length):\n",
    "    bin_data = bin_data.reshape(len(bin_data), 1)\n",
    "    sign_length = np.ones(sign_length, dtype=int).reshape(1, sign_length)\n",
    "    signal = (bin_data*sign_length).ravel()\n",
    "    return signal"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "@nb.njit\n",
    "def fast_repeat(bin_data, sign_length):\n",
    "    n = len(bin_data)\n",
    "    out = np.empty(n * sign_length, dtype=np.uint8)\n",
    "    for i in range(n):\n",
    "        row = i*sign_length\n",
    "        out[row:row+sign_length] = bin_data[i]\n",
    "    return out"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Establish cost of generating data (used for calculating bandwidth)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8.59 µs ± 14.1 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%%timeit\n",
    "bin_data, sign_length = setup()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Original (~1.45 MB/s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "684 ms ± 31.6 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "%%timeit\n",
    "bin_data, sign_length = setup()\n",
    "\n",
    "signal = np.array([])\n",
    "for bit in bin_data:\n",
    "    temp = np.full(sign_length,int(bit))\n",
    "    signal = np.append(signal,temp)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### np.repeat (~400 MB/s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.63 ms ± 61.3 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
     ]
    }
   ],
   "source": [
    "%%timeit\n",
    "bin_data, sign_length = setup()\n",
    "signal = bin_data.repeat(sign_length)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### numpy broadcast (~625 MB/s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.61 ms ± 8.64 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%%timeit\n",
    "bin_data, sign_length = setup()\n",
    "func_b(bin_data, sign_length) # primitive_screwhead"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Numba compiled (~60 GB/s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "25.1 µs ± 77.7 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%%timeit\n",
    "bin_data, sign_length = setup()\n",
    "fast_repeat(bin_data, sign_length) # numba"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Confirm that the output is correct"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(True, True, True)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bin_data, sign_length = setup()\n",
    "a = bin_data.repeat(sign_length)\n",
    "b = func_b(bin_data, sign_length)\n",
    "c = fast_repeat(bin_data, sign_length)\n",
    "\n",
    "(a == b).all(), (a == c).all(), (b == c).all()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.8.0 64-bit",
   "language": "python",
   "name": "python38064bitce923c4d35cd4821adec5b4dd96a59a2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}