Skip to content

Instantly share code, notes, and snippets.

@genkuroki

genkuroki/sum of sin(x)／n.ipynb

Last active May 21, 2021 03:17
Show Gist options
  • Save genkuroki/e73671622c6ee038286ce0fdb18a7407 to your computer and use it in GitHub Desktop.
Save genkuroki/e73671622c6ee038286ce0fdb18a7407 to your computer and use it in GitHub Desktop.
Download ZIP
sum of sin(x)／n
Display the source blob
Display the rendered blob
Raw
sum of sin(x)／n.ipynb
{
"cells": [
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "versioninfo()",
"execution_count": 1,
"outputs": [
{
"output_type": "stream",
"text": "Julia Version 1.7.0-DEV.1112\nCommit 8a0a9a7388 (2021-05-18 18:42 UTC)\nPlatform Info:\n OS: Windows (x86_64-w64-mingw32)\n CPU: Intel(R) Core(TM) i7-10750H CPU @ 2.60GHz\n WORD_SIZE: 64\n LIBM: libopenlibm\n LLVM: libLLVM-11.0.1 (ORCJIT, skylake)\nEnvironment:\n JULIA_NUM_THREADS = 12\n JULIA_PYTHONCALL_EXE = C:\\Users\\genkuroki\\.julia\\conda\\3\\python.exe\n",
"name": "stdout"
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "using LoopVectorization, SLEEFPirates, BenchmarkTools",
"execution_count": 2,
"outputs": []
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "sin_ccall(x::Float64) = ccall(:sin, Float64, (Float64,), x)\n\nfunction f_ccall(N)\n y = 0.0\n for n in 1:N\n y += sin_ccall(Float64(n))/n\n end\n 2y + 1\nend\n\n@btime f_ccall(10^6)",
"execution_count": 3,
"outputs": [
{
"output_type": "stream",
"text": " 31.395 ms (0 allocations: 0 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 3,
"data": {
"text/plain": "3.141590588883843"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "function f_naive(N)\n y = 0.0\n for n in 1:N\n y += sin(n)/n\n end\n 2y + 1\nend\n\n@btime f_naive(10^6)",
"execution_count": 4,
"outputs": [
{
"output_type": "stream",
"text": " 11.739 ms (0 allocations: 0 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 4,
"data": {
"text/plain": "3.141590588883843"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "function f_simd(N)\n y = 0.0\n @simd for n in 1:N\n y += sin(n)/n\n end\n 2y + 1\nend\n\n@btime f_simd(10^6)",
"execution_count": 5,
"outputs": [
{
"output_type": "stream",
"text": " 11.932 ms (0 allocations: 0 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 5,
"data": {
"text/plain": "3.141590588883843"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "function f_sleef(N)\n y = 0.0\n for n in 1:N\n y += SLEEFPirates.sin_fast(float(n))/n\n end\n 2y + 1\nend\n\n@btime f_sleef(10^6)",
"execution_count": 6,
"outputs": [
{
"output_type": "stream",
"text": " 6.512 ms (0 allocations: 0 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 6,
"data": {
"text/plain": "3.1415905888838433"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "function f_simd_sleef(N)\n y = 0.0\n @simd for n in 1:N\n y += SLEEFPirates.sin_fast(float(n))/n\n end\n 2y + 1\nend\n\n@btime f_simd_sleef(10^6)",
"execution_count": 7,
"outputs": [
{
"output_type": "stream",
"text": " 3.026 ms (0 allocations: 0 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 7,
"data": {
"text/plain": "3.141590588883785"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "function f_avx(N)\n y = 0.0\n @avx for n in 1:N\n y += sin(n)/n\n end\n 2y + 1\nend\n\n@btime f_avx(10^6)",
"execution_count": 8,
"outputs": [
{
"output_type": "stream",
"text": " 1.609 ms (0 allocations: 0 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 8,
"data": {
"text/plain": "3.1415905888837843"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "function f_avx_sleef(N)\n y = 0.0\n @avx for n in 1:N\n y += SLEEFPirates.sin_fast(float(n))/n\n end\n 2y + 1\nend\n\n@btime f_avx_sleef(10^6)",
"execution_count": 9,
"outputs": [
{
"output_type": "stream",
"text": " 2.347 ms (0 allocations: 0 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 9,
"data": {
"text/plain": "3.141590588883785"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "function f_avxt(N)\n y = 0.0\n @avxt for n in 1:N\n y += sin(n)/n\n end\n 2y + 1\nend\n\n@btime f_avxt(10^6)",
"execution_count": 10,
"outputs": [
{
"output_type": "stream",
"text": " 375.300 μs (24 allocations: 464 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 10,
"data": {
"text/plain": "3.141590588883795"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "function f_avxt_sleef(N)\n y = 0.0\n @avxt for n in 1:N\n y += SLEEFPirates.sin_fast(float(n))/n\n end\n 2y + 1\nend\n\n@btime f_avxt_sleef(10^6)",
"execution_count": 11,
"outputs": [
{
"output_type": "stream",
"text": " 476.200 μs (24 allocations: 464 bytes)\n",
"name": "stdout"
},
{
"output_type": "execute_result",
"execution_count": 11,
"data": {
"text/plain": "3.14159058888379"
},
"metadata": {}
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "print(\"f_ccall(10^6): \")\n@btime f_ccall(10^6)\nprint(\"f_naive(10^6): \")\n@btime f_naive(10^6)\nprint(\"f_simd(10^6): \")\n@btime f_simd(10^6)\nprint(\"f_sleef(10^6): \")\n@btime f_sleef(10^6)\nprint(\"f_simd_sleef(10^6): \")\n@btime f_simd_sleef(10^6)\nprint(\"f_avx(10^6): \")\n@btime f_avx(10^6)\nprint(\"f_avx_sleef(10^6): \")\n@btime f_avx_sleef(10^6)\nprint(\"f_avxt(10^6): \")\n@btime f_avxt(10^6)\nprint(\"f_avxt_sleef(10^6): \")\n@btime f_avxt_sleef(10^6);",
"execution_count": 12,
"outputs": [
{
"output_type": "stream",
"text": "f_ccall(10^6): 31.138 ms (0 allocations: 0 bytes)\nf_naive(10^6): 11.924 ms (0 allocations: 0 bytes)\nf_simd(10^6): 11.913 ms (0 allocations: 0 bytes)\nf_sleef(10^6): 6.439 ms (0 allocations: 0 bytes)\nf_simd_sleef(10^6): 2.974 ms (0 allocations: 0 bytes)\nf_avx(10^6): 1.613 ms (0 allocations: 0 bytes)\nf_avx_sleef(10^6): 2.349 ms (0 allocations: 0 bytes)\nf_avxt(10^6): 353.200 μs (24 allocations: 464 bytes)\nf_avxt_sleef(10^6): 566.400 μs (24 allocations: 464 bytes)\n",
"name": "stdout"
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "log_ccall(x::Float64) = ccall(:log, Float64, (Float64,), x)\n\nfunction g_naive(N)\n y = 0.0\n for n in 1:N\n y += log(n)\n end\n y\nend\n\nfunction g_ccall(N)\n y = 0.0\n for n in 1:N\n y += log_ccall(float(n))\n end\n y\nend\n\nfunction g_sleef(N)\n y = 0.0\n for n in 1:N\n y += SLEEFPirates.log_fast(float(n))\n end\n y\nend\n\nfunction g_simd_sleef(N)\n y = 0.0\n @simd for n in 1:N\n y += SLEEFPirates.log_fast(float(n))\n end\n y\nend\n\nfunction g_avx(N)\n y = 0.0\n @avx for n in 1:N\n y += log(n)\n end\n y\nend\n\nfunction g_avx_sleef(N)\n y = 0.0\n @avx for n in 1:N\n y += SLEEFPirates.log_fast(float(n))\n end\n y\nend\n\nfunction g_avxt(N)\n y = 0.0\n @avxt for n in 1:N\n y += log(n)\n end\n y\nend\n\nfunction g_avxt_sleef(N)\n y = 0.0\n @avxt for n in 1:N\n y += SLEEFPirates.log_fast(float(n))\n end\n y\nend\n\ng_stirling(N) = N*log(N) - N + log(N)/2 + log(2π)/2 + 1/(12N)\n\n@show g_naive(10^6)\n@show g_ccall(10^6)\n@show g_sleef(10^6)\n@show g_simd_sleef(10^6)\n@show g_avx(10^6)\n@show g_avx_sleef(10^6)\n@show g_avxt(10^6)\n@show g_avxt_sleef(10^6)\n@show g_stirling(10^6)\nprintln()\n\nprint(\"g_naive(10^6): \")\n@btime g_naive(10^6)\nprint(\"g_ccall(10^6): \")\n@btime g_ccall(10^6)\nprint(\"g_sleef(10^6): \")\n@btime g_sleef(10^6)\nprint(\"g_simd_sleef(10^6): \")\n@btime g_simd_sleef(10^6)\nprint(\"g_avx(10^6): \")\n@btime g_avx(10^6)\nprint(\"g_avx_sleef(10^6): \")\n@btime g_avx_sleef(10^6)\nprint(\"g_avxt(10^6): \")\n@btime g_avxt(10^6)\nprint(\"g_avxt_sleef(10^6): \")\n@btime g_avxt_sleef(10^6)\nprint(\"g_stirling(10^6): \")\n@btime g_stirling(10^6);",
"execution_count": 13,
"outputs": [
{
"output_type": "stream",
"text": "g_naive(10 ^ 6) = 1.2815518384657895e7\ng_ccall(10 ^ 6) = 1.2815518384657895e7\ng_sleef(10 ^ 6) = 1.2815518384657895e7\ng_simd_sleef(10 ^ 6) = 1.2815518384658162e7\ng_avx(10 ^ 6) = 1.281551838465816e7\ng_avx_sleef(10 ^ 6) = 1.2815518384658225e7\ng_avxt(10 ^ 6) = 1.2815518384658162e7\ng_avxt_sleef(10 ^ 6) = 1.2815518384658169e7\ng_stirling(10 ^ 6) = 1.2815518384658167e7\n\ng_naive(10^6): 10.776 ms (0 allocations: 0 bytes)\ng_ccall(10^6): 8.716 ms (0 allocations: 0 bytes)\ng_sleef(10^6): 4.236 ms (0 allocations: 0 bytes)\ng_simd_sleef(10^6): 1.539 ms (0 allocations: 0 bytes)\ng_avx(10^6): 2.289 ms (0 allocations: 0 bytes)\ng_avx_sleef(10^6): 1.859 ms (0 allocations: 0 bytes)\ng_avxt(10^6): 464.400 μs (24 allocations: 464 bytes)\ng_avxt_sleef(10^6): 346.100 μs (24 allocations: 464 bytes)\ng_stirling(10^6): 22.992 ns (0 allocations: 0 bytes)\n",
"name": "stdout"
}
]
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "",
"execution_count": null,
"outputs": []
}
],
"metadata": {
"@webio": {
"lastKernelId": null,
"lastCommId": null
},
"_draft": {
"nbviewer_url": "https://gist.github.com/e73671622c6ee038286ce0fdb18a7407"
},
"gist": {
"id": "e73671622c6ee038286ce0fdb18a7407",
"data": {
"description": "sum of sin(x)／n",
"public": true
}
},
"kernelspec": {
"name": "julia-1.7-depwarn-o3",
"display_name": "Julia 1.7.0-DEV depwarn -O3",
"language": "julia"
},
"language_info": {
"file_extension": ".jl",
"name": "julia",
"mimetype": "application/julia",
"version": "1.7.0"
},
"toc": {
"nav_menu": {},
"number_sections": true,
"sideBar": true,
"skip_h1_title": false,
"base_numbering": 1,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment